June 24, 2025 • 56 mins
On this episode of The Founder's Sandbox, Brenda speaks with David Hirschfeld, owner of 18 year old business Tekyz, that boasts a hyperexceptional development team building high “ticket” products in the B2B space. They speak about ways in which AI is a gamechanger, how Tekyz backs their work for clients with relentless pursuit of quality, and how Tekyz practices ruthless compassion,to protect the company and enable it to grow
Having collaborated with over 90 startups, he developed the Launch 1st Method—a systematic approach that minimizes risks and accelerates software company success with reduced reliance on investor funding, after observing that many companies launch a product first and then fail at a later stage – With Tekyz approach of Launch 1st exceptional founders are in love with the problem not the product.
David's expertise bridges cutting-edge AI technologies, workflow optimization, and startup ecosystem dynamics. When not transforming business strategies, he enjoys woodworking, golfing, and drawing leadership insights from his experience raising four successful sons.
You can find out more about David and Tekyz at:
https://sites.google.com/tekyz.com/david-hirschfeld?usp=sharing
https://tekyz.podbean.com/ - Scaling Smarter Episodes.
www.scalingsmarter.net - Schedule an interview
https://www.linkedin.com/in/dhirschfeld/
https://www.linkedin.com/in/dhirschfeld/
https://www.facebook.com/dmhirschfeld
transcription:
00:04
Welcome back to the Founders Sandbox. I am Brenda McCabe, the host here on this monthly podcast, now in its third season. This podcast reaches entrepreneurs, business owners that are scaling.
00:31
professional service providers that provide services to these entrepreneurs, and corporate board directors who, like me, are building resilient, purpose-driven, and scalable businesses with great corporate governance. My guests to this podcast are business owners themselves, professional service providers, and corporate directors who, like me, want to use the power of the private company to build a better
01:01
world through storytelling with each of my guests in the sandbox. My goal is to provide a fun sandbox environment where we can equip one founder at a time to build a better world through great corporate governance. So today I'm absolutely delighted to have as my guest, David Hirschfeld. David is the owner and CEO of Techies, 17 or 18 year old business now that boasts
01:29
a hyper exceptional development team that are building high ticket products in the B2B space. Welcome David to the Founder Sandbox. Hi Brenda and thanks for having me. Great. So I'm delighted that we actually did a dry run in February. We've known each other for some time and AI, we're going to be touching on AI. And I think that the world of AI
01:58
particularly in software development, has changed significantly since we last spoke in February. So we're going to be getting into some, I think, novel concepts for the listeners of the Founder Sandbox. So I wanted to, you I always talk about how I like to work with growth stage companies that typically are bootstrapped and
02:26
It's only at a later stage do they seek institutional investment by building great corporate governance and reducing the reliance on investor
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:04):
Welcome back to the Founders Sandbox. I amBrenda McCabe, the host here on this monthly
podcast, now in its third season. This podcastreaches entrepreneurs, business owners that
(00:30):
are scaling. professional service providersthat provide services to these entrepreneurs,
and corporate board directors who, like me,are building resilient, purpose-driven, and
scalable businesses with great corporate governance.My guests to this podcast are business owners
themselves, professional service providers,and corporate directors who, like me, want
(00:54):
to use the power of the private company tobuild a better world through storytelling with
each of my guests in the sandbox. My goal isto provide a fun sandbox environment where
we can equip one founder at a time to builda better world through great corporate governance.
So today I'm absolutely delighted to have asmy guest, David Hirschfeld. David is the owner
(01:22):
and CEO of Techies, 17 or 18 year old businessnow that boasts a hyper exceptional development
team that are building high ticket productsin the B2B space. Welcome David to the Founder
Sandbox. Hi Brenda and thanks for having me.Great. So I'm delighted that we actually did
(01:45):
a dry run in February. We've known each otherfor some time and AI, we're going to be touching
on AI. And I think that the world of AI particularlyin software development, has changed significantly
since we last spoke in February. So we're goingto be getting into some, I think, novel concepts
(02:10):
for the listeners of the Founder Sandbox. SoI wanted to, you I always talk about how I
like to work with growth stage companies that typically are bootstrapped and It's only
at a later stage do they seek institutionalinvestment by building great corporate governance
(02:33):
and reducing the reliance on investor funding until such a time that they choose the right
type of investors that can help them scale.So when I found out what you do at Techies
with Launch First and the type of work youdo in B2B businesses, I absolutely wanted to
(02:53):
have you here on the founder sandbox. So let'sjump right in, right? I think I'm eager to
learn more about how to scale your bespoke developmentat Techies, right? To scale my own business?
(03:13):
Okay. So there's a lot of different aspectsto scaling my business and I bootstrapped for
the last 18 years. I've never taken any investment with techies. And I've done that very specifically
(03:33):
because it gives me a lot of freedom. I don'thave a reporting structure that I have to
worry about. That doesn't mean that I can belazy with my team. To grow my team, I have
a philosophy that I only hire people that aresmarter than I am. And the ones that are
(04:01):
in a position to hire, they can only hire peoplethat are smarter than them. And by really
sticking to this philosophy, even though sometimesit makes us grow a little slower than we would
like, it means that when we bring in people,those people contribute immediately and contribute
in a way that it's our job to get the impedimentsout of their way and to facilitate them so
(04:28):
that they can contribute and help us grow thecompany. So I call it the ball rolls uphill
here because my job is to support everybodythat is above me, which is everybody. And then
the people that I support directly, their jobis to support the people that are above them.
(04:51):
Because if we're hiring correctly, then peoplethat we bring in can contribute in the area
that we're bringing them in way more than theperson that's hiring them. Okay. Thank you
for that. So before you launched Techies, youhad a career in companies like, I believe,
(05:13):
Computer Associates, right? Texas Experimentsand TelaMotorola. There was a period of time
between your experience in these large corporationsbefore your launch tech is where you actually
had your own startup and you sold it in 2000,right? And I believe you also learned perhaps
(05:36):
with the second startup about how hard it isto find product market fit. Can you talk to
that for my listeners, please? I don't knowthat it's that hard to find product market
fit. It depends if that's your focus or not.If your focus is to nail down product market
fit, then it's not that hard to determine whetheryou can achieve that or not fairly quickly.
(06:04):
You can do that by selling your product topotential customers. That sounds strange.
Of course, we all want to sell our products,but What I'm suggesting is you start selling
your product before you have a product, beforeyou have a full product. And I don't mean
an MVP, but a design prototype. You go out tothe market and you start to sell it. If you
(06:29):
have product market fit and you've identifiedthe early adopter in your market and you know
that they have a very high need from a perceptionperspective and there's a big cost to the
problem that you're solving. then you can offerthem a big enough value upfront that they'll
buy your product early and you can prove thatthere's a market for your product and they'll
(06:52):
buy it in enough numbers that you can achievea measurable metric, which I kind of call
the golden ratio, which is three to one in termsof what is the lifetime value of a customer
versus what does it cost to acquire that customer?And you can get to that three to one ratio.
(07:13):
in a prelaunch sale model before you ever starteddeveloping your product as a way of proving
product market fit. Or you pivot quickly andcheaply because you're not having to rebuild
a product that you've built in the wrong way.Or you fail fast and cheap. And every entrepreneur's
(07:34):
first goal should be to fail fast and cheap.know that sounds backwards, but that should
be your goal is that you can fail fast and cheapor if you If you fail to fail fast and cheap,
that means you've found a path to revenue and product market fit. And now you know you have
a viable business. making the investment tobuild the product is a no brainer. And you
(07:59):
came upon this methodology, right? Yes. becauseyou did yourself when you had your first company,
you did not understand the funding part, right? Can you talk? a bit about your specific example
and then how that's informed now 17 years oftechies and over 90 projects with startups.
