AI isn’t the question anymore. Execution is.
In this episode of AI for the C Suite, Chad Harvey sits down with Chris Happ, CEO of Virtuous AI, to break down why most companies are stuck between knowing AI matters and actually making it work.
98% of CEOs say AI matters. Only 7% have a strategy.
So what’s going wrong?
This conversation goes beyond tools and hype. It focuses on what’s actually happening inside organizations and why so many AI efforts stall before they ever scale.
We cover: • Why most AI pilots fail • The gap between strategy and execution • Why buying AI tools doesn’t create an advantage • The role of trust in AI adoption • How companies should rethink workflows instead of optimizing broken ones • A real example of freeing up millions in cash through better decision making
If you’re a CEO or business leader trying to figure out how to move from experimentation to real impact with AI, this is a practical place to start.
Watch the full episode: https://youtu.be/sUcoylSLSzE Learn more about Virtuous AI: https://www.virtuousai.com/
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:01):
And looks like we are live here.
oh
I'm Chad Harvey and this is AI for the C-suite, the show for senior leaders who know AImatters and need to figure out what to do about it.
Each episode we dig into what it all actually means for mid-market organizations, how it'schanging decisions, strategy, leadership, and the nature of work.
(00:27):
Let's get into it.
My guest today has spent his career building companies that solve real operationalproblems in supply chain, in commercial real estate, in marketplaces, and then selling
them.
So when he tells you what works and what doesn't in AI adoption, he's not theorizing.
He's lived it on both sides of the table.
Chris Hap is now CEO of Virtuous AI, where he's focused on a problem that I suspect soundsfamiliar to many of you.
(00:54):
The gap between knowing AI matters and actually making it work inside your organization.
Before taking this role, Chris built and exited multiple businesses, most recently growinga marketplace to $10 billion in gross merchandise volume.
His organization has also recently completed a major research study on AI adoption in themid-market with Chief Executive Group.
(01:17):
And some of those findings challenge the conventional wisdom on where companies areactually succeeding and failing.
Chris, welcome to AI for the C-suite.
Hey Chad, how are ya?
I couldn't be doing any better.
How about yourself?
The same, off to a great start for the year.
People are excited.
(01:38):
AI's top of mind, so it's an exciting place to be right now.
It certainly is.
So since it's top of mind, let's talk about what's top of mind in that survey that Imentioned during the intro.
So you and your organization, I think it's called the state of AI adoption.
ah It's a mid market company survey that you conducted with chief executive group.
(01:59):
I'd like to learn a little bit more about that.
And I'd also like to know, was there anything in the findings that surprised you?
Yeah, so we polled the Chief Executive Network audience.
uh It's a uh group comprised of, I'll say, 10 million to a billion in revenue,cross-section of verticals, so a pretty good, we thought a very good view across the
(02:24):
mid-market writ large.
And uh so got a...
Quite a good response rate, more than we even expected.
I think a lot of people were really interested in not only uh learning about others, butbenchmarking where they themselves are at.
So got a great response rate.
(02:47):
the most, sort of the punchline that came out of this was 98 % of the CEOs, focused juston the CEOs.
98 % Yep, this is critical.
I know that AI is going to impact my business.
The flip side though, and probably what's most interesting to unpack here on this episode,7%.
(03:12):
So one out of 10 or one out of 12 said, yeah, we have a strategy.
So we all know it's important, but we don't really know what to do about it.
So that that's what I think if we, if we just
focus on what did this study uncover?
It's probably what a lot in the audience, why they're listening is what do I actually do?
(03:32):
I've seen this, I've lived this, I've used the individual tools and I've seen thepotential, but how do I unleash that in my organization?
And so you said one of ten.
want to make sure I get those numbers right.
One of ten has a strategy, but they're still flailing.
(03:53):
Yeah, so m it's 7 % of the respondents said we have a clear AI strategy.
We all think it's important, but yeah, so I guess it's more like one in 12 actually says,actually codified some sort of strategy or have a way or framework to adopt this in our
organization.
(04:15):
And what we, um the,
primary use case that they were saying that came back in the response was processautomation.
That's where we've started.
That's where we've had some sort of pilot success.
So I think of that as uh moving data between systems, taking rules-based process and moveit from sort of A to B to C.
(04:43):
Mm-hmm.
I really don't even think that's AI.
I think that's more RPA, robotic process automation that's been around for a while.
But I think that's really where people are at.
And it feels a little bit magical because it's, instead of people moving piles of paper ontheir desk, their machines are doing it.
(05:03):
But you and I know that's not really AI.
And that's really what they're seeing the most success because I think it's pretty easystill.
Do you think this opens up a larger question in my mind?
Do you think that the rise of AI and now the fact that this is being discussed at theboard and the CEO level in ways that previous technological advancements had not?
(05:27):
Do you think that that is showcasing or exposing maybe is a better word the lack ofadvancement in RPA adoption and some of the other techniques that have been around for 20
years in the mid market?
Yeah, actually, I hadn't thought of it that way until you just asked it.
But I think that's a great point.
And the fact that everybody's bored.
(05:49):
I 98 % of the CEOs say this is critical.
the board, this is there's board pressure.
Certainly even just individual pressure of, I know this could work, but that is forcingpeople, forcing executives to um take action.
And probably the easy action is I'm going to go buy some point off the shelf solution.
(06:13):
I'm going to buy something to sort of stop the immediate pressure that's being put on me.
And so, yeah, it's probably a forcing function for
some of these technologies that you and I know have worked for many, years, just becausethey're easy and you can have quick success with some of them.
(06:35):
Well, it's not the long-term strategy.
Yeah, we're clearly seeing that in the results.
People are saying this isn't a strategy, hey, checked a box at least on the, m in terms ofat a board level getting, um I now have an answer to what I've.
you know, to the question, what are you doing in AI?
(06:56):
Well, I've got some process automation happening.
Yeah, that's slightly terrifying that that's the answer uh from a lot of folks in that.
Well, you know, we're just going to go buy something and we're going to throw it in thereand then we're going to say box checked.
uh That sounds to me like a future zombie company.
(07:17):
Yeah.
Well, and if you, let's just play this out.
Um, so if I buy a tactical AI solution, meaning it's an off the shelf solution that, um,let's say my CRM or my ERP vendor provides by definition, everybody else has access to
that exact same solution.
(07:39):
So, um, what am I actually achieving?
So let's, let's take like a sales AI agent.
So now it can, it's going to sell better, faster, you know, send more emails, what haveyou.
Now my competitor bought the same thing.
So now they're both competing with each other.
So I'm burning a bunch of compute and time and effort, but I essentially have the exactsame solution.
(08:04):
So, um,
Not only do I have this problem where I have siloed data because I have a point solutionrunning and it's running a process, but my competitor is running the same thing.
So what did I accomplish there?
Really?
As you said, I probably created these like zombie AI agents that are off really making abunch of money for sort of traditional software vendors, but not really solving the
(08:30):
problem, which is do I have an AI brain that's forming in my business?
It's this is fascinating and I didn't anticipate our conversation going down this roadeven five minutes ago and we started but what you've just unlocked for me here.
Is yeah, I just concluded an interview uh coming out probably within a few weeks of uh ofwhen your episode publishes with pit Bingaman down in Australia and he's got a company
(08:59):
that's focused on providing and unlocking creative potential in AI and one of hissuppositions is that if we're all using these same AI tools, we're all going to end up
with the same default crap quite frankly and.
I just heard the software equivalency of that with you with respect to organizationsbuying these tools from their vendors and I hadn't thought about it.
(09:23):
You know, I think about the content homogenization.
I hadn't thought about the process and the technical kind of homogenization, which youjust shown a light on.
Yeah, yeah, I, I, from where I sit the, and what, um, I advise companies to do is it'sreally why we're focused on the mid market as well.
(09:43):
What, if you look at the enterprise with a massive IT budget and spend they're buildingtowards a, their own AI, their own, I really call it a brain.
Mm-hmm.
really building their own contextual layer of the decisions and the cause effects thathappen in their business that are relevant to, or they believe are relevant to their
(10:06):
success.
That's really tough for the mid market to be able to do because of the budgets, thetalent, oh the pace at which this moves.
So you're focused on running your business.
You can't be focused on building all of the infrastructure that is required to stand thatup.
But it doesn't change the fundamental problem that AI is better when it has contextualawareness and you want to amplify your special sauce or your business differences,
(10:38):
especially in the mid-market.
That's why you're there.
That's why you love doing what you do is that you're different by definition.
You're not a homogenous product that has sort of achieved mass market appeal that way.
And um
It's about amplifying that.
And how do you get the contextual awareness into a brain that runs?
(11:06):
We call it uh connected intelligence.
But how do you start to build that?
such that you don't look like everyone else and your AI and your processes reflect who youreally are and not what the software vendors or what the um off the shelf solutions want
(11:27):
you to be.
Cause that's really what we did that em it's been called the software industrial complex.
They'll push a solution on you, sell it at scale, and then say everyone's process shouldlook like this, which is fine.
But then every business starts to look the same and you start to morph to what they wantas opposed to what should my business actually do and where do I get leverage as a
(11:53):
business?
Well, I think that falls right into something else that you and I talked about in terms ofthe execution gap, the distance between your strategic goals and the actual delivery
required to meet those goals.
And I'm wondering what your thoughts are about that as it pertains to this concept thatwe're unpacking here around the software industrial complex and some of these kind of
(12:19):
pre-formatted canned processes and systems that are out there, be they AI or otherwise.
Yeah, yeah, I'm curious as well your thoughts on what you're seeing just in talking to thegroup too.
So from where my perspective, the execution gap is I know as a CEO, I need to be usingthese tools.
I can see the power of AI.
(12:41):
I know what I want my business to achieve.
but I don't have the resources or the capability.
I can't invest ahead of the curve the way others can.
If I could just put 10 more, what historically bodies on this or just some more technologyhere, I could get ahead of that.
um And it's just growing because the larger companies have bigger budgets and can attractmore talent.
(13:09):
So.
How do we close that?
And so it's not about throwing more at it.
It's about being smarter and creating leverage.
So for me, it's very much this walk, get my data into a system, a brain, run.
How do I start to use that to create automations and then really scale it?
(13:31):
How do I unleash this?
system that I've trained this AI superpower and then close the gap because I don't neednow I've created a resource leverage engine, which is now at the core of my business.
Let's talk a little bit more about that idea of connected intelligence and getting thatdata in the right format, the right structure, getting it organized to populate that brain
(13:56):
because that's an area where I see an awful lot of mid-market organizations.
I don't want to say fall down, but it just hasn't been a priority.
And we're generating mountains and mountains of data, even, you know, smallerorganizations do this, yet it's not organized very well.
So what are you finding as you walk into companies uh in terms of their data hygiene andtheir data organization and how well or not well that's preparing them for this next level
(14:25):
of connected intelligence?
Yeah.
The default answer I get, is unfortunate, the way to clean up data is an ERP upgrade,which is terrifying, right?
We both know the cost overruns, the...
um
(14:47):
Effort it's always too long.
It's I've never met anyone who wants to do another ERP upgrade and so if that's the solveWhat you're basically saying is okay.
I'm gonna go put make my business look like Whatever the ERP wants it to look like and I'mgonna In essence take all the old mess and clean it and put it into there and move forward
(15:16):
The way I advise people is doesn't matter.
Absolutely doesn't matter.
I don't, I think the role of the ERP as a system that stores source data and puts a webinterface on it, it'll get, it'll move back to that sort of relegation.
be relegated back to that as opposed to this center of the organization tool.
(15:41):
um And so,
Where I advise people is it is actually really easy to get data now.
So the state of the hygiene is poor, but just that's what it is.
There's on average in the mid market studies will tell you that you have 50 differentapplications running.
(16:02):
So 50 silos of data costing roughly $2,000 per person per year.
So you can quickly do the math on what it's costing you to have all these silos andpockets of data, let alone the
hygiene that you asked about.
The great news is it's really easy now to pull data out of all of these systems and use AIto pull it in and make sense of it.
(16:31):
So a brain is actually really easy to create.
So what do we want to take orders, customers, cause effect, things that matter in thebusiness.
I don't need all of Slack or all of my email.
I want key decisions and key things.
And I can pull that into a semantic or contextually aware layer that I can start tointeract with.
(16:58):
there was this push where everybody had to have a data lake.
And it's still happening, data warehouse, lake house, whatever name the vendor's pushing.
To me, it's completely irrelevant to do that.
The tools we now have access to with AI make it that easy to pull your data in.
(17:23):
And if you look at what a dashboard signified, it's basically, I know what problem I'mlooking for.
I'm going to build a report that looks for it.
I'm going to make it red when it happens.
And then when I see it, I'm going to have to click three times in to sort of diagnose it.
So if you know all that ahead of time,
Why wouldn't you just ask an AI to look for that?
(17:46):
So seek out that data and then send you an email or a text or something.
So we've, we, I don't think dashboards even work.
look backwards.
if the implementation plan is, I'm going to put all my data.
I'm going to upgrade an ERP.
Then I'm going to put a business intelligence layer on top of that.
Cause the ERP doesn't do that very well.
(18:08):
And then I'm going to look backwards at data that I already know is a problem.
you problems I've already solved.
just going to look for them.
Now I'm two and a half years into something that, you know, we all agree is going tochange in the next three or six months.
So we've, we've created a concept in our, our, um, Bayo platform called a flash board.
(18:29):
So it's real time.
It's dynamic.
It's fluid information.
You asked the system, what should I know about today?
And it's just connected intelligence.
looks at your business.
It says, here's what
Inventory there's an issue.
see an inventory.
Here's a visual.
Here's what I think we should do about it um And that's just by wiring data and let alonesort of automations and agents all working together but the I think people uh companies I
(18:57):
talked to Have been Scared away from or or the salespeople.
I'll just say maybe put a pause
There's some great salespeople in the business, uh, intelligence world and ERP world whomake that the solve as opposed to let's just go at these sources.
(19:19):
There might be 50 pull the raw data in and start to look at it.
So let's uh let's make this real and uh correct me if I've got the company wrong, but Ibelieve you worked with Mugsy direct to consumer.
OK, OK, good.
I got that right.
So I I think what you're talking about in some ways, maybe this would be a good uh exampleto make this more concrete.
(19:46):
So Mugsy direct to consumer fashion brand.
uh I think you worked with them on inventory demand planning.
What did the problem look like for them before AI?
And then what actually changed?
And what was the measurable outcome?
So maybe let's use this as almost a little mini example or case study here.
Yeah.
So in talking with our CEO, had an initial, his initial thesis was AI, not too dissimilarthan we started the conversation.
(20:20):
AI is important, but I don't, I'm not really sure how to begin this journey.
So we take them through a walk, run scale.
Walk is all right.
What are the most, what are your 50 systems?
They had 14.
You could think Shopify, Amazon, Meta, Google, email, automation, uh inventory.
So they had 14 systems and something as simple as what was sales yesterday.
(20:46):
If you're not talking to all those systems, you're not going to get the same answer.
I sold something in Amazon.
I sold something in Shopify.
I sold something in my retail location.
And those are multiple different systems.
I don't need a business warehouse.
They didn't have a business warehouse.
We connect individually to those systems.
(21:10):
genius of AI is that it's been trained.
It can pull data from thousands of systems without any code.
So it's easy to connect to those.
And what are you fundamentally pulling in, in order?
Inventory, a customer.
you right there because I'm sorry to interrupt your flow, but what you just said remindsme of an objection.
I see red flags getting thrown all the time by CEOs based on the type of statement thatyou just made where AI doesn't need to know or it pulls these thousands and everybody
(21:39):
always says some version to me of, how can I trust that?
Yeah, yeah, exactly.
Well, um it's a great question.
how does it, um there's a couple of layers to in that.
um I would not trust my critical business data with a um AI platform or a sort of an openAI, you know, off the shelf AI platform.
(22:12):
I just wouldn't put my data in there because they're gonna learn, I don't suggest theylearn on your specific data, but they learn how your business runs.
So maybe less the price, but more importantly, how do you sell, you know, in this case, Dto C genes?
Now it doesn't mean they're gonna build a business from that, but as they watch thepatterns, the AI gets smarter and smarter.
(22:37):
So I personally don't allow the AI to do that.
What we do do though is we don't write code anymore.
So it used to be, you would have a programmer.
need to connect to HubSpot to pull my customer data.
And then I need to pull connect the net suite to pull my sales data, whatever it is.
(22:59):
Because that those APIs are available and open AI knows how to write the connection to goback and forth.
So I don't need massive teams.
programmers, it's not a three month, Hey, connect net suite and other three months connecthub spot.
So that's easy.
Um, so that, that is no longer a skill.
(23:21):
That's a commodity that, that we can, I don't need to pay developers for, but the datathat comes through that connection, I think that's really important.
That's your data.
That's private.
That should live private.
That's, that's part of our, the way we solve it for the mid market is that all that data,all of that polling of that data.
runs in a private SOC 2 data center where you're not sharing that.
(23:45):
So I think it's a really good point that the actual code that gets written or the way topull the data AI is doing, but the actual pulling of the data is private and secure.
And so it shouldn't be a concern, but I certainly wouldn't just go into ChatGPT or Geminiand say, here's my password, go pull all my HubSpot data, go pull all my sales data.
(24:09):
Here's my pricing strategy because what would you do?
You know, ask the New York times how this goes, right?
Like what would you do?
You are training on your IP and that's just the way AI needs to work is that it needsdata.
It's very data hungry to, to, um, to get smarter.
Alright, I appreciate you taking that detour with me.
(24:30):
I know I interrupted your flow.
uh I had asked you about outcomes and what actually changed as a case study with Muggsy.
So thank you for that because that again, that is a question that comes up all the time.
So kind of back to you on the journey with Muggsy.
Yeah.
yeah, I'm glad you asked that.
I think it's critical that this runs, that people understand the um potential there for umprivacy concerns.
(24:55):
um Yeah, so as we thought about the Muggsy example, so step one is let's connect ourbusiness and end up with a, I don't like this term, but
This is the way people often think about it is I've chat GPT for my business.
Now I have this AI.
(25:17):
can ask any question.
Hey, what was sales yesterday?
And not only does it pull the data from each of the different sources, it knows how Icalculate sales.
So just something fundamental is everyone does it a little bit different.
And, you know, it's funny, a lot of these like.
Harking back to when they started the business, there was some quirk and you know, that'sjust permeate.
(25:40):
throughout the business.
um So that is step one.
And a lot of times what happens in that step one connected intelligence is we take thespreadsheet that every CEO has that secretly runs the business.
And there's data from different sources.
I've got oh sales, I've got marketing, Roat, those different things.
(26:02):
And that now becomes this flash board where I can ask, hey, what's going on with thebusiness?
I don't need to spend.
an hour each morning compiling the Google sheet or the spreadsheet with all the differentdata sources.
So now I have, I can make decisions faster.
I'm confident.
I've built trust that I'm actually looking at the right information.
So for Muggsy, that was step one is before I'm going to make any decision, I got to justfundamentally trust that this, information I'm getting and the, um, it has access to
(26:38):
my data and the way I think.
So that's where the brain really is formed.
And while it's not hard to get the data in, that's probably a, you know, if I'm honest, a30 step process or 30 day process where you're just sort of having a back and forth and
sort of building trust at an executive level that you're getting the answers and theinformation that you would expect from a
(27:07):
senior data person.
And I think in the mid market, what happens is that people are siloed.
It's actually in the enterprise as well.
So the benefit here too, is that you get the same answer every single time, because thesedon't live in silos the same way.
When you ask marketing, what was sales, you get a different answer.
(27:28):
um Then when I ask the COO, I'm going to get a different answer.
And it's just skewed.
It was actually funny.
Our entire company, they live in AI.
We use it.
It's just part of who we are.
If we don't embrace it, then how can we expect our customers?
(27:49):
Even when my head of sales built a plan for me and my head of marketing built a plan forme, when I asked the plan sort of came back in terms of like we're in the funnel and the
percentages where the things move through the funnel.
Even building with the same AI, I got a marketing skewed answer for marketing, i.e.
(28:11):
more deals fell out in the middle of the funnel where sales takes over, then sales lesscomes in the top of the funnel and they're more effective.
So just how they were working with the AI in their own way skewed what the AI gave it andskewed their version of the plan that ostensibly came from ChatGPT or Gemini or whatever.
(28:31):
So
It's really important to break that down and get a consistent view of your business that'sunbiased.
And that's really what happens with this connected intelligence phase.
So that problem is solved.
Now all the while we had a use case in mind, which was, hey, we have too much inventory.
(28:53):
The easy way to plan is to say, here was sales last year.
I want to beat that by 30%.
So let's just take
and multiply every number, everything we bought last year times 1.3.
And now that's going to be our way to plan inventory.
And you can put some logic around, you know, we know that um in the fall around postThanksgiving to kind of the end of the year, we have more sales.
(29:21):
So we'll do a little bit more then.
But that's sort of the planning process.
And um as you start to m
go, what is the impact of buying that way?
Now I start to have millions of dollars in inventory that um wasn't very strategicallyplanned.
(29:43):
What could I do instead?
If I had more cash sitting around, I could put it in marketing.