(08:23):
Okay. So my first company was Bootstrap. Okay. And that one was successful and we grew it
despite me, it was me and a partner. And despiteourselves, we grew it over eight years. where
he ended up with 800 customers in 22 countriesand sold it to a publicly traded firm out of
(08:47):
Toronto. That was in the product food, snackfood distribution business because that was
what our product was focused on. So I startedanother company about five years later, not
realizing the things that I did the first time.that made it so successful, which really
(09:13):
fit the launch first model to a large degree. But the second time I built a product that
would have been successful had I followed myfirst model, but I didn't. So I went the route
of building an MVP and getting customers ona free version of it, and then going out and
trying to raise money, which is the very classicapproach that the SaaS products take now.
(09:40):
And the problem is with that approach is thatyou end up digging a really deep hole in terms
of the investment that you make to build theproduct with enough functionality that you
can convince people it's worth putting an investmentin and you're not generating any revenue at
the time. And I should have just started sellingthe product and generating subscription revenue
(10:01):
right from the beginning. First of all, I wouldhave been able to raise money much more easily.
Secondly, I would have not needed to raisemoney as much if I'd focused on sales. The
problem with a lot of founders is they fallin love with their product. They believe that
people will buy it at enough numbers and thatinvestors will see the potential. they're afraid
(10:29):
of sales. I've fallen into this trap beforetoo. I've done it both ways. And I can tell
you selling early and staying focused on thecustomer and the problem are the way to be
successful. So founders who I find are consistentlysuccessful, they are focused on the problem,
(10:50):
they love the problem. The product is just thenatural conclusion to solving the problem,
not something to be in love with. They spendtheir time talking to customers about the problems.
So how does a potential customer find you andwork with you? Oh, they can find me at Techies
(11:11):
or they can find me at LaunchFirst, was spelledlaunch1st.com. And they can find me on LinkedIn.
And then to work with me, it's just give mea call, send me an email, we'll set up a Zoom.
I'll start to learn about what you're tryingto accomplish and what your requirements are.
(11:33):
And I'll typically spend quite a bit of timewith any potential clients. in one to usually
multiple calls or Zooms, learning and creatingestimates and doing a lot of work in advance
with the idea that there'll be a natural conclusionat the end of this that they'll wanna start
(11:54):
working with me in a paid fashion. So there'sa lot of value that my clients get from me
whether they end up contracting me or not. And how, again, back to, thank you for that
and that. how to contact you will be in theshow notes. But what types of sectors do you
work in? You know, in your introduction, Italk about high ticket B2B, right? who are
(12:21):
the, so what founder that's has some ideatoday? What would be their call to action
to find techies? And what would you, is it launchfirst before you go down? No, it's not necessarily.
It may be an existing company that is tryingto implement AI or implement workflow automation,
(12:46):
or they have a project and they don't have theIT team or capacity to handle it. We love
those types of projects. It might be an existingstartup that is struggling with their software
development team and they're not getting tothe end goal that they're expecting and the
(13:07):
product's buggy, it's taking too long, there'sconstant delays, they're way over budget and
they need to get this thing done. And I callthose recovery projects, they're probably
my favorite because people recognize very quickly the difference that we bring. and they really,
(13:34):
really appreciate us. As far as what sectors, business sectors, healthcare, law enforcement,
prop tech, real estate, finance, entertainment,I mean, we work in many, many different sectors
over the last 18 years. So regardless in B2B, B2B2C, not so much e-commerce unless
(14:01):
there's some complex workflow associated withyour particular e-commerce, but there's lots
of really good solutions for e-commerce that don't require developers to be involved.
But mobile, web, IoT, definitely everythingis AI now. Absolutely. And in fact, when we
(14:25):
last spoke, I'd like to say that you startedto drink your own Kool-Aid at Techies. you're
starting to actually use AI automation for internalfunctions as well as projects at Techies. So
can you walk my listeners through how you'reusing AI automation and what's the latest
(14:52):
with agentic AI? So let's do the first. Yeah, okay. So there are a bunch of questions there.
So let me start with that we're building productsinternally at Techies to help us with our
own workflows. These products though are applicable to almost any development company
(15:17):
or any company with a development team. Someof them are, and some of them are applicable
to companies that are, well, so one product is putting voice capability in front of project
management tool. and we use JIRA and JIRA isan incredibly technical tool for project managers
(15:37):
and development teams to use to their projects,requirements, their track bugs, all of that.
And so your relationship with what I call relationshipwith project management is very technical one.
If you're a client, some clients are willingto go through the learning curve so that they
(15:57):
can enter their own... bugs and feature requestsand things like that directly into JIRA. Most
don't. They want to send us emails, whichis fine, and just give us a list of what's
going on and the problems that they're findingor the things that they need for a future
version and the planning and the documentation,everything else. This is a real technical thing.
(16:20):
We're going to make it a very natural personalrelationship by adding voice in front of all
this so that you can be sharing your screenwith your little voice app and say, just found
a problem on the screen. And the voice appcan see the screen. It knows your project.
It knows your requirements. And it can identifyproblems on the screen that you may not have
(16:43):
even noticed. And it can also prevent you fromreporting bugs that have already been reported
and tell you when they're planned to be built. And all of this just with a verbal discussion
with the app. that basically knows your project. Kind of like talking to a project manager
(17:03):
in real time, but they don't have to write downnotes and they can instantly look up anything
about your project in terms of what's been reportedin terms of bugs or feature requests and update
them or create new ones for you or just reportthem to you and tell you when things are planned
to be built and released or. where they've alreadybeen released and maybe you need to clear your
(17:27):
cache so you can see the change, whatever. Yeah. So it be like an avatar, but it's trained
and it's specific to Jira in your case? Inthe first version, it's actually being built
architected so that we'll be able to add otherproject management tools to it besides Jira
in the future. to begin with, because we useJira, it's going to work directly with Jira
(17:52):
to start. And this, by the way, you askedabout agentic workflows, right? So we're
building an agentic workflow in this tool wherewe have more different agents that work together
to resolve these issues. so we have an agentthat reads and writes documentation to JIRA.
(18:16):
We have an agent that communicates with theuser and the user might be the programmer might
be a person in QA, it might be a client fora lot of different things. And we have an analyst
agent that when the person talks, the voiceagent says to the analyst agent, here's what
(18:37):
I understand. Here's the information I justgot. Go do your work and come back and get
me the answer. And it'll speak to the JIRA agentto get the information. It will also speak
directly to us. a vector database, which isa database where all the documentation from
that project is ingested into our own separateAI model so that the context of all the communication
(19:06):
is about their project and doesn't go off intoother directions. And then can get back.
So this is an agentic workflow. The idea ofagents is like everybody keeps talking about
agents. Not everybody is really clear on whatthat even means. Can you define that? an
(19:29):
agent is an AI model that you can interactwith that is focused on one specific area
of expertise. So if it's a travel agent, theword agent fits very well there, then their
expertise would be on everything related totravel and booking travel and looking up options
(19:52):
and comparing prices. And that would be anAI travel agent. So that's very different
from an AI project management agent, very differentfrom an AI financial analyst agent. So each
agent specializes in its own area of expertiseand may draw from specific repositories of
(20:19):
information that are specific to that particularagent's area of expertise. And they actually
look from the perspective of that type of person,if it was a person. So, and so they'll respond
in a way that is consistent with how somebodywho is a project manager would respond to you
(20:41):
when you're talking to them, asking you questionsabout your requirements, knows what information
it needs to be able to assess it properly, thingslike that. wouldn't be very good about travel
because that's not its area of expertise. Right. So is it common to have companies that are
(21:02):
creating with their own large language model,right? Or their workflow processes internally
to the company to create their own agent AI? Or is there a marketplace now where you can
say, want this type of agent to get in. Thisis a very basic question, but do build it?
(21:23):
Right. Or do you buy it? Or is it somethingin between? It's something in between. So
there are tools that allow you to basicallycollect agents out there. And there's a difference
between an agent and a context. Cause you heara lot about model context switching and things
like, don't know. if your audience knows thesethings. Or model context protocol. A context
(21:51):
is not an agent, but it has some agent capabilitiesbecause it's kind of specializing your model
in a certain area. But you would use this, butyou're not, if it's a true agent, then it's
probably tied to its own vector database. thatgets trained with specific information. It
(22:15):
might be company's information. It might beinformation, let's say if I'm a security agent,
then I'm going to be trained on the entire NISTsystem as well as all of my security architecture
that's currently in place. And that so thatit could monitor and assess instantly whether
(22:43):
there's security vulnerabilities, which youwouldn't ask Chet GPT to do that. No. Right?
Because it couldn't. Because it doesn't know anything about your organization or environment.
And it really also doesn't know how to prioritize what matters and what doesn't at any given
moment. Whereas a security agent, that wouldbe what it does.
(23:10):
I don't know if I answered that question. Oh,bad thing about building or buying. there
are- Or something in between, Yeah. So thereare tools that you can use to build workflows
and bring in different agents that alreadyexist. And you can use something like OpenAI
(23:32):
or Claude and use it to create an agent andgive it some intelligence and- give it a specific,
in this case, you're giving it a specific context. You could even tie a special machine learning
database to it and make it even more agenticin that way. And then build these workflows
(23:54):
where you're like, let's say a marketing workflow, where you're saying you first go out and research
all the people who are your ideal customerprofile. I was going to say ICP, but I'm trying
not to use acronyms because not everybody knowsevery acronym. Ideal customer profile. And
(24:16):
then it finds all these people that fit yourideal customer profile. Then it says, well,
which of these people are in the countriesthat I do business? And then it illuminates
the ones that aren't. then which ones, and itmay be using the same agent or different agents
to do this. Then once it's nailed it down tothe very discrete set of customers. Now the
(24:39):
next step in the workflow is, okay, now enrichtheir data of these people to find their email
and other ways of contacting them as well asother information about them so that I have
a really full picture of what kind of activityare they active socially? they speak? Do they
(25:01):
post? What are they speaking about? What arethey posting about? What events are they going
to? Things like that. So that would be thenext step and that'd be an agent that's doing
all the enriching. And then after that, thenext step would be to call basically call a
writing agent to go do, am I writing an email?Am I writing a LinkedIn connection post? Am
(25:22):
I doing both? Set up a drip campaign and startreaching out to these people one at a time
with very customized specific language, right?That is in your voice. It doesn't sound like
it's written by a typical AI outreach thing.All right, so these would be steps in a workflow
that you could use with several different toolsto build the workflows and then calling these
(25:45):
different agents. Let's go back to the launchedfirst. What would be a typical engagement with
a company? you know, they, um, the foundersthat have the greatest success in your experiences
(26:07):
are the ones that love the problem space andnot the product. All right. So walk my listeners
through. What a typical engagement. it's staffaugmentation. it full out outsourcing? it
tech? because it's very complex. I can touchso many. can touch high tech and high ticket
(26:34):
B2B products, sector agnostic. what, put somelegs on this for my listeners, please. Sure,
sure. We're not. so much a staff augmentationcompany, although we'll do that if asked to,
but that's not the kind of business that we look for. We look for project type work.
(26:58):
So a typical engagement for launch first wouldbe somebody wants to launch a product, they're
in the concept phase. We help refine the conceptand we build out, help that we do the design
and then we build a high fidelity prototype,which is a design prototype. When I demo a
design prototype to somebody, they think thatthey're looking at a finished product, but
(27:24):
it's not. It doesn't actually do anything.It just looks like it does everything. So
it's very animated set of mock-ups is anotherway to look at it. And it's important because
you can build out the big vision of the productthis way in a couple of months, whereas it
(27:46):
takes instead of, you so you're looking at thetwo year roadmap when we're done of the product.
If we were to build an MVP, then you're goingto see a very limited view of the product and
it's going to cost a lot more to build thatMVP than it takes to build this design prototype.
Now we're in the process of doing this. We'realso nailing down who that early adopter is.
(28:12):
And there's a, there's a very, metrics drivenmethodology for doing this. your launch first.
Within launch first, right. Okay. All right.And then we'll help the client build a marketing
funnel and help them start to generate sales. We're not doing the selling, they're doing
the selling. And it's important that foundersdo the selling because they need to hear what
(28:38):
customers are saying about the thing they'redemoing, why they want it, why they don't.
So that if we need to pivot, which we cando easily and quickly with a design prototype,
then we can pivot and then go and test themodel again, two or three or four times in
the space of a couple of months. And we'lleither find a path to revenue or accept the
(29:01):
fact that this probably isn't the right productfor the right time. But in the process of
doing this, you're learning a lot about themarket and about the potential customer. I
want to be clear about something. Almost everyfounder that comes to that I meet with, they
love the product, not the problem. They startedout with a problem that they realized they
(29:24):
had a good solution for and they forgot allabout the problem at that point. And so I spend
a lot of time with founders reminding themwhy the problem is all that matters and what
that means and how to approach customers, potentialcustomers so that you're syncing with their
problems, not telling them about this productthat you're building because nobody cares about
(29:47):
your product. All they care about is what they'restruggling with. And if they believe that
you really understand that, then they careabout whether you can solve that problem for
them or And can I be audacious and ask youwhat a typical engagement duration is like?
(30:09):
So this would be for launch first. Yes. If it'sa, and our hope is that they'll find a path
to revenue and start building the product andengage us for the development. Cause that's
really our business is building the products. So, but it's not a requirement. And, and
our typical engagement with our clients areseveral years. Not all of them, but most of
(30:34):
them, would say. Once they start working withus, they just continue to work with us until
they decide to bring in their own in-house team or they fail eventually, which many of our
clients do, which is why I created Launch First.Right. You often talk about your hyper exceptional
(30:54):
team at Techies. What is it that's so highlyexceptional? Talk to me about your team. Where
are they? Yeah. And if you go to my website,which is tekyz.com, you'll see at the very
top of it in the header above the fold, itsays hyper exceptional development team. And
I don't expect people to believe me becauseI write that down or I tell them that I expect
(31:19):
them to ask me, well, what does that mean? Doyou have evidence? And that's the question
I want to get because I do. Because when youwork in an exceptional manner, as a natural
consequence of working that way, you producecertain artifacts that the typical development
teams don't produce. And I'm not saying therearen't other exceptional teams, but they're
(31:43):
really few and far between. And what makes ateam exceptional is a constant need to improve
their ability to deliver and the level of qualitythat they deliver as well and the speed at
which they develop. It's all of these things.So, and, you know, after 18 years, we've done
(32:03):
a lot of improving and a lot of automation internally, because that allows our team to work in a
really disciplined protocol manner without havingto feel like they're under the strict discipline
and protocol of, you know, a difficult environmentto work in. And so we create automation everywhere
(32:26):
we can. The voice... tool is one of those automations. The way we do status reports, it's very clear
at the level of detail that we provide everyweek to every client in terms of status reports
where we're showing here's what we estimated,here's the actual, here's our percent variance
(32:49):
on how much time we spent and how much it'scosting. We want to always be within 10 %
above or below. Either being above or belowis not, know, the fact that we're ahead of
that doesn't necessarily mean that's a goodthing, right? So we want to be accurate with
our estimates. And we are typically within10%. In fact, our largest customer last year,
(33:14):
we did a retrospective and we were within sixand a half percent of what our estimates were
for the whole year. and that's a, we're prettyhappy with that number. I think most teams
are looking at many, many times that in termsof variance. it's not that uncommon for teams
(33:35):
to be double or triple what they're or evenhigher what the actual estimate was. So when
we do invoicing, we invoice for each personat their rate. based on their level of expertise,
which is all part of our agreement upfront.So the client is very transparent every month
(33:57):
for the hours that they work. And we attachthe daily time sheets to every invoice. I'm
the only company I know of right now that doesthat. I know there are others. I've seen monthly,
but I've never seen daily. Yeah. Yeah. Becausefor me, if I could ask, well, why did this
(34:19):
person ask a work that many hours that lastmonth? What did they do? I hate that feeling
that I get when somebody asks that question.I know they're only asking because they have
to justify it to somebody else or whatever thereason, but I don't like the way it feels because
it feels like my integrity is being questioned.I don't get upset at people for asking me that.