So if I can be smarter about inventory, particularly in retail, I can make massive impactson the business because there's an obvious.
correlation between the amount of money I spend in marketing and then when it sells.
So the more cash I can free up, we have a massive impact.
(30:07):
Particularly as you think about sort of the times we're in now.
uh
you know, tariffs and inflation, the value of getting this right becomes more moreimportant.
So that was, that was the goal.
we, didn't start saying, Hey, we just want this brain that's smart.
want to start to be able to solve problems like this.
(30:28):
And that's sort of where, as we get to this run stage and the way um we're able to startgetting there is let's look at, let's let the model start to predict inventory.
instead of this top down,
sort of goal-based approach.
Let's look at the history of the business and then ask a machine to predict how inventoryshould look.
(30:55):
So the result for them was, well, hey, we've freed up several million dollars in cash.
As we back test it, it's about 95 % accurate to what actually happens, what it thinks willhappen to what actually happens.
And so it frees up this
sort of 70 % sort of gut check accurate to this very close to reality.
(31:22):
And just enables different business decisions to be made.
that's what he said.
Yeah.
Yeah.
that's huge.
have come across, and I don't want to make it sound like an overblown number, but I'vecome across numerous businesses over the years, profitable businesses that went out of
business in the mid-market space because of cash management.
(31:42):
And what you're saying here is, I know we're talking about a retail example, but anybodythat's got inventory, if you can free up several million dollars of extra cash uh by...
using the types of systems and putting the type of uh business intelligence in place thatyou're talking about, man, that's huge.
It's amazing.
It's amazing.
And it's just sitting there.
(32:04):
And it's really not that hard.
We connect in the data.
Now what do I have?
I have my order history.
I have my marketing history.
I have the different things that have happened.
And I have time.
So I have a history with that.
And now we ask the machine to run through that in a way only a machine can.
It can run massive compute in ways that our minds can't.
(32:27):
And then it says, here's what I think is going to happen.
And it gets smarter and smarter as you can add more more factors.
But it turns out in less than a couple of months, you can have a very accurate predictionof what's going to happen.
Then you could start to add in more and more more things.
Hey, this what happens when I have an influencer wear a pair of jeans?
(32:50):
Let's introduce that to the model.
We don't need to start with that.
We can still be 95 plus percent accurate, free up millions of dollars.
But
then you can really start to do some exciting things.
think sometimes people kill these projects with...
over complication or perfects the enemy of good enough kind of thinking, you can getreally accurate and much better than you were doing.
(33:12):
Just think of the difference we said there.
Yeah, and let's talk about that here briefly because I have seen a lot of AI pilots thatgo nowhere.
And oh, and by the way, I wrote something down here before I get to that.
uh Just in case anybody uh kind of drove by the point that you made, I wrote this quotedown.
You said the spreadsheet that every CEO has that actually runs the business.
(33:35):
And when you said that, it just rang so true to me because I walked into so manyorganizations.
They got these complicated dashboards.
They got metrics.
And then here's this Excel sheet of this Google Doc that the CEO is actually using to runthe business.
So you're definitely on point with that.
um
for most people.
like, let's just get that into a AI system you can ask questions of.
(33:58):
Exactly.
All right.
So uh with that little detour aside, let's talk about AI pilots that go nowhere becauseyou kind of referred to that a moment ago as we were concluding some remarks there.
I think you've got the idea about having a right to left or an outcomes based approach.
um But let's talk about AI pilots that go nowhere.
(34:19):
Let's talk about what the right type of approach is um for
Folks out there, maybe they've got AI experiments that have failed or pilots that wentnowhere.
What can you speak to about that here?
Because I'm seeing more and more of that.
Yeah, I do like the um right to left approach.
Start with an outcome that matters and then work it back.
(34:44):
And so that could be the inventory example we gave.
It could be the Excel spreadsheet.
that, let's have a system I can interact with that understands that, key pieces ofinformation or data sources.
um Why I think they fail,
I'll give two reasons.
(35:04):
One, there's not CEO buying.
Mm-hmm.
If you are checking the box, like we said before, and the organization knows that I thinkit's going to end the way we all were most of these pilots end, which is you're going to,
(35:25):
it's going to die in sort of organizational inertia because this is different.
This is not implementing a new, report or an ERP system.
AI fundamentally is a
change to the way you do business if you believe it at a strategic level.
It's a thought partner.
(35:45):
It's a strategic resource that's always on with the ability to run massive compute andsimulate a lot of different examples and then start to give you options, which is just
different to the way businesses run.
We come in, we work nine to five, we
(36:05):
have silos of information, we operate on that and sort of, so if the CEO doesn't buy in, Ithink that's where these will die.
A lot of times you have people sort of in the middle of the company who really don't wantthese to be successful because it's gonna change what they do.
(36:26):
I don't mean they're gonna be out of a job, it's just gonna change the way they functionby the nature of it, which.
You could view as very powerful as an individual or you could view it as very scary and Iunderstand that.
So that's one reason.
The other reason is what we said before.
If you're tactical about this.
(36:47):
Mm-hmm.
You're going to end up looking like everyone else and you're going to end up with, Ithink, mediocre C plus results.
AI is trained on a crowd sourced set of data.
So if you buy that off the shelf product, you're going to get a crowd sourced answer toyour business, which is fine.
(37:09):
AI is smart.
It knows how to build a sales plan and a marketing plan, but it doesn't know how to buildyour sales or marketing plan.
If you just look like everyone else, then what do you have?
And so then the pilot doesn't feel very good.
You can see the potential, but when you actually say, well, people are better because theyknow the nuance, they know how to get it from C plus or B minus to A.
(37:33):
We did a pilot.
was, my mind was blown in this one um for a customer that was doing.
It was a drone company that was working on helping farmers regenerative farming and AIcould run a farm just with satellite images to I would say like a C plus B minus level,
(37:58):
like an individual farm.
actually told a farmer they had pumpkins growing on the, that they didn't realize the sunhad like planted these pumpkins, this little patch of pumpkins.
And it knew it just from the satellite images.
I figured that out.
It still wasn't going to beat the farmer individually.
But if that farmer would then say, I've got this really smart thing that if I, if I workwith it a little bit, wow, the power I could, I could have, but that's where the pilot
(38:25):
sort of moves from.
Um, it's really pretty good, but how do I make it great?
There's work there.
And, and if you just buy something off the shelf and expect it to do better than all thepeople you have and
Importantly, it doesn't have all the history of your business and understanding and thecontext of your business.
If you don't give it that it's never gonna move off the, you know, pilot phase.
(38:50):
And that's where I think people expect it's easier to just buy something than to say thisis strategic.
I'm gonna
wire in the intelligence as part of the phase one, because you don't get that sexyoutcome.
So if you start with the outcome and say, this is what I want to achieve, but I'm going toalso lay the foundation.
(39:11):
What we'll actually do with our customers, we don't let them move on and don't let thempay us more money until they move through the phases, the walk, the run phase, and the
scale phase.
So in walk, we say, we're not going to write data to a lot of other systems.
We're going to
help you build trust and get the right answers.
(39:31):
Then as you run, we're going to start to uh automatically write information from system Ato system B.
And then the final stages, you're starting to let the agents make decisions and reason andstart to do things on your behalf.
But you have to build trust to do that.
So if you come in and just say, I'm going to buy a solution and I'm going have all theseagents running and my business is going to look better in
(39:59):
overnight, it's just not going to work.
It's going to look like every other business.
It's going to look like the software company and those agents are optimized to sell youmore licenses.
It's kind of the end story, right?
It's going to use more compute and they're going to optimize for that silo, but notunderstand your business.
(40:21):
So everything that you just walked through, I would imagine that you went through withyour own company as well.
I believe you had said to me you went from a target of a million dollars in revenue peremployee to a target of about 10 million per employee with AI.
Is that right?
Okay.
All right.
Wanted to make sure I had that right.
So as you went through that, right?
(40:45):
No, no, no.
Yeah, I recognize that.
And that's the interesting part about so much of the AI stuff is many of it is or much ofit is still aspirational in nature, um yet it's it's aspirational um at a level that is
obtainable as I guess is what I would say uh from my observations with folks.
(41:07):
Yeah.
Well said.
Well said.
Well, and I love your philosophy on this, the learning philosophy.
This is so new.
We're all learning and I make mistakes all the time.
And, but I think the biggest you can make mistake you can make an AI is not getting in thegame and learning.
Wanted to start building this brain.
(41:29):
You take, know, my kids now are one just turned seven.
So I'm, four and seven.
Imagine if you didn't start teaching them anything until they're four, you've just pushedthe 18 years out.
Now they're 22, you you've got to start teaching these things.
You've got to start making mistakes.
There's no, um, there's no other way to start working with this because the system cravesyour data to be, uh, able to best help you as a business.
(41:59):
I love that example and kind uh of drawing your kids in and the idea of education andlearning delays there.
And the first thing that popped in my mind when you were talking about that would be.
As if I didn't start speaking or teaching someone English and I'm just going to throw outsome some random dates here and it's 1752 right?
(42:21):
And with the pace of AI, if I wait a couple years to dip my toe in when I'm ready to teachthem English, they're going to be speaking in 2026 vernacular, which sounds nothing like
300 years ago.
Yeah, isn't that a hundred great point great point and you know what I hear to follow thaton is I'm just gonna wait because the AI is gonna get better.
(42:43):
All right.
Well now the AI the gen AI tool got better, but it has zero context about your business.
So again, I think a lot of people
And I don't expect they would understand it.
I've naturally had to learn a lot about how this works.
But if you have a generative AI model, which is amazing at predicting words and is amazingat human language, I'm mesmerized to the conversations and the back and forth and the way
(43:12):
it um uses our language with us.
But um all that said, it has zero awareness unless you give it that.
As you're asking questions and having a back and forth that has zero transactional historywith you.
So to your point, how do you build that up?
(43:34):
And then as the Oracle that you're asking is that keeps getting smarter.
You want to keep a constant set of learning such that you're not getting a differentanswer every single day because that that field goal or that
ball is constantly moving.
And but you want to have a way of getting that same answer.
(43:56):
Sales yesterday can't change because the Oracle changed, right?
So you have to have a way of teaching it and then using knowing that it's going to getsmarter, still get the same answer.
So it's a really interesting paradigm shift there too.
It absolutely is.
And I think it also opens up this line of discussion opens up the larger question ofstrategy in an organization and the shelf life for that, because I was just a guest uh on
(44:27):
a pod the other day.
I don't do a lot of those, but they're wanting to ask me about strategy and timelines.
So we were talking about some of this.
And one of my views is that most strategic plans.
um
If it's an 18 month shelf life uh and that that assumes it's dynamic, that's a long shelflife anymore.