(34:41):
I just feel like I'm not giving them enoughinformation if they have to ask me that question.
So we started about eight years ago. providingthe daily time sheets because I don't like
that question. And we never get questioned onour invoices ever anymore. I bet you it's
informed you as well in future projects, maybe on including workflow automation in
(35:07):
your own internal processes, right? When yousee people's time sheets, right? And you've
gone over budget. So it informs you internally.So it's not only for the client. I suspect,
right? No, it's not. Right. And we use it ourselvesto also, because it also helps us looking at
our overhead costs because not everything getsbuilt to the client. And so we track all our
(35:28):
own times, you know, what we're spending doingwhat. And we don't get to, it's not like a
developer has to spend a lot of time or a QAperson or whatever, putting in a lot of detail.
We just need a couple of bullets, you know,every day in the time sheet with the, whatever
they spend. If they spent four hours on onething and three on another, they'll just break
(35:49):
it into two entries just to make it easy. Andthat's important for us, or they may be working
on two different projects and each project.So when we do the timesheets also every month,
we give our clients a breakdown by project.So if we're working on four different projects
for a client or even one project, but it hasfour different really functional elements that
(36:17):
are very clearly different. Like let's say amobile app and a web app and a particular
client implementation. Each one of those getsassigned its own project and we break down
summaries of the time spent on each of thoseevery month and who spent the time on those,
along with the daily time sheets, along withthe invoice. And nobody else does that because
(36:39):
it takes a lot of discipline and protocol andyou have to have lot of systems in place to
do that without literally getting everybodyto quit, right? That works for you. And nobody
minds doing it because it's easy because ofall the systems we put in place to do that.
That's the whole point, right? Right. were not particularly happy of getting asked that
(37:02):
question oftentimes. So eight years ago, youset out to provide the information on a daily
basis, which is incredible. We started thatwith blended rates like a lot of companies
do. And then I didn't like that because at theend of a project when most of it's QA, people
would start to get frustrated that they're stillgetting billed the same blended rate, even
(37:24):
though for the more expensive period at thebeginning of the project, I thought, okay,
forget this. Well, just bill based on individual. And then I didn't get those questions anymore,
but then I would get questions about individualson the month. And that's when I started doing
the time sheets. And like I said, I'm surethere's other companies that do it, but I haven't
(37:50):
run into one or somebody that works with one.So that's an exceptional thing that we do.
But it also allows us to do really, reallygood reporting to the client on status on what
we've spent our time on, what we're expectingto spend our time on next week, what we just
(38:10):
spent our time on this week, where we are. interms of our plan for the month, things like
that. So let's switch gears, David. Yeah.Back to actually the podcast and some of
my guests and listeners are corporate boarddirectors. So they're sitting on either advisory
(38:31):
boards or fiduciary corporate boards. And withall the hype around AI. it's not uncommon
for them to be asking, what are we doing, right?For existing companies, right? And I'd like
you to walk my listeners through while it'sin the, you know, in the imaginary realm,
(38:56):
what is it? I think any founder today that'sactually scaling, right? Has to have some AI
element. At least I've even heard you need tohave it. an AI officer in the company. So
what's your take on that? What would you respondto either to your board of advisors, your advisory
(39:17):
board, or your board of directors? So, andof course, a lot of it depends on the type
of company you are. Absolutely. Right. If you'remaking alternative material I-beams, for example,
for skyscraper construction, then AI, otherthan maybe in the design process of these specialized
(39:42):
materials, AI may not be as big a criticalfactor, although for invoice reconciliation
and distribution and scheduling and all that,AI could be a huge value to you if you don't
have super efficient systems already. For mosteverybody else though, if you have not embraced
(40:05):
the need to leverage AI and everything you'redoing, then you're way behind already. That
doesn't mean you have to be in a race to dothis. just, because I'm of the belief that
you have to slow down to speed up. But youdo need to make it a priority. And in a lot
(40:34):
of different ways. Number one is, The most obviousis workflow automation. You should be probably
tackling workflow automation as just a partof your constant improvement program to become
more efficient, whether it's with AI or not. But AI is particularly good at workflow automation
(40:55):
because it can tackle steps in that workflowthat couldn't be tackled without AI. So the
first thing the companies should be doing ifthey're not doing it is documenting all of
their processes, all of their tribal knowledgeinto playbooks. So when you have somebody who's
(41:17):
an expert in something in your company and they'rethe person who's the only one that knows how
to do it and so we can't live without them,that's a bottleneck for scaling. Because if
you bring somebody else in to expand their capacity,they're going to... put a big dependency on
that person with all the expertise, which isgoing to cause problems. So anybody in a
(41:42):
position like that should be documenting allof their procedures and protocols and especially
all the nuances and all the edge cases intoplaybooks. And there should be some centralized
playbook repository for the company. And thisbecomes part of your intellectual property
and part of your value if you ever you're tryingto raise money or you're trying to sell your
(42:04):
company. So it increases your value. So youdo that, then AI, you start to look at automating
those workflows because now they're documented.So now what can be automated in them from just
a workflow automation perspective. And thenhow much can you implement AI in there? Because
(42:25):
now AI can learn to make the same kinds of decisionsthat this person is making. And this is like
the low hanging fruit that I'm talking aboutright now. Right. Exactly. Right. Because the
bigger stuff is if we implement AI in here,what workflows would we totally throw away
and start from scratch? Because we can thinkof way more sophisticated ways of addressing
(42:50):
this now that we have intelligence involvedin all these steps. But that's later. worry
about that once you get your arms around implementingAI, automated workflows and then- So workflow
automation. So playbooks, workflows and AI inyour automated workflows. That's sort of the
stepped wise process. Excellent. You heard ithere on the founder sandbox. Thank you, David.
(43:20):
And if you're not sure how to do all that,ask AI, okay, here's my company. What should
I be focusing on if I wanna implement playbooks,workflow automation and AI? And AI will help
you figure this all out. Right. That's a jewelhere. So what'd you do? Chat GBT, co-pilot,
(43:42):
what's your complexity? Where would you go to?All right. Well, it just depends on the flavor
of the day. Right now. I was using chat GPTprimarily for this stuff just because it was
a first and I'm very comfortable with the apps.have them everywhere. And Claude's recently
(44:02):
come out with a new version and it's in someways I'm just finding the output way more organized
and smarter. And so I've been using Claude morein the last couple of weeks, but that'll change
in another week or two. Any one of them willdo a pretty decent job. I'm not using perplexity
(44:23):
because it's built on top of the other ones. But perplexity is a great tool if you're newer
with this because it makes some of the... It'sa little bit more accessible for somebody who
doesn't know how to use AI. Gemini is also really good, but that's more of a technical...
(44:46):
And there's so many things you can do. withAI that you wouldn't even think about. And
I'll give you an example, more as a brain openingexercise for everybody than anything else.
Because this is something I did about sevenweeks ago. I, chat GPT had just come out
a week or two before with their vision capabilityin the mobile app. And for those of you who
(45:12):
don't know it, with chat GPT, there's a talkbutton. It's not the microphone. It's the
one that looks like a sound wave in the mobileapp. You tap that, and now you have a voice
conversation with chat, which I use this constantly.Even when I'm working with, I've got some
(45:35):
contractors at my house whose English isn'tvery good, so I ask it to do real-time translation
for me. And it does matter the language. AndI start talking, and it translates to their
language. And they respond in their languageand it translates to English and it's doing
it perfectly. And so I can have a very naturalconversation with anybody just holding my phone
(45:57):
up in front of them now. Right? But it hasthis vision capability where when you go into
that voice mode, you tap the camera next toit, and now it's looking out the front of your
screen while you're talking to it. And so I'llgive you a couple of examples where I've used
it six weeks ago and again, like weeks laterand I now used it many times like this. I
(46:24):
was in Lowe's, which is a store for home improvement. And for some project I was on, my wife calls
me and says, I need fertilizer for a hibiscus.And I say, well, what do I get? She says, anything
that says hibiscus on it, it'll be fine. I said,okay, fine. And if anybody that knows these
big box stores, there's like hundreds of bagsof fertilizer of different brands. And I couldn't
(46:49):
find one that said hibiscus. This is a typicalthing with my wife. Oh, just look for this.
And of course, there isn't that. So I askedChess GPT, okay, I'm in Lowe's and I'm looking
for a fertilizer for hibiscus. What would yousuggest? And it said, oh, there's a number
of brands that are high acid. And I said, we'llrecommend a brand. Tonal is a really good brand.
(47:14):
And I said, okay. So I'm looking and I can'tfind it. So I walked 30 feet back and I'm talking,
right? I'm having this, know, people are lookingat me like, what the hell is he doing? And
I walked 30 feet back because there's many,many shelves, you know, columns of shelves
with fertilizer. I walked back and I turnedon the vision and I say, okay, there's all
(47:39):
the fertilizers. And I'm moving my phone acrossall these shelves. say, do you see tonal here?
And it says, yes, look for the one in the redand white bag. And I see it on the shelf.
So I walk straight forward. see a red and whitebag. That's not tonal. said, this isn't it.
And she, cause it's a woman's voice that I have,she says, it's two shelves to the left, second
(48:04):
from the top. I walk over there and it's rightwhere she said it was. Crazy. And you're not
a beta user. So this is available today. Thisis available. It's been available for a couple
of months. And then My daughter-in-law askedme to get something from the pharmacy, from
CVS, another big box pharmacy store, right?And this is something I don't even know if
(48:27):
I'm in the right aisle because it's somethingI've never bought. So I ask it, I say, I'm
looking for this brand and I'm not sure ifI'm in the right aisle or not, but I'm going
to walk down the aisle and tell me if you seeit. As I'm walking down the aisle, holding
it straight forward so it can see both sides. And it says, well, Yes, I'm familiar with
the brand. You should look for it in a greenand white box. then she goes like this. Oh,
(48:51):
I see it. It's down there on the right on thebottom shelf. And I turn and I look and it's
right by my right foot. You heard it here. Thisis crazy. think it's a bit creepy. How many
times have you been looking for something ona shelf? You know, and you're like, oh, how
long, how many hours is this going to take meto spot it? Good internet connection and all
(49:14):
that. So, oh my goodness. It's creepy and it'swonderful. So same time. the same time. Yeah.
Yeah. For quality of life and even for, um,yeah. So That's a mind opening thing is all
the reason I bring that up. Excellent. Hey,let's go. Let's continue on in the founder
sandbox. I'd like to ask each of my guests to share with me. I'm all about working with
(49:40):
resilient, purpose driven and scalable companiesin the growth phase. So what does resilience
mean to you? You can either answer, you know,what's the first thing that comes out of your,
you cannot use chat, GBT. I'm not fancy. Nohands. No hands, and I don't have the voice
version going because you'd hear it. Podcastwe could do it. And we are real. We're not.
(50:04):
Yeah, we are real. We're not. So I think that's,I don't think that's a difficult question to
answer. Resilience means opportunity. So nomatter what happens, even if it seems terrible,
what opportunity does that create? Excellent.If you ask that. keep reframing everything
from that perspective, it creates resilience.Right. Thank you. What about purpose-driven?
(50:33):
Purpose-driven means having a clear long-termpath and goal and asking yourself if the
things you're doing keep you on purpose to that.
(50:56):
Scalable. What's scalable mean for you? Scalablefor me means eliminating tribal knowledge
or not eliminating it, but documenting tribalknowledge. First of all, figuring out how
(51:18):
you generate revenue and then how you expandyour ability to generate revenue, which means
growing your growing your team, growing yourcapacity and identifying the bottlenecks and
focusing all your energy on the bottlenecks.And usually the bottlenecks have to do with
tribal knowledge or with lack of workflow automation.Wow, you know, it's easier said than done though,
(51:45):
that tribal knowledge, it is resistant, right?Oh yeah, because it's career, what's the
word I'm trying to think of? It keeps youin your job forever if you're the only one
that knows how to do the thing. Absolutely.That's for another podcast, David. My final
(52:05):
question today is, did you have fun in theFounder Sandbox? Oh, yes. I had a lot of fun.
Thanks. That's a great question too. Thank you,Brenda. Did you have fun? Did you? I had had
fun. And particularly in this last part, right?Cause we're talking about some heavy duty,
(52:26):
you know, uses of, um, agentic AI, right. Andscalable, you know, LTV, CAC and all that.
And then we get to hear these real life, youknow, kind of creepy, um, uh, uses of, um,
on our phones today with, um, with AI, whichis, which is quite amazing. But I also know
(52:48):
that in your world of techies, your team, whichis distributed, have a lot of fun events too.
So you probably- have one more thing on thewhole scalable thing. You have to be compassionately
ruthless or ruthlessly compassionate, howeveryou want to say it. Okay. So that the people,
(53:09):
every, and the ruthless is anything that's goingto get in the way of you growing your company,
which benefits everybody in the company. itneeds to be addressed in a ruthless way. But
if you build a culture of ruthlessly compassionate,then all the people that work for you feel
that same level of ruthlessness to protect thecompany and make it grow. And you practice
(53:36):
what you preach, I suspect, at Techies. Yes.Yes. It took me a while, but if we accidentally
hire the wrong person, either because we madea mistake in the process or they faked us out
and we recognize they're not smart enough. Literally,that's usually the problem. They're not smart
enough to carry their weight. We fire them immediately.We don't try to bring them along because you
(54:01):
can't improve somebody's IQ. You can improveany other aspect, but their IQ is their IQ.
And that will be a bottleneck forever. inour team and it'll require other people to
carry that person. And it sends the wrong messageto the team that I don't value them enough
to make sure that we only surround them withpeople that are going to inspire them and help
(54:26):
them grow. Excellent. And I suspect they arenot fungible by AI, your employees, not techies.
I mean, we've gotten better and better. at notmaking those mistakes over the years. So that
doesn't typically happen. takes us, we're muchmore careful about how we hire. AI gives us
(54:50):
the ability to recruit faster, more broadly, along with workflow automation. But what
I mean by real, this is the compassionate. Oncemy team understood this, now they embody that
and they will get rid of somebody if they madea mistake. I don't have to force the issue
ever anymore because they recognize how much,important it is to protect their teams. So
(55:14):
to my listeners, if you liked this episode todaywith the CEO and founder of Techies, sign up
for the monthly release of founders, businessowners, corporate directors, and professional
service providers who provide their examplesof how they're building companies or consulting
with companies to make them more resilient,scalable, and purpose-driven. to make profits
(55:40):
for good. Signing off for today. See you nextmonth in the Founder Sandbox. Thank you.
Welcome back to the Founders Sandbox. I amBrenda McCabe, the host here on this monthly
podcast, now in its third season. This podcastreaches entrepreneurs, business owners that
(00:30):
are scaling. professional service providersthat provide services to these entrepreneurs,
and corporate board directors who, like me,are building resilient, purpose-driven, and
scalable businesses with great corporate governance.My guests to this podcast are business owners
themselves, professional service providers,and corporate directors who, like me, want
(00:54):
to use the power of the private company tobuild a better world through storytelling with
each of my guests in the sandbox. My goal isto provide a fun sandbox environment where
we can equip one founder at a time to builda better world through great corporate governance.