There are no 10 year, five year, three year plans anymore.
(44:47):
And I think the point that I'm driving to here with a question based on what you said isthat if you wait to get in the game with AI, you're not only sacrificing momentum, you're
sacrificing agility because the speed at which uh changes are coming down the pike now issuch that there are decisions
that are happening outside of your purview based on AI within your company if you'veimplemented it, right?
(45:13):
So if you are not implementing AI, if you're not using these tools, you're sacrificingspeed and you're increasing the level of complexity regarding decisions that your mid and
senior line uh supervisors and managers need to make in my view.
And that's gonna further drag you down.
and compound your uh or hamstring your ability to be agile and to move quicker based onmarket forces that are changing extremely rapidly.
(45:42):
So I'm not sure I got myself to a question with that, but that was kind of top of mindbased on what you were talking about a moment ago.
Yeah, I think you described better than I did the execution gap.
You just keep falling behind.
Other people are getting ahead faster and it's compounding because it's smarter and fasterand they can make decisions faster and you're still doing it the old way.
(46:06):
If you go in wanting perfection, you won't achieve it without some failure.
So you should go in saying I'm going to start now and I'm going to fail along the way, butit's going to close the gap because others are doing it.
And if I'm not, look, 98 % of CEOs said, I need this.
(46:26):
7 % have a strategy.
do you, could plan for 18 months and build a strategy or you could get in and do, and thestrategy will form because this is, this is new.
There is nobody, there's not a
silver bullet strategy.
there was this, we wouldn't be seeing that execution gap forming.
(46:46):
Right.
What's that quote from Patton?
always butcher it, but it's something along the lines of an imperfect plan violentlyexecuted today is far superior to a perfect plan executed, you know, far down the road.
Right.
So same idea.
So.
Go ahead.
(47:07):
I was just going to ask you, since we're talking about this change and we were talking ata strategic level, we touched on the human component earlier and I know you referenced
we're not here to automate you out of a job.
I think 26 is the year we're going to start to see an awful lot of pushback.
We're already seeing it.
And I've had some conversations with other folks that specialize in AI training fororganizations.
(47:31):
They focus a lot on mindset shifts.
They focus a lot on
uh demystifying the technology and reducing the level of fear.
I would imagine that um all of these things come up when you're working with your clientsthough.
it's one thing to talk about this in the boardroom and to set the course of the company.
It's another when you've got to handle those actual conversations with the workforce andthe line level team members.
(47:56):
So I'm interested, what's been your experience so far with those types of conversations?
with the folks that are on the front lines doing the work.
Yeah, I suppose there's two, there's a carrot and stick approach.
You could see the CEOs who will take the approach of you will use this tool.
(48:20):
You will report back to me if you didn't, because I'll see that you have zero chat historyand we know how that ends.
So that can work.
And then there's more of the carrot.
How do I lead someone?
How do I, I, um, show them or, um, start to pull them towards how this can make their lifeeasier, better, um, help them use it to achieve the outcomes they want or the business
(48:50):
wants.
think largely the people I talked to don't want to do the repetitive.
manual non-creative tasks So I think there's a natural inclination to want to work withthis stuff Once you get over the initial of of wow, it just did that um I Think if you can
(49:20):
get people there they want to work with it that that's generally my my belief I will say
It's changing for me, what I value in the workforce.
resumes are less important to me than they ever were.
Uh, raw curiosity and creativity, um, given that the answers already exist in these tools,the sort of book answers, but it's, do you do with that?
(49:53):
And how creative, how can you 10 X yourself with these tools?
Mm-hmm.
personally value that in our business.
Now granted, we're in the business of bringing AI to other people.
So we're at the tip of the spear there.
But I do think that's really important as people are thinking about careers and careerchanges and as we um work with interns.
(50:19):
We do a lot of interns just to try to pull them through and teach them how this works inum business.
But I think that skill set
of being told what to do versus I'm going to be creative and I've got this set of tools towork with.
That's a shift.
And so that I don't know that there's a.
(50:39):
em
Certain people are gonna love that, other people aren't, and there's different points inthe career where you may say, I don't wanna do that anymore, I don't wanna shift, I just
wanna kinda see this through, I've got three more years left.
And so that piece of the workforce, I think, gonna, that'll be interesting how that playsout.
(50:59):
Mm-hmm.
I couldn't agree more.
Our time is uh starting to get close to an end here.
And so I want to shift gears for maybe our last subject uh point of conversation and askyou about values, because this is something that I've been tracking very closely since the
(51:20):
fourth quarter of 2025.
uh Anthropic, the company behind Claude, came out with the sole document, as they calledit in late 25 and in early 26 in January, they came out and published their 77 page, what
they call the
Constitution.
This is an area that fascinates me.
um
And it fascinates me precisely because of what we talked about earlier with respect.
(51:44):
And there's a parallel here to what you talked about when companies buy ERPs, they buylarge software systems.
You're asking for the same pre-formatted processes that everybody else is using.
And I think there's a parallel here with the values and the judgment calls that AI ismaking with respect to the values that are inculcated into those different models.
(52:04):
uh In my view, think Anthropic is doing a very good job of
trying to grapple with this question and being relatively transparent.
But I am curious whether this conversation comes up much with your clients in terms oforganizational values.
How do we infuse those into the tools that we're using?
How do we make sure that uh what is important to us and
(52:28):
what we value as an organization, mission, vision, values is then seeded into theautomated AI systems and tools that we're using.
I don't expect that you've got a silver bullet answer, but I'm interested in yourexperience on this, because I think this is an undeveloped conversation that is super
critical that not enough people are having yet.
Yeah, I love that you're you've always been I think thought leading on where does ethicsvalue where does that meet AI because to your point we're trusting the answers that come
(53:01):
from these oracles and whoever's trained there's a bias inherently in how it's trained andfor example we won't we don't use any of the Chinese models I don't
Hmm.
Yep.
what's in there.
um And that's just a stake we've put in the ground.
(53:24):
um Because of the reasons we said, the model is trained by people and processes and ondata.
I don't know.
And so I don't trust that the answers are going to be based in fact.
Now.
that it doesn't change that you can have biases by the programming team and the data setsthat we're training these on, even in the, you know, Anthropic or open AI or Gemini or any
(53:50):
of those.
So um the question is, is real.
And that's where the most important thing you can do as a business.
And this is different than a consumer level.
As a, when I'm interacting with a
Gemini or AI on my desktop, I have my own values to gut check it and I have my own versionof the fact that I encourage everybody not to take what it says at face value.
(54:22):
Their aim is their aim, their design to make
you feel like you're right and kind of pull you through a conversation and please you.
So I encourage people individually to challenge them and ask, is that real?
Do I believe that?
I mean, there's still prediction engines, so they'll get the wrong answer.
And if you go with it, it's just going to perpetuate that.
(54:46):
So what does a business do?
Because me as an individual, I can do that.
I have that choice.
But a business can't.
put those guardrails in.
So if you don't have core values and vision and mission and all that stuff as guardrails,your agent and employees who are interacting with this are going to biased answers or
(55:09):
they're going to get their own answers, which don't necessarily reflect the values of thecompany.
So it all goes back to building a corporate, a corpus of information, a brain.
That is absolutely critical that sits in front and independent of any of these modelswhere that lives and those guardrails live and they're indisputable.
(55:34):
You cannot deviate from those.
That way we don't have, are stories you can flip open the news.
just saw one, Air Canada was held liable by the Supreme Court for their agents.
They said, well, we didn't write them.
Well, sorry.
They're acting on your behalf.
They gave $2,000 refunds.
That's you're liable for that.
(55:55):
So how do you build?
And it's critical.
What is the brain that you're building that guards against that from one of these modelsthat we're all using?
Eventually, I think everyone will start to build their own smaller AI models that we usethat have that inherently built in.
But right now, what is it?
(56:17):
The cost of doing that is just
It's just not feasible for anybody.
So we all use these massive models.
They're great, but we get their biases.
So I love that you've brought this up because this is where it's going is how do we knowwhat's coming?
How do we know why we got the answer?
What data was was put in there?
(56:39):
And so that's where as a business leader, I think you need a business AI, which isseparate than.
what we use like a co-pilot on your desktop or something that helps me build a betterPowerPoint or write emails faster.
I need a business rules that it follows.
Well said.
All right, let me bring us the full circle background or we started here with a partingshot for you.
(57:05):
We started talking about the the one in the 12.
Let's talk about the other 11 uh as a final thought here for our listeners.
So if I'm uh 11 of the 12 and I am not doing what I probably should be doing uh in AI landright now, and I'm a CEO listening to this.
What's something you'd recommend if I'm in that 11 category that I should do in the next30 days to start to move the needle in my organization?
(57:32):
Yep.
Well, one, take ownership yourself.
head of sales, of course, this comes from the sales guys said what either you think thisis strategic or not.
If you don't think it is, so you're in the 2 % even, I don't think this matters.
Put your head in the sand.
Don't let's don't have the conversation.
(57:53):
If you think it's strategic and you believe this has the potential to impact your businessin a massive way.
When, where else would you have that happen and not own that as the CEO?
So make that decision that you're going to own it and not punt it to it.
Cause this is not an IT problem.
(58:14):
The IT parts easy.
So the next 30 days, I would say, do I care about this?
Is it strategic?
Will it disrupt my business?
And if so, I'm going to own it.
At that point, you'd say,
Do I want to own it by buying a bunch of piecemeal tactical solutions or am I going tobuild or buy a solution that is a strategic platform, a brain, an AI system that runs in
(58:42):
my company and learns what my company operates like, how I want it to operate, andeverybody can use it.
I would get to that in the next 30 days.
So where that ends for the mid-market is, do I build one?
Because you could build what we've done.
You absolutely could.
Or do I buy one?
We're the only system out there that does what we do.
(59:06):
We call it a BAO.
There'll be others that do it.
But if you get to, want this to be a strategic component of my organization, I'm going toown it as a CEO.
Now I got to go solve that problem.
That's what I would do in the next 30 days is get.
Reason your way through those answers and then start making decisions around that whichwill lead you to all right now I have a strategic plan and it won't be I'm gonna start
(59:37):
buying things It'll be I'm gonna start with a plan in mind and we're gonna ebb and flow onthis journey because it will It will change as technologies emerge
Well said and sage parting advice from Chris Hap, CEO of Virtuous AI.
Chris, I appreciated our time together today.
If someone wants to learn more about Virtuous AI and your offerings, if they want to getin touch with you, what is the best way for them to do that?
(01:00:04):
Yep, it's VirtuousAI.com.
And if they want to email me, I'm chapp, C-H-A-P-P, at VirtuousAI.com.
Very good.
Thank you so much.
This was a great conversation.
I appreciated your perspective and your time today.
Likewise.
And for our listeners out there, thank you for once again listening to AI for the C-suite.
(01:00:27):
If this episode was useful, subscribe wherever you get your podcasts, follow us onLinkedIn, and check out AIfortheC-suite.com.
Until next time, keep your algorithms running, your leadership evolving, and your AI incheck.
Take care, everybody.
And looks like we are live here.
oh
I'm Chad Harvey and this is AI for the C-suite, the show for senior leaders who know AImatters and need to figure out what to do about it.
Each episode we dig into what it all actually means for mid-market organizations, how it'schanging decisions, strategy, leadership, and the nature of work.
(00:27):
Let's get into it.
My guest today has spent his career building companies that solve real operationalproblems in supply chain, in commercial real estate, in marketplaces, and then selling
them.
So when he tells you what works and what doesn't in AI adoption, he's not theorizing.
He's lived it on both sides of the table.
Chris Hap is now CEO of Virtuous AI, where he's focused on a problem that I suspect soundsfamiliar to many of you.
(00:54):
The gap between knowing AI matters and actually making it work inside your organization.
Before taking this role, Chris built and exited multiple businesses, most recently growinga marketplace to $10 billion in gross merchandise volume.
His organization has also recently completed a major research study on AI adoption in themid-market with Chief Executive Group.
(01:17):
And some of those findings challenge the conventional wisdom on where companies areactually succeeding and failing.
Chris, welcome to AI for the C-suite.
Hey Chad, how are ya?
I couldn't be doing any better.
How about yourself?
The same, off to a great start for the year.
People are excited.
(01:38):
AI's top of mind, so it's an exciting place to be right now.
It certainly is.
So since it's top of mind, let's talk about what's top of mind in that survey that Imentioned during the intro.
So you and your organization, I think it's called the state of AI adoption.
ah It's a mid market company survey that you conducted with chief executive group.
(01:59):
I'd like to learn a little bit more about that.
And I'd also like to know, was there anything in the findings that surprised you?
Yeah, so we polled the Chief Executive Network audience.
uh It's a uh group comprised of, I'll say, 10 million to a billion in revenue,cross-section of verticals, so a pretty good, we thought a very good view across the
(02:24):
mid-market writ large.
And uh so got a...
Quite a good response rate, more than we even expected.
I think a lot of people were really interested in not only uh learning about others, butbenchmarking where they themselves are at.
So got a great response rate.
(02:47):
the most, sort of the punchline that came out of this was 98 % of the CEOs, focused juston the CEOs.
98 % Yep, this is critical.
I know that AI is going to impact my business.
The flip side though, and probably what's most interesting to unpack here on this episode,7%.
(03:12):
So one out of 10 or one out of 12 said, yeah, we have a strategy.
So we all know it's important, but we don't really know what to do about it.
So that that's what I think if we, if we just
focus on what did this study uncover?
It's probably what a lot in the audience, why they're listening is what do I actually do?
(03:32):
I've seen this, I've lived this, I've used the individual tools and I've seen thepotential, but how do I unleash that in my organization?
And so you said one of ten.
want to make sure I get those numbers right.
One of ten has a strategy, but they're still flailing.
(03:53):
Yeah, so m it's 7 % of the respondents said we have a clear AI strategy.
We all think it's important, but yeah, so I guess it's more like one in 12 actually says,actually codified some sort of strategy or have a way or framework to adopt this in our
organization.
(04:15):
And what we, um the,
primary use case that they were saying that came back in the response was processautomation.
That's where we've started.
That's where we've had some sort of pilot success.
So I think of that as uh moving data between systems, taking rules-based process and moveit from sort of A to B to C.
(04:43):
Mm-hmm.
I really don't even think that's AI.
I think that's more RPA, robotic process automation that's been around for a while.
But I think that's really where people are at.
And it feels a little bit magical because it's, instead of people moving piles of paper ontheir desk, their machines are doing it.
(05:03):
But you and I know that's not really AI.
And that's really what they're seeing the most success because I think it's pretty easystill.
Do you think this opens up a larger question in my mind?
Do you think that the rise of AI and now the fact that this is being discussed at theboard and the CEO level in ways that previous technological advancements had not?
(05:27):
Do you think that that is showcasing or exposing maybe is a better word the lack ofadvancement in RPA adoption and some of the other techniques that have been around for 20
years in the mid market?
Yeah, actually, I hadn't thought of it that way until you just asked it.
But I think that's a great point.
And the fact that everybody's bored.
(05:49):
I 98 % of the CEOs say this is critical.
the board, this is there's board pressure.
Certainly even just individual pressure of, I know this could work, but that is forcingpeople, forcing executives to um take action.
And probably the easy action is I'm going to go buy some point off the shelf solution.
(06:13):
I'm going to buy something to sort of stop the immediate pressure that's being put on me.
And so, yeah, it's probably a forcing function for
some of these technologies that you and I know have worked for many, years, just becausethey're easy and you can have quick success with some of them.
(06:35):
Well, it's not the long-term strategy.
Yeah, we're clearly seeing that in the results.
People are saying this isn't a strategy, hey, checked a box at least on the, m in terms ofat a board level getting, um I now have an answer to what I've.
you know, to the question, what are you doing in AI?
(06:56):
Well, I've got some process automation happening.
Yeah, that's slightly terrifying that that's the answer uh from a lot of folks in that.
Well, you know, we're just going to go buy something and we're going to throw it in thereand then we're going to say box checked.
uh That sounds to me like a future zombie company.
(07:17):
Yeah.
Well, and if you, let's just play this out.
Um, so if I buy a tactical AI solution, meaning it's an off the shelf solution that, um,let's say my CRM or my ERP vendor provides by definition, everybody else has access to
that exact same solution.
(07:39):
So, um, what am I actually achieving?
So let's, let's take like a sales AI agent.
So now it can, it's going to sell better, faster, you know, send more emails, what haveyou.
Now my competitor bought the same thing.
So now they're both competing with each other.
So I'm burning a bunch of compute and time and effort, but I essentially have the exactsame solution.
(08:04):
So, um,
Not only do I have this problem where I have siloed data because I have a point solutionrunning and it's running a process, but my competitor is running the same thing.
So what did I accomplish there?
Really?
As you said, I probably created these like zombie AI agents that are off really making abunch of money for sort of traditional software vendors, but not really solving the
(08:30):
problem, which is do I have an AI brain that's forming in my business?
It's this is fascinating and I didn't anticipate our conversation going down this roadeven five minutes ago and we started but what you've just unlocked for me here.
Is yeah, I just concluded an interview uh coming out probably within a few weeks of uh ofwhen your episode publishes with pit Bingaman down in Australia and he's got a company
(08:59):
that's focused on providing and unlocking creative potential in AI and one of hissuppositions is that if we're all using these same AI tools, we're all going to end up
with the same default crap quite frankly and.
I just heard the software equivalency of that with you with respect to organizationsbuying these tools from their vendors and I hadn't thought about it.
(09:23):
You know, I think about the content homogenization.
I hadn't thought about the process and the technical kind of homogenization, which youjust shown a light on.
Yeah, yeah, I, I, from where I sit the, and what, um, I advise companies to do is it'sreally why we're focused on the mid market as well.
(09:43):
What, if you look at the enterprise with a massive IT budget and spend they're buildingtowards a, their own AI, their own, I really call it a brain.
Mm-hmm.
really building their own contextual layer of the decisions and the cause effects thathappen in their business that are relevant to, or they believe are relevant to their
(10:06):
success.
That's really tough for the mid market to be able to do because of the budgets, thetalent, oh the pace at which this moves.
So you're focused on running your business.
You can't be focused on building all of the infrastructure that is required to stand thatup.
But it doesn't change the fundamental problem that AI is better when it has contextualawareness and you want to amplify your special sauce or your business differences,
(10:38):
especially in the mid-market.
That's why you're there.
That's why you love doing what you do is that you're different by definition.
You're not a homogenous product that has sort of achieved mass market appeal that way.
And um
It's about amplifying that.
And how do you get the contextual awareness into a brain that runs?
(11:06):
We call it uh connected intelligence.
But how do you start to build that?
such that you don't look like everyone else and your AI and your processes reflect who youreally are and not what the software vendors or what the um off the shelf solutions want
(11:27):
you to be.
Cause that's really what we did that em it's been called the software industrial complex.
They'll push a solution on you, sell it at scale, and then say everyone's process shouldlook like this, which is fine.
But then every business starts to look the same and you start to morph to what they wantas opposed to what should my business actually do and where do I get leverage as a
(11:53):
business?
Well, I think that falls right into something else that you and I talked about in terms ofthe execution gap, the distance between your strategic goals and the actual delivery
required to meet those goals.
And I'm wondering what your thoughts are about that as it pertains to this concept thatwe're unpacking here around the software industrial complex and some of these kind of
(12:19):
pre-formatted canned processes and systems that are out there, be they AI or otherwise.
Yeah, yeah, I'm curious as well your thoughts on what you're seeing just in talking to thegroup too.
So from where my perspective, the execution gap is I know as a CEO, I need to be usingthese tools.
I can see the power of AI.
(12:41):
I know what I want my business to achieve.
but I don't have the resources or the capability.
I can't invest ahead of the curve the way others can.
If I could just put 10 more, what historically bodies on this or just some more technologyhere, I could get ahead of that.
um And it's just growing because the larger companies have bigger budgets and can attractmore talent.
(13:09):
So.
How do we close that?
And so it's not about throwing more at it.
It's about being smarter and creating leverage.
So for me, it's very much this walk, get my data into a system, a brain, run.
How do I start to use that to create automations and then really scale it?
(13:31):
How do I unleash this?
system that I've trained this AI superpower and then close the gap because I don't neednow I've created a resource leverage engine, which is now at the core of my business.
Let's talk a little bit more about that idea of connected intelligence and getting thatdata in the right format, the right structure, getting it organized to populate that brain
(13:56):
because that's an area where I see an awful lot of mid-market organizations.
I don't want to say fall down, but it just hasn't been a priority.
And we're generating mountains and mountains of data, even, you know, smallerorganizations do this, yet it's not organized very well.
So what are you finding as you walk into companies uh in terms of their data hygiene andtheir data organization and how well or not well that's preparing them for this next level
(14:25):
of connected intelligence?
Yeah.
The default answer I get, is unfortunate, the way to clean up data is an ERP upgrade,which is terrifying, right?
We both know the cost overruns, the...
um
(14:47):
Effort it's always too long.
It's I've never met anyone who wants to do another ERP upgrade and so if that's the solveWhat you're basically saying is okay.
I'm gonna go put make my business look like Whatever the ERP wants it to look like and I'mgonna In essence take all the old mess and clean it and put it into there and move forward
(15:16):
The way I advise people is doesn't matter.