So today I'm absolutely delighted to have asmy guest, David Hirschfeld. David is the owner
(01:22):
and CEO of Techies, 17 or 18 year old businessnow that boasts a hyper exceptional development
team that are building high ticket productsin the B2B space. Welcome David to the Founder
Sandbox. Hi Brenda and thanks for having me.Great. So I'm delighted that we actually did
(01:45):
a dry run in February. We've known each otherfor some time and AI, we're going to be touching
on AI. And I think that the world of AI particularlyin software development, has changed significantly
since we last spoke in February. So we're goingto be getting into some, I think, novel concepts
(02:10):
for the listeners of the Founder Sandbox. SoI wanted to, you I always talk about how I
like to work with growth stage companies that typically are bootstrapped and It's only
at a later stage do they seek institutionalinvestment by building great corporate governance
(02:33):
and reducing the reliance on investor funding until such a time that they choose the right
type of investors that can help them scale.So when I found out what you do at Techies
with Launch First and the type of work youdo in B2B businesses, I absolutely wanted to
(02:53):
have you here on the founder sandbox. So let'sjump right in, right? I think I'm eager to
learn more about how to scale your bespoke developmentat Techies, right? To scale my own business?
(03:13):
Okay. So there's a lot of different aspectsto scaling my business and I bootstrapped for
the last 18 years. I've never taken any investment with techies. And I've done that very specifically
(03:33):
because it gives me a lot of freedom. I don'thave a reporting structure that I have to
worry about. That doesn't mean that I can belazy with my team. To grow my team, I have
a philosophy that I only hire people that aresmarter than I am. And the ones that are
(04:01):
in a position to hire, they can only hire peoplethat are smarter than them. And by really
sticking to this philosophy, even though sometimesit makes us grow a little slower than we would
like, it means that when we bring in people,those people contribute immediately and contribute
in a way that it's our job to get the impedimentsout of their way and to facilitate them so
(04:28):
that they can contribute and help us grow thecompany. So I call it the ball rolls uphill
here because my job is to support everybodythat is above me, which is everybody. And then
the people that I support directly, their jobis to support the people that are above them.
(04:51):
Because if we're hiring correctly, then peoplethat we bring in can contribute in the area
that we're bringing them in way more than theperson that's hiring them. Okay. Thank you
for that. So before you launched Techies, youhad a career in companies like, I believe,
(05:13):
Computer Associates, right? Texas Experimentsand TelaMotorola. There was a period of time
between your experience in these large corporationsbefore your launch tech is where you actually
had your own startup and you sold it in 2000,right? And I believe you also learned perhaps
(05:36):
with the second startup about how hard it isto find product market fit. Can you talk to
that for my listeners, please? I don't knowthat it's that hard to find product market
fit. It depends if that's your focus or not.If your focus is to nail down product market
fit, then it's not that hard to determine whetheryou can achieve that or not fairly quickly.
(06:04):
You can do that by selling your product topotential customers. That sounds strange.
Of course, we all want to sell our products,but What I'm suggesting is you start selling
your product before you have a product, beforeyou have a full product. And I don't mean
an MVP, but a design prototype. You go out tothe market and you start to sell it. If you
(06:29):
have product market fit and you've identifiedthe early adopter in your market and you know
that they have a very high need from a perceptionperspective and there's a big cost to the
problem that you're solving. then you can offerthem a big enough value upfront that they'll
buy your product early and you can prove thatthere's a market for your product and they'll
(06:52):
buy it in enough numbers that you can achievea measurable metric, which I kind of call
the golden ratio, which is three to one in termsof what is the lifetime value of a customer
versus what does it cost to acquire that customer?And you can get to that three to one ratio.
(07:13):
in a prelaunch sale model before you ever starteddeveloping your product as a way of proving
product market fit. Or you pivot quickly andcheaply because you're not having to rebuild
a product that you've built in the wrong way.Or you fail fast and cheap. And every entrepreneur's
(07:34):
first goal should be to fail fast and cheap.know that sounds backwards, but that should
be your goal is that you can fail fast and cheapor if you If you fail to fail fast and cheap,
that means you've found a path to revenue and product market fit. And now you know you have
a viable business. making the investment tobuild the product is a no brainer. And you
(07:59):
came upon this methodology, right? Yes. becauseyou did yourself when you had your first company,
you did not understand the funding part, right? Can you talk? a bit about your specific example
and then how that's informed now 17 years oftechies and over 90 projects with startups.
(08:23):
Okay. So my first company was Bootstrap. Okay. And that one was successful and we grew it
despite me, it was me and a partner. And despiteourselves, we grew it over eight years. where
he ended up with 800 customers in 22 countriesand sold it to a publicly traded firm out of
(08:47):
Toronto. That was in the product food, snackfood distribution business because that was
what our product was focused on. So I startedanother company about five years later, not
realizing the things that I did the first time.that made it so successful, which really
(09:13):
fit the launch first model to a large degree. But the second time I built a product that
would have been successful had I followed myfirst model, but I didn't. So I went the route
of building an MVP and getting customers ona free version of it, and then going out and
trying to raise money, which is the very classicapproach that the SaaS products take now.
(09:40):
And the problem is with that approach is thatyou end up digging a really deep hole in terms
of the investment that you make to build theproduct with enough functionality that you
can convince people it's worth putting an investmentin and you're not generating any revenue at
the time. And I should have just started sellingthe product and generating subscription revenue
(10:01):
right from the beginning. First of all, I wouldhave been able to raise money much more easily.
Secondly, I would have not needed to raisemoney as much if I'd focused on sales. The
problem with a lot of founders is they fallin love with their product. They believe that
people will buy it at enough numbers and thatinvestors will see the potential. they're afraid
(10:29):
of sales. I've fallen into this trap beforetoo. I've done it both ways. And I can tell
you selling early and staying focused on thecustomer and the problem are the way to be
successful. So founders who I find are consistentlysuccessful, they are focused on the problem,
(10:50):
they love the problem. The product is just thenatural conclusion to solving the problem,
not something to be in love with. They spendtheir time talking to customers about the problems.
So how does a potential customer find you andwork with you? Oh, they can find me at Techies
(11:11):
or they can find me at LaunchFirst, was spelledlaunch1st.com. And they can find me on LinkedIn.
And then to work with me, it's just give mea call, send me an email, we'll set up a Zoom.
I'll start to learn about what you're tryingto accomplish and what your requirements are.
(11:33):
And I'll typically spend quite a bit of timewith any potential clients. in one to usually
multiple calls or Zooms, learning and creatingestimates and doing a lot of work in advance
with the idea that there'll be a natural conclusionat the end of this that they'll wanna start
(11:54):
working with me in a paid fashion. So there'sa lot of value that my clients get from me
whether they end up contracting me or not. And how, again, back to, thank you for that
and that. how to contact you will be in theshow notes. But what types of sectors do you
work in? You know, in your introduction, Italk about high ticket B2B, right? who are
(12:21):
the, so what founder that's has some ideatoday? What would be their call to action
to find techies? And what would you, is it launchfirst before you go down? No, it's not necessarily.
It may be an existing company that is tryingto implement AI or implement workflow automation,
(12:46):
or they have a project and they don't have theIT team or capacity to handle it. We love
those types of projects. It might be an existingstartup that is struggling with their software
development team and they're not getting tothe end goal that they're expecting and the
(13:07):
product's buggy, it's taking too long, there'sconstant delays, they're way over budget and
they need to get this thing done. And I callthose recovery projects, they're probably
my favorite because people recognize very quickly the difference that we bring. and they really,
(13:34):
really appreciate us. As far as what sectors, business sectors, healthcare, law enforcement,
prop tech, real estate, finance, entertainment,I mean, we work in many, many different sectors
over the last 18 years. So regardless in B2B, B2B2C, not so much e-commerce unless
(14:01):
there's some complex workflow associated withyour particular e-commerce, but there's lots
of really good solutions for e-commerce that don't require developers to be involved.
But mobile, web, IoT, definitely everythingis AI now. Absolutely. And in fact, when we
(14:25):
last spoke, I'd like to say that you startedto drink your own Kool-Aid at Techies. you're
starting to actually use AI automation for internalfunctions as well as projects at Techies. So
can you walk my listeners through how you'reusing AI automation and what's the latest
(14:52):
with agentic AI? So let's do the first. Yeah, okay. So there are a bunch of questions there.
So let me start with that we're building productsinternally at Techies to help us with our
own workflows. These products though are applicable to almost any development company
(15:17):
or any company with a development team. Someof them are, and some of them are applicable
to companies that are, well, so one product is putting voice capability in front of project
management tool. and we use JIRA and JIRA isan incredibly technical tool for project managers
(15:37):
and development teams to use to their projects,requirements, their track bugs, all of that.
And so your relationship with what I call relationshipwith project management is very technical one.
If you're a client, some clients are willingto go through the learning curve so that they
(15:57):
can enter their own... bugs and feature requestsand things like that directly into JIRA. Most
don't. They want to send us emails, whichis fine, and just give us a list of what's
going on and the problems that they're findingor the things that they need for a future
version and the planning and the documentation,everything else. This is a real technical thing.
(16:20):
We're going to make it a very natural personalrelationship by adding voice in front of all
this so that you can be sharing your screenwith your little voice app and say, just found
a problem on the screen. And the voice appcan see the screen. It knows your project.
It knows your requirements. And it can identifyproblems on the screen that you may not have
(16:43):
even noticed. And it can also prevent you fromreporting bugs that have already been reported
and tell you when they're planned to be built. And all of this just with a verbal discussion
with the app. that basically knows your project. Kind of like talking to a project manager
(17:03):
in real time, but they don't have to write downnotes and they can instantly look up anything
about your project in terms of what's been reportedin terms of bugs or feature requests and update
them or create new ones for you or just reportthem to you and tell you when things are planned
to be built and released or. where they've alreadybeen released and maybe you need to clear your
(17:27):
cache so you can see the change, whatever. Yeah. So it be like an avatar, but it's trained
and it's specific to Jira in your case? Inthe first version, it's actually being built
architected so that we'll be able to add otherproject management tools to it besides Jira
in the future. to begin with, because we useJira, it's going to work directly with Jira
(17:52):
to start. And this, by the way, you askedabout agentic workflows, right? So we're
building an agentic workflow in this tool wherewe have more different agents that work together
to resolve these issues. so we have an agentthat reads and writes documentation to JIRA.
(18:16):
We have an agent that communicates with theuser and the user might be the programmer might
be a person in QA, it might be a client fora lot of different things. And we have an analyst
agent that when the person talks, the voiceagent says to the analyst agent, here's what
(18:37):
I understand. Here's the information I justgot. Go do your work and come back and get
me the answer. And it'll speak to the JIRA agentto get the information. It will also speak
directly to us. a vector database, which isa database where all the documentation from
that project is ingested into our own separateAI model so that the context of all the communication
(19:06):
is about their project and doesn't go off intoother directions. And then can get back.
So this is an agentic workflow. The idea ofagents is like everybody keeps talking about
agents. Not everybody is really clear on whatthat even means. Can you define that? an
(19:29):
agent is an AI model that you can interactwith that is focused on one specific area
of expertise. So if it's a travel agent, theword agent fits very well there, then their
expertise would be on everything related totravel and booking travel and looking up options
(19:52):
and comparing prices. And that would be anAI travel agent. So that's very different
from an AI project management agent, very differentfrom an AI financial analyst agent. So each
agent specializes in its own area of expertiseand may draw from specific repositories of
(20:19):
information that are specific to that particularagent's area of expertise. And they actually
look from the perspective of that type of person,if it was a person. So, and so they'll respond
in a way that is consistent with how somebodywho is a project manager would respond to you
(20:41):
when you're talking to them, asking you questionsabout your requirements, knows what information
it needs to be able to assess it properly, thingslike that. wouldn't be very good about travel
because that's not its area of expertise. Right. So is it common to have companies that are
(21:02):
creating with their own large language model,right? Or their workflow processes internally
to the company to create their own agent AI? Or is there a marketplace now where you can
say, want this type of agent to get in. Thisis a very basic question, but do build it?
(21:23):
Right. Or do you buy it? Or is it somethingin between? It's something in between. So
there are tools that allow you to basicallycollect agents out there. And there's a difference
between an agent and a context. Cause you heara lot about model context switching and things
like, don't know. if your audience knows thesethings. Or model context protocol. A context
(21:51):
is not an agent, but it has some agent capabilitiesbecause it's kind of specializing your model
in a certain area. But you would use this, butyou're not, if it's a true agent, then it's
probably tied to its own vector database. thatgets trained with specific information. It
(22:15):
might be company's information. It might beinformation, let's say if I'm a security agent,
then I'm going to be trained on the entire NISTsystem as well as all of my security architecture
that's currently in place. And that so thatit could monitor and assess instantly whether
(22:43):
there's security vulnerabilities, which youwouldn't ask Chet GPT to do that. No. Right?
Because it couldn't. Because it doesn't know anything about your organization or environment.
And it really also doesn't know how to prioritize what matters and what doesn't at any given
moment. Whereas a security agent, that wouldbe what it does.
(23:10):
I don't know if I answered that question. Oh,bad thing about building or buying. there
are- Or something in between, Yeah. So thereare tools that you can use to build workflows
and bring in different agents that alreadyexist. And you can use something like OpenAI
(23:32):
or Claude and use it to create an agent andgive it some intelligence and- give it a specific,
in this case, you're giving it a specific context. You could even tie a special machine learning
database to it and make it even more agenticin that way. And then build these workflows
(23:54):
where you're like, let's say a marketing workflow, where you're saying you first go out and research
all the people who are your ideal customerprofile. I was going to say ICP, but I'm trying
not to use acronyms because not everybody knowsevery acronym. Ideal customer profile. And
(24:16):
then it finds all these people that fit yourideal customer profile. Then it says, well,
which of these people are in the countriesthat I do business? And then it illuminates
the ones that aren't. then which ones, and itmay be using the same agent or different agents
to do this. Then once it's nailed it down tothe very discrete set of customers. Now the
(24:39):
next step in the workflow is, okay, now enrichtheir data of these people to find their email
and other ways of contacting them as well asother information about them so that I have
a really full picture of what kind of activityare they active socially? they speak? Do they
(25:01):
post? What are they speaking about? What arethey posting about? What events are they going
to? Things like that. So that would be thenext step and that'd be an agent that's doing
all the enriching. And then after that, thenext step would be to call basically call a
writing agent to go do, am I writing an email?Am I writing a LinkedIn connection post? Am
(25:22):
I doing both? Set up a drip campaign and startreaching out to these people one at a time
with very customized specific language, right?That is in your voice. It doesn't sound like
it's written by a typical AI outreach thing.All right, so these would be steps in a workflow
that you could use with several different toolsto build the workflows and then calling these
(25:45):
different agents. Let's go back to the launchedfirst. What would be a typical engagement with
a company? you know, they, um, the foundersthat have the greatest success in your experiences
(26:07):
are the ones that love the problem space andnot the product. All right. So walk my listeners
through. What a typical engagement. it's staffaugmentation. it full out outsourcing? it
tech? because it's very complex. I can touchso many. can touch high tech and high ticket
(26:34):
B2B products, sector agnostic. what, put somelegs on this for my listeners, please. Sure,
sure. We're not. so much a staff augmentationcompany, although we'll do that if asked to,
but that's not the kind of business that we look for. We look for project type work.