Absolutely doesn't matter.
I don't, I think the role of the ERP as a system that stores source data and puts a webinterface on it, it'll get, it'll move back to that sort of relegation.
be relegated back to that as opposed to this center of the organization tool.
(15:41):
um And so,
Where I advise people is it is actually really easy to get data now.
So the state of the hygiene is poor, but just that's what it is.
There's on average in the mid market studies will tell you that you have 50 differentapplications running.
(16:02):
So 50 silos of data costing roughly $2,000 per person per year.
So you can quickly do the math on what it's costing you to have all these silos andpockets of data, let alone the
hygiene that you asked about.
The great news is it's really easy now to pull data out of all of these systems and use AIto pull it in and make sense of it.
(16:31):
So a brain is actually really easy to create.
So what do we want to take orders, customers, cause effect, things that matter in thebusiness.
I don't need all of Slack or all of my email.
I want key decisions and key things.
And I can pull that into a semantic or contextually aware layer that I can start tointeract with.
(16:58):
there was this push where everybody had to have a data lake.
And it's still happening, data warehouse, lake house, whatever name the vendor's pushing.
To me, it's completely irrelevant to do that.
The tools we now have access to with AI make it that easy to pull your data in.
(17:23):
And if you look at what a dashboard signified, it's basically, I know what problem I'mlooking for.
I'm going to build a report that looks for it.
I'm going to make it red when it happens.
And then when I see it, I'm going to have to click three times in to sort of diagnose it.
So if you know all that ahead of time,
Why wouldn't you just ask an AI to look for that?
(17:46):
So seek out that data and then send you an email or a text or something.
So we've, we, I don't think dashboards even work.
look backwards.
if the implementation plan is, I'm going to put all my data.
I'm going to upgrade an ERP.
Then I'm going to put a business intelligence layer on top of that.
Cause the ERP doesn't do that very well.
(18:08):
And then I'm going to look backwards at data that I already know is a problem.
you problems I've already solved.
just going to look for them.
Now I'm two and a half years into something that, you know, we all agree is going tochange in the next three or six months.
So we've, we've created a concept in our, our, um, Bayo platform called a flash board.
(18:29):
So it's real time.
It's dynamic.
It's fluid information.
You asked the system, what should I know about today?
And it's just connected intelligence.
looks at your business.
It says, here's what
Inventory there's an issue.
see an inventory.
Here's a visual.
Here's what I think we should do about it um And that's just by wiring data and let alonesort of automations and agents all working together but the I think people uh companies I
(18:57):
talked to Have been Scared away from or or the salespeople.
I'll just say maybe put a pause
There's some great salespeople in the business, uh, intelligence world and ERP world whomake that the solve as opposed to let's just go at these sources.
(19:19):
There might be 50 pull the raw data in and start to look at it.
So let's uh let's make this real and uh correct me if I've got the company wrong, but Ibelieve you worked with Mugsy direct to consumer.
OK, OK, good.
I got that right.
So I I think what you're talking about in some ways, maybe this would be a good uh exampleto make this more concrete.
(19:46):
So Mugsy direct to consumer fashion brand.
uh I think you worked with them on inventory demand planning.
What did the problem look like for them before AI?
And then what actually changed?
And what was the measurable outcome?
So maybe let's use this as almost a little mini example or case study here.
Yeah.
So in talking with our CEO, had an initial, his initial thesis was AI, not too dissimilarthan we started the conversation.
(20:20):
AI is important, but I don't, I'm not really sure how to begin this journey.
So we take them through a walk, run scale.
Walk is all right.
What are the most, what are your 50 systems?
They had 14.
You could think Shopify, Amazon, Meta, Google, email, automation, uh inventory.
So they had 14 systems and something as simple as what was sales yesterday.
(20:46):
If you're not talking to all those systems, you're not going to get the same answer.
I sold something in Amazon.
I sold something in Shopify.
I sold something in my retail location.
And those are multiple different systems.
I don't need a business warehouse.
They didn't have a business warehouse.
We connect individually to those systems.
(21:10):
genius of AI is that it's been trained.
It can pull data from thousands of systems without any code.
So it's easy to connect to those.
And what are you fundamentally pulling in, in order?
Inventory, a customer.
you right there because I'm sorry to interrupt your flow, but what you just said remindsme of an objection.
I see red flags getting thrown all the time by CEOs based on the type of statement thatyou just made where AI doesn't need to know or it pulls these thousands and everybody
(21:39):
always says some version to me of, how can I trust that?
Yeah, yeah, exactly.
Well, um it's a great question.
how does it, um there's a couple of layers to in that.
um I would not trust my critical business data with a um AI platform or a sort of an openAI, you know, off the shelf AI platform.
(22:12):
I just wouldn't put my data in there because they're gonna learn, I don't suggest theylearn on your specific data, but they learn how your business runs.
So maybe less the price, but more importantly, how do you sell, you know, in this case, Dto C genes?
Now it doesn't mean they're gonna build a business from that, but as they watch thepatterns, the AI gets smarter and smarter.
(22:37):
So I personally don't allow the AI to do that.
What we do do though is we don't write code anymore.
So it used to be, you would have a programmer.
need to connect to HubSpot to pull my customer data.
And then I need to pull connect the net suite to pull my sales data, whatever it is.
(22:59):
Because that those APIs are available and open AI knows how to write the connection to goback and forth.
So I don't need massive teams.
programmers, it's not a three month, Hey, connect net suite and other three months connecthub spot.
So that's easy.
Um, so that, that is no longer a skill.
(23:21):
That's a commodity that, that we can, I don't need to pay developers for, but the datathat comes through that connection, I think that's really important.
That's your data.
That's private.
That should live private.
That's, that's part of our, the way we solve it for the mid market is that all that data,all of that polling of that data.
runs in a private SOC 2 data center where you're not sharing that.
(23:45):
So I think it's a really good point that the actual code that gets written or the way topull the data AI is doing, but the actual pulling of the data is private and secure.
And so it shouldn't be a concern, but I certainly wouldn't just go into ChatGPT or Geminiand say, here's my password, go pull all my HubSpot data, go pull all my sales data.
(24:09):
Here's my pricing strategy because what would you do?
You know, ask the New York times how this goes, right?
Like what would you do?
You are training on your IP and that's just the way AI needs to work is that it needsdata.
It's very data hungry to, to, um, to get smarter.
Alright, I appreciate you taking that detour with me.
(24:30):
I know I interrupted your flow.
uh I had asked you about outcomes and what actually changed as a case study with Muggsy.
So thank you for that because that again, that is a question that comes up all the time.
So kind of back to you on the journey with Muggsy.
Yeah.
yeah, I'm glad you asked that.
I think it's critical that this runs, that people understand the um potential there for umprivacy concerns.
(24:55):
um Yeah, so as we thought about the Muggsy example, so step one is let's connect ourbusiness and end up with a, I don't like this term, but
This is the way people often think about it is I've chat GPT for my business.
Now I have this AI.
(25:17):
can ask any question.
Hey, what was sales yesterday?
And not only does it pull the data from each of the different sources, it knows how Icalculate sales.
So just something fundamental is everyone does it a little bit different.
And, you know, it's funny, a lot of these like.
Harking back to when they started the business, there was some quirk and you know, that'sjust permeate.
(25:40):
throughout the business.
um So that is step one.
And a lot of times what happens in that step one connected intelligence is we take thespreadsheet that every CEO has that secretly runs the business.
And there's data from different sources.
I've got oh sales, I've got marketing, Roat, those different things.
(26:02):
And that now becomes this flash board where I can ask, hey, what's going on with thebusiness?
I don't need to spend.
an hour each morning compiling the Google sheet or the spreadsheet with all the differentdata sources.
So now I have, I can make decisions faster.
I'm confident.
I've built trust that I'm actually looking at the right information.
So for Muggsy, that was step one is before I'm going to make any decision, I got to justfundamentally trust that this, information I'm getting and the, um, it has access to
(26:38):
my data and the way I think.
So that's where the brain really is formed.
And while it's not hard to get the data in, that's probably a, you know, if I'm honest, a30 step process or 30 day process where you're just sort of having a back and forth and
sort of building trust at an executive level that you're getting the answers and theinformation that you would expect from a
(27:07):
senior data person.
And I think in the mid market, what happens is that people are siloed.
It's actually in the enterprise as well.
So the benefit here too, is that you get the same answer every single time, because thesedon't live in silos the same way.
When you ask marketing, what was sales, you get a different answer.
(27:28):
um Then when I ask the COO, I'm going to get a different answer.
And it's just skewed.
It was actually funny.
Our entire company, they live in AI.
We use it.
It's just part of who we are.
If we don't embrace it, then how can we expect our customers?
(27:49):
Even when my head of sales built a plan for me and my head of marketing built a plan forme, when I asked the plan sort of came back in terms of like we're in the funnel and the
percentages where the things move through the funnel.
Even building with the same AI, I got a marketing skewed answer for marketing, i.e.
(28:11):
more deals fell out in the middle of the funnel where sales takes over, then sales lesscomes in the top of the funnel and they're more effective.
So just how they were working with the AI in their own way skewed what the AI gave it andskewed their version of the plan that ostensibly came from ChatGPT or Gemini or whatever.
(28:31):
So
It's really important to break that down and get a consistent view of your business that'sunbiased.
And that's really what happens with this connected intelligence phase.
So that problem is solved.
Now all the while we had a use case in mind, which was, hey, we have too much inventory.
(28:53):
The easy way to plan is to say, here was sales last year.
I want to beat that by 30%.
So let's just take
and multiply every number, everything we bought last year times 1.3.
And now that's going to be our way to plan inventory.
And you can put some logic around, you know, we know that um in the fall around postThanksgiving to kind of the end of the year, we have more sales.
(29:21):
So we'll do a little bit more then.
But that's sort of the planning process.
And um as you start to m
go, what is the impact of buying that way?
Now I start to have millions of dollars in inventory that um wasn't very strategicallyplanned.
(29:43):
What could I do instead?
If I had more cash sitting around, I could put it in marketing.
So if I can be smarter about inventory, particularly in retail, I can make massive impactson the business because there's an obvious.
correlation between the amount of money I spend in marketing and then when it sells.
So the more cash I can free up, we have a massive impact.
(30:07):
Particularly as you think about sort of the times we're in now.
uh
you know, tariffs and inflation, the value of getting this right becomes more moreimportant.
So that was, that was the goal.
we, didn't start saying, Hey, we just want this brain that's smart.
want to start to be able to solve problems like this.
(30:28):
And that's sort of where, as we get to this run stage and the way um we're able to startgetting there is let's look at, let's let the model start to predict inventory.
instead of this top down,
sort of goal-based approach.