(26:58):
So a typical engagement for launch first wouldbe somebody wants to launch a product, they're
in the concept phase. We help refine the conceptand we build out, help that we do the design
and then we build a high fidelity prototype,which is a design prototype. When I demo a
design prototype to somebody, they think thatthey're looking at a finished product, but
(27:24):
it's not. It doesn't actually do anything.It just looks like it does everything. So
it's very animated set of mock-ups is anotherway to look at it. And it's important because
you can build out the big vision of the productthis way in a couple of months, whereas it
(27:46):
takes instead of, you so you're looking at thetwo year roadmap when we're done of the product.
If we were to build an MVP, then you're goingto see a very limited view of the product and
it's going to cost a lot more to build thatMVP than it takes to build this design prototype.
Now we're in the process of doing this. We'realso nailing down who that early adopter is.
(28:12):
And there's a, there's a very, metrics drivenmethodology for doing this. your launch first.
Within launch first, right. Okay. All right.And then we'll help the client build a marketing
funnel and help them start to generate sales. We're not doing the selling, they're doing
the selling. And it's important that foundersdo the selling because they need to hear what
(28:38):
customers are saying about the thing they'redemoing, why they want it, why they don't.
So that if we need to pivot, which we cando easily and quickly with a design prototype,
then we can pivot and then go and test themodel again, two or three or four times in
the space of a couple of months. And we'lleither find a path to revenue or accept the
(29:01):
fact that this probably isn't the right productfor the right time. But in the process of
doing this, you're learning a lot about themarket and about the potential customer. I
want to be clear about something. Almost everyfounder that comes to that I meet with, they
love the product, not the problem. They startedout with a problem that they realized they
(29:24):
had a good solution for and they forgot allabout the problem at that point. And so I spend
a lot of time with founders reminding themwhy the problem is all that matters and what
that means and how to approach customers, potentialcustomers so that you're syncing with their
problems, not telling them about this productthat you're building because nobody cares about
(29:47):
your product. All they care about is what they'restruggling with. And if they believe that
you really understand that, then they careabout whether you can solve that problem for
them or And can I be audacious and ask youwhat a typical engagement duration is like?
(30:09):
So this would be for launch first. Yes. If it'sa, and our hope is that they'll find a path
to revenue and start building the product andengage us for the development. Cause that's
really our business is building the products. So, but it's not a requirement. And, and
our typical engagement with our clients areseveral years. Not all of them, but most of
(30:34):
them, would say. Once they start working withus, they just continue to work with us until
they decide to bring in their own in-house team or they fail eventually, which many of our
clients do, which is why I created Launch First.Right. You often talk about your hyper exceptional
(30:54):
team at Techies. What is it that's so highlyexceptional? Talk to me about your team. Where
are they? Yeah. And if you go to my website,which is tekyz.com, you'll see at the very
top of it in the header above the fold, itsays hyper exceptional development team. And
I don't expect people to believe me becauseI write that down or I tell them that I expect
(31:19):
them to ask me, well, what does that mean? Doyou have evidence? And that's the question
I want to get because I do. Because when youwork in an exceptional manner, as a natural
consequence of working that way, you producecertain artifacts that the typical development
teams don't produce. And I'm not saying therearen't other exceptional teams, but they're
(31:43):
really few and far between. And what makes ateam exceptional is a constant need to improve
their ability to deliver and the level of qualitythat they deliver as well and the speed at
which they develop. It's all of these things.So, and, you know, after 18 years, we've done
(32:03):
a lot of improving and a lot of automation internally, because that allows our team to work in a
really disciplined protocol manner without havingto feel like they're under the strict discipline
and protocol of, you know, a difficult environmentto work in. And so we create automation everywhere
(32:26):
we can. The voice... tool is one of those automations. The way we do status reports, it's very clear
at the level of detail that we provide everyweek to every client in terms of status reports
where we're showing here's what we estimated,here's the actual, here's our percent variance
(32:49):
on how much time we spent and how much it'scosting. We want to always be within 10 %
above or below. Either being above or belowis not, know, the fact that we're ahead of
that doesn't necessarily mean that's a goodthing, right? So we want to be accurate with
our estimates. And we are typically within10%. In fact, our largest customer last year,
(33:14):
we did a retrospective and we were within sixand a half percent of what our estimates were
for the whole year. and that's a, we're prettyhappy with that number. I think most teams
are looking at many, many times that in termsof variance. it's not that uncommon for teams
(33:35):
to be double or triple what they're or evenhigher what the actual estimate was. So when
we do invoicing, we invoice for each personat their rate. based on their level of expertise,
which is all part of our agreement upfront.So the client is very transparent every month
(33:57):
for the hours that they work. And we attachthe daily time sheets to every invoice. I'm
the only company I know of right now that doesthat. I know there are others. I've seen monthly,
but I've never seen daily. Yeah. Yeah. Becausefor me, if I could ask, well, why did this
(34:19):
person ask a work that many hours that lastmonth? What did they do? I hate that feeling
that I get when somebody asks that question.I know they're only asking because they have
to justify it to somebody else or whatever thereason, but I don't like the way it feels because
it feels like my integrity is being questioned.I don't get upset at people for asking me that.
(34:41):
I just feel like I'm not giving them enoughinformation if they have to ask me that question.
So we started about eight years ago. providingthe daily time sheets because I don't like
that question. And we never get questioned onour invoices ever anymore. I bet you it's
informed you as well in future projects, maybe on including workflow automation in
(35:07):
your own internal processes, right? When yousee people's time sheets, right? And you've
gone over budget. So it informs you internally.So it's not only for the client. I suspect,
right? No, it's not. Right. And we use it ourselvesto also, because it also helps us looking at
our overhead costs because not everything getsbuilt to the client. And so we track all our
(35:28):
own times, you know, what we're spending doingwhat. And we don't get to, it's not like a
developer has to spend a lot of time or a QAperson or whatever, putting in a lot of detail.
We just need a couple of bullets, you know,every day in the time sheet with the, whatever
they spend. If they spent four hours on onething and three on another, they'll just break
(35:49):
it into two entries just to make it easy. Andthat's important for us, or they may be working
on two different projects and each project.So when we do the timesheets also every month,
we give our clients a breakdown by project.So if we're working on four different projects
for a client or even one project, but it hasfour different really functional elements that
(36:17):
are very clearly different. Like let's say amobile app and a web app and a particular
client implementation. Each one of those getsassigned its own project and we break down
summaries of the time spent on each of thoseevery month and who spent the time on those,
along with the daily time sheets, along withthe invoice. And nobody else does that because
(36:39):
it takes a lot of discipline and protocol andyou have to have lot of systems in place to
do that without literally getting everybodyto quit, right? That works for you. And nobody
minds doing it because it's easy because ofall the systems we put in place to do that.
That's the whole point, right? Right. were not particularly happy of getting asked that
(37:02):
question oftentimes. So eight years ago, youset out to provide the information on a daily
basis, which is incredible. We started thatwith blended rates like a lot of companies
do. And then I didn't like that because at theend of a project when most of it's QA, people
would start to get frustrated that they're stillgetting billed the same blended rate, even
(37:24):
though for the more expensive period at thebeginning of the project, I thought, okay,
forget this. Well, just bill based on individual. And then I didn't get those questions anymore,
but then I would get questions about individualson the month. And that's when I started doing
the time sheets. And like I said, I'm surethere's other companies that do it, but I haven't
(37:50):
run into one or somebody that works with one.So that's an exceptional thing that we do.
But it also allows us to do really, reallygood reporting to the client on status on what
we've spent our time on, what we're expectingto spend our time on next week, what we just
(38:10):
spent our time on this week, where we are. interms of our plan for the month, things like
that. So let's switch gears, David. Yeah.Back to actually