Let's look at the history of the business and then ask a machine to predict how inventoryshould look.
(30:55):
So the result for them was, well, hey, we've freed up several million dollars in cash.
As we back test it, it's about 95 % accurate to what actually happens, what it thinks willhappen to what actually happens.
And so it frees up this
sort of 70 % sort of gut check accurate to this very close to reality.
(31:22):
And just enables different business decisions to be made.
that's what he said.
Yeah.
Yeah.
that's huge.
have come across, and I don't want to make it sound like an overblown number, but I'vecome across numerous businesses over the years, profitable businesses that went out of
business in the mid-market space because of cash management.
(31:42):
And what you're saying here is, I know we're talking about a retail example, but anybodythat's got inventory, if you can free up several million dollars of extra cash uh by...
using the types of systems and putting the type of uh business intelligence in place thatyou're talking about, man, that's huge.
It's amazing.
It's amazing.
And it's just sitting there.
(32:04):
And it's really not that hard.
We connect in the data.
Now what do I have?
I have my order history.
I have my marketing history.
I have the different things that have happened.
And I have time.
So I have a history with that.
And now we ask the machine to run through that in a way only a machine can.
It can run massive compute in ways that our minds can't.
(32:27):
And then it says, here's what I think is going to happen.
And it gets smarter and smarter as you can add more more factors.
But it turns out in less than a couple of months, you can have a very accurate predictionof what's going to happen.
Then you could start to add in more and more more things.
Hey, this what happens when I have an influencer wear a pair of jeans?
(32:50):
Let's introduce that to the model.
We don't need to start with that.
We can still be 95 plus percent accurate, free up millions of dollars.
But
then you can really start to do some exciting things.
think sometimes people kill these projects with...
over complication or perfects the enemy of good enough kind of thinking, you can getreally accurate and much better than you were doing.
(33:12):
Just think of the difference we said there.
Yeah, and let's talk about that here briefly because I have seen a lot of AI pilots thatgo nowhere.
And oh, and by the way, I wrote something down here before I get to that.
uh Just in case anybody uh kind of drove by the point that you made, I wrote this quotedown.
You said the spreadsheet that every CEO has that actually runs the business.
(33:35):
And when you said that, it just rang so true to me because I walked into so manyorganizations.
They got these complicated dashboards.
They got metrics.
And then here's this Excel sheet of this Google Doc that the CEO is actually using to runthe business.
So you're definitely on point with that.
um
for most people.
like, let's just get that into a AI system you can ask questions of.
(33:58):
Exactly.
All right.
So uh with that little detour aside, let's talk about AI pilots that go nowhere becauseyou kind of referred to that a moment ago as we were concluding some remarks there.
I think you've got the idea about having a right to left or an outcomes based approach.
um But let's talk about AI pilots that go nowhere.
(34:19):
Let's talk about what the right type of approach is um for
Folks out there, maybe they've got AI experiments that have failed or pilots that wentnowhere.
What can you speak to about that here?
Because I'm seeing more and more of that.
Yeah, I do like the um right to left approach.
Start with an outcome that matters and then work it back.
(34:44):
And so that could be the inventory example we gave.
It could be the Excel spreadsheet.
that, let's have a system I can interact with that understands that, key pieces ofinformation or data sources.
um Why I think they fail,
I'll give two reasons.
(35:04):
One, there's not CEO buying.
Mm-hmm.
If you are checking the box, like we said before, and the organization knows that I thinkit's going to end the way we all were most of these pilots end, which is you're going to,
(35:25):
it's going to die in sort of organizational inertia because this is different.
This is not implementing a new, report or an ERP system.
AI fundamentally is a
change to the way you do business if you believe it at a strategic level.
It's a thought partner.
(35:45):
It's a strategic resource that's always on with the ability to run massive compute andsimulate a lot of different examples and then start to give you options, which is just
different to the way businesses run.
We come in, we work nine to five, we
(36:05):
have silos of information, we operate on that and sort of, so if the CEO doesn't buy in, Ithink that's where these will die.
A lot of times you have people sort of in the middle of the company who really don't wantthese to be successful because it's gonna change what they do.
(36:26):
I don't mean they're gonna be out of a job, it's just gonna change the way they functionby the nature of it, which.
You could view as very powerful as an individual or you could view it as very scary and Iunderstand that.
So that's one reason.
The other reason is what we said before.
If you're tactical about this.
(36:47):
Mm-hmm.
You're going to end up looking like everyone else and you're going to end up with, Ithink, mediocre C plus results.
AI is trained on a crowd sourced set of data.
So if you buy that off the shelf product, you're going to get a crowd sourced answer toyour business, which is fine.
(37:09):
AI is smart.
It knows how to build a sales plan and a marketing plan, but it doesn't know how to buildyour sales or marketing plan.
If you just look like everyone else, then what do you have?
And so then the pilot doesn't feel very good.
You can see the potential, but when you actually say, well, people are better because theyknow the nuance, they know how to get it from C plus or B minus to A.
(37:33):
We did a pilot.
was, my mind was blown in this one um for a customer that was doing.
It was a drone company that was working on helping farmers regenerative farming and AIcould run a farm just with satellite images to I would say like a C plus B minus level,
(37:58):
like an individual farm.
actually told a farmer they had pumpkins growing on the, that they didn't realize the sunhad like planted these pumpkins, this little patch of pumpkins.
And it knew it just from the satellite images.
I figured that out.
It still wasn't going to beat the farmer individually.
But if that farmer would then say, I've got this really smart thing that if I, if I workwith it a little bit, wow, the power I could, I could have, but that's where the pilot
(38:25):
sort of moves from.
Um, it's really pretty good, but how do I make it great?
There's work there.
And, and if you just buy something off the shelf and expect it to do better than all thepeople you have and
Importantly, it doesn't have all the history of your business and understanding and thecontext of your business.
If you don't give it that it's never gonna move off the, you know, pilot phase.
(38:50):
And that's where I think people expect it's easier to just buy something than to say thisis strategic.
I'm gonna
wire in the intelligence as part of the phase one, because you don't get that sexyoutcome.
So if you start with the outcome and say, this is what I want to achieve, but I'm going toalso lay the foundation.
(39:11):
What we'll actually do with our customers, we don't let them move on and don't let thempay us more money until they move through the phases, the walk, the run phase, and the
scale phase.
So in walk, we say, we're not going to write data to a lot of other systems.
We're going to
help you build trust and get the right answers.
(39:31):
Then as you run, we're going to start to uh automatically write information from system Ato system B.
And then the final stages, you're starting to let the agents make decisions and reason andstart to do things on your behalf.
But you have to build trust to do that.
So if you come in and just say, I'm going to buy a solution and I'm going have all theseagents running and my business is going to look better in
(39:59):
overnight, it's just not going to work.
It's going to look like every other business.
It's going to look like the software company and those agents are optimized to sell youmore licenses.
It's kind of the end story, right?
It's going to use more compute and they're going to optimize for that silo, but notunderstand your business.
(40:21):
So everything that you just walked through, I would imagine that you went through withyour own company as well.
I believe you had said to me you went from a target of a million dollars in revenue peremployee to a target of about 10 million per employee with AI.
Is that right?
Okay.
All right.
Wanted to make sure I had that right.
So as you went through that, right?
(40:45):
No, no, no.
Yeah, I recognize that.
And that's the interesting part about so much of the AI stuff is many of it is or much ofit is still aspirational in nature, um yet it's it's aspirational um at a level that is
obtainable as I guess is what I would say uh from my observations with folks.
(41:07):
Yeah.
Well said.
Well said.
Well, and I love your philosophy on this, the learning philosophy.
This is so new.
We're all learning and I make mistakes all the time.
And, but I think the biggest you can make mistake you can make an AI is not getting in thegame and learning.
Wanted to start building this brain.
(41:29):
You take, know, my kids now are one just turned seven.
So I'm, four and seven.
Imagine if you didn't start teaching them anything until they're four, you've just pushedthe 18 years out.
Now they're 22, you you've got to start teaching these things.
You've got to start making mistakes.
There's no, um, there's no other way to start working with this because the system cravesyour data to be, uh, able to best help you as a business.
(41:59):
I love that example and kind uh of drawing your kids in and the idea of education andlearning delays there.
And the first thing that popped in my mind when you were talking about that would be.
As if I didn't start speaking or teaching someone English and I'm just going to throw outsome some random dates here and it's 1752 right?
(42:21):
And with the pace of AI, if I wait a couple years to dip my toe in when I'm ready to teachthem English, they're going to be speaking in 2026 vernacular, which sounds nothing like
300 years ago.
Yeah, isn't that a hundred great point great point and you know what I hear to follow thaton is I'm just gonna wait because the AI is gonna get better.
(42:43):
All right.
Well now the AI the gen AI tool got better, but it has zero context about your business.
So again, I think a lot of people
And I don't expect they would understand it.
I've naturally had to learn a lot about how this works.
But if you have a generative AI model, which is amazing at predicting words and is amazingat human language, I'm mesmerized to the conversations and the back and forth and the way
(43:12):
it um uses our language with us.
But um all that said, it has zero awareness unless you give it that.
As you're asking questions and having a back and forth that has zero transactional historywith you.
So to your point, how do you build that up?
(43:34):
And then as the Oracle that you're asking is that keeps getting smarter.
You want to keep a constant set of learning such that you're not getting a differentanswer every single day because that that field goal or that
ball is constantly moving.
And but you want to have a way of getting that same answer.
(43:56):
Sales yesterday can't change because the Oracle changed, right?
So you have to have a way of teaching it and then using knowing that it's going to getsmarter, still get the same answer.
So it's a really interesting paradigm shift there too.
It absolutely is.
And I think it also opens up this line of discussion opens up the larger question ofstrategy in an organization and the shelf life for that, because I was just a guest uh on
(44:27):
a pod the other day.
I don't do a lot of those, but they're wanting to ask me about strategy and timelines.
So we were talking about some of this.
And one of my views is that most strategic plans.
um
If it's an 18 month shelf life uh and that that assumes it's dynamic, that's a long shelflife anymore.
There are no 10 year, five year, three year plans anymore.
(44:47):
And I think the point that I'm driving to here with a question based on what you said isthat if you wait to get in the game with AI, you're not only sacrificing momentum, you're
sacrificing agility because the speed at which uh changes are coming down the pike now issuch that there are decisions
that are happening outside of your purview based on AI within your company if you'veimplemented it, right?
(45:13):
So if you are not implementing AI, if you're not using these tools, you're sacrificingspeed and you're increasing the level of complexity regarding decisions that your mid and
senior line uh supervisors and managers need to make in my view.
And that's gonna further drag you down.
and compound your uh or hamstring your ability to be agile and to move quicker based onmarket forces that are changing extremely rapidly.
(45:42):
So I'm not sure I got myself to a question with that, but that was kind of top of mindbased on what you were talking about a moment ago.
Yeah, I think you described better than I did the execution gap.
You just keep falling behind.
Other people are getting ahead faster and it's compounding because it's smarter and fasterand they can make decisions faster and you're still doing it the old way.
(46:06):
If you go in wanting perfection, you won't achieve it without some failure.
So you should go in saying I'm going to start now and I'm going to fail along the way, butit's going to close the gap because others are doing it.
And if I'm not, look, 98 % of CEOs said, I need this.
(46:26):
7 % have a strategy.
do you, could plan for 18 months and build a strategy or you could get in and do, and thestrategy will form because this is, this is new.
There is nobody, there's not a
silver bullet strategy.
there was this, we wouldn't be seeing that execution gap forming.
(46:46):
Right.
What's that quote from Patton?
always butcher it, but it's something along the lines of an imperfect plan violentlyexecuted today is far superior to a perfect plan executed, you know, far down the road.
Right.
So same idea.
So.
Go ahead.
(47:07):
I was just going to ask you, since we're talking about this change and we were talking ata strategic level, we touched on the human component earlier and I know you referenced
we're not here to automate you out of a job.
I think 26 is the year we're going to start to see an awful lot of pushback.
We're already seeing it.
And I've had some conversations with other folks that specialize in AI training fororganizations.
(47:31):
They focus a lot on mindset shifts.
They focus a lot on
uh demystifying the technology and reducing the level of fear.
I would imagine that um all of these things come up when you're working with your clientsthough.
it's one thing to talk about this in the boardroom and to set the course of the company.
It's another when you've got to handle those actual conversations with the workforce andthe line level team members.
(47:56):
So I'm interested, what's been your experience so far with those types of conversations?
with the folks that are on the front lines doing the work.
Yeah, I suppose there's two, there's a carrot and stick approach.
You could see the CEOs who will take the approach of you will use this tool.
(48:20):
You will report back to me if you didn't, because I'll see that you have zero chat historyand we know how that ends.
So that can work.
And then there's more of the carrot.
How do I lead someone?
How do I, I, um, show them or, um, start to pull them towards how this can make their lifeeasier, better, um, help them use it to achieve the outcomes they want or the business
(48:50):
wants.
think largely the people I talked to don't want to do the repetitive.
manual non-creative tasks So I think there's a natural inclination to want to work withthis stuff Once you get over the initial of of wow, it just did that um I Think if you can
(49:20):
get people there they want to work with it that that's generally my my belief I will say
It's changing for me, what I value in the workforce.
resumes are less important to me than they ever were.
Uh, raw curiosity and creativity, um, given that the answers already exist in these tools,the sort of book answers, but it's, do you do with that?
(49:53):
And how creative, how can you 10 X yourself with these tools?
Mm-hmm.
personally value that in our business.
Now granted, we're in the business of bringing AI to other people.
So we're at the tip of the spear there.
But I do think that's really important as people are thinking about careers and careerchanges and as we um work with interns.
(50:19):
We do a lot of interns just to try to pull them through and teach them how this works inum business.
But I think that skill set
of being told what to do versus I'm going to be creative and I've got this set of tools towork with.
That's a shift.
And so that I don't know that there's a.
(50:39):
em
Certain people are gonna love that, other people aren't, and there's different points inthe career where you may say, I don't wanna do that anymore, I don't wanna shift, I just
wanna kinda see this through, I've got three more years left.
And so that piece of the workforce, I think, gonna, that'll be interesting how that playsout.
(50:59):
Mm-hmm.
I couldn't agree more.
Our time is uh starting to get close to an end here.
And so I want to shift gears for maybe our last subject uh point of conversation and askyou about values, because this is something that I've been tracking very closely since the
(51:20):
fourth quarter of 2025.
uh Anthropic, the company behind Claude, came out with the sole document, as they calledit in late 25 and in early 26 in January, they came out and published their 77 page, what
they call the
Constitution.
This is an area that fascinates me.
um
And it fascinates me precisely because of what we talked about earlier with respect.
(51:44):
And there's a parallel here to what you talked about when companies buy ERPs, they buylarge software systems.
You're asking for the same pre-formatted processes that everybody else is using.
And I think there's a parallel here with the values and the judgment calls that AI ismaking with respect to the values that are inculcated into those different models.
(52:04):
uh In my view, think Anthropic is doing a very good job of
trying to grapple with this question and being relatively transparent.
But I am curious whether this conversation comes up much with your clients in terms oforganizational values.
How do we infuse those into the tools that we're using?
How do we make sure that uh what is important to us and
(52:28):
what we value as an organization, mission, vision, values is then seeded into theautomated AI systems and tools that we're using.
I don't expect that you've got a silver bullet answer, but I'm interested in yourexperience on this, because I think this is an undeveloped conversation that is super
critical that not enough people are having yet.
Yeah, I love that you're you've always been I think thought leading on where does ethicsvalue where does that meet AI because to your point we're trusting the answers that come
(53:01):
from these oracles and whoever's trained there's a bias inherently in how it's trained andfor example we won't we don't use any of the Chinese models I don't
Hmm.
Yep.
what's in there.
um And that's just a stake we've put in the ground.
(53:24):
um Because of the reasons we said, the model is trained by people and processes and ondata.
I don't know.
And so I don't trust that the answers are going to be based in fact.
Now.
that it doesn't change that you can have biases by the programming team and the data setsthat we're training these on, even in the, you know, Anthropic or open AI or Gemini or any
(53:50):
of those.
So um the question is, is real.
And that's where the most important thing you can do as a business.
And this is different than a consumer level.
As a, when I'm interacting with a
Gemini or AI on my desktop, I have my own values to gut check it and I have my own versionof the fact that I encourage everybody not to take what it says at face value.
(54:22):
Their aim is their aim, their design to make
you feel like you're right and kind of pull you through a conversation and please you.
So I encourage people individually to challenge them and ask, is that real?
Do I believe that?
I mean, there's still prediction engines, so they'll get the wrong answer.
And if you go with it, it's just going to perpetuate that.
(54:46):
So what does a business do?
Because me as an individual, I can do that.
I have that choice.
But a business can't.
put those guardrails in.
So if you don't have core values and vision and mission and all that stuff as guardrails,your agent and employees who are interacting with this are going to biased answers or
(55:09):
they're going to get their own answers, which don't necessarily reflect the values of thecompany.
So it all goes back to building a corporate, a corpus of information, a brain.
That is absolutely critical that sits in front and independent of any of these modelswhere that lives and those guardrails live and they're indisputable.
(55:34):
You cannot deviate from those.
That way we don't have, are stories you can flip open the news.
just saw one, Air Canada was held liable by the Supreme Court for their agents.
They said, well, we didn't write them.
Well, sorry.
They're acting on your behalf.
They gave $2,000 refunds.
That's you're liable for that.
(55:55):
So how do you build?
And it's critical.
What is the brain that you're building that guards against that from one of these modelsthat we're all using?
Eventually, I think everyone will start to build their own smaller AI models that we usethat have that inherently built in.
But right now, what is it?
(56:17):
The cost of doing that is just
It's just not feasible for anybody.
So we all use these massive models.
They're great, but we get their biases.
So I love that you've brought this up because this is where it's going is how do we knowwhat's coming?
How do we know why we got the answer?
What data was was put in there?
(56:39):
And so that's where as a business leader, I think you need a business AI, which isseparate than.
what we use like a co-pilot on your desktop or something that helps me build a betterPowerPoint or write emails faster.
I need a business rules that it follows.
Well said.
All right, let me bring us the full circle background or we started here with a partingshot for you.
(57:05):
We started talking about the the one in the 12.
Let's talk about the other 11 uh as a final thought here for our listeners.
So if I'm uh 11 of the 12 and I am not doing what I probably should be doing uh in AI landright now, and I'm a CEO listening to this.
What's something you'd recommend if I'm in that 11 category that I should do in the next30 days to start to move the needle in my organization?
(57:32):
Yep.
Well, one, take ownership yourself.
head of sales, of course, this comes from the sales guys said what either you think thisis strategic or not.
If you don't think it is, so you're in the 2 % even, I don't think this matters.
Put your head in the sand.
Don't let's don't have the conversation.
(57:53):
If you think it's strategic and you believe this has the potential to impact your businessin a massive way.
When, where else would you have that happen and not own that as the CEO?
So make that decision that you're going to own it and not punt it to it.
Cause this is not an IT problem.
(58:14):
The IT parts easy.
So the next 30 days, I would say, do I care about this?
Is it strategic?
Will it disrupt my business?
And if so, I'm going to own it.
At that point, you'd say,
Do I want to own it by buying a bunch of piecemeal tactical solutions or am I going tobuild or buy a solution that is a strategic platform, a brain, an AI system that runs in
(58:42):
my company and learns what my company operates like, how I want it to operate, andeverybody can use it.
I would get to that in the next 30 days.
So where that ends for the mid-market is, do I build one?
Because you could build what we've done.
You absolutely could.
Or do I buy one?
We're the only system out there that does what we do.
(59:06):
We call it a BAO.
There'll be others that do it.
But if you get to, want this to be a strategic component of my organization, I'm going toown it as a CEO.
Now I got to go solve that problem.
That's what I would do in the next 30 days is get.
Reason your way through those answers and then start making decisions around that whichwill lead you to all right now I have a strategic plan and it won't be I'm gonna start
(59:37):
buying things It'll be I'm gonna start with a plan in mind and we're gonna ebb and flow onthis journey because it will It will change as technologies emerge
Well said and sage parting advice from Chris Hap, CEO of Virtuous AI.
Chris, I appreciated our time together today.
If someone wants to learn more about Virtuous AI and your offerings, if they want to getin touch with you, what is the best way for them to do that?
(01:00:04):
Yep, it's VirtuousAI.com.
And if they want to email me, I'm chapp, C-H-A-P-P, at VirtuousAI.com.
Very good.
Thank you so much.
This was a great conversation.
I appreciated your perspective and your time today.
Likewise.
And for our listeners out there, thank you for once again listening to AI for the C-suite.
(01:00:27):
If this episode was useful, subscribe wherever you get your podcasts, follow us onLinkedIn, and check out AIfortheC-suite.com.
Until next time, keep your algorithms running, your leadership evolving, and your AI incheck.
Take care, everybody.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Betrayal Weekly
Betrayal Weekly is back for a new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. If you would like to share your story, you can reach out to the Betrayal Team by emailing them at betrayalpod@gmail.com and follow us on Instagram at @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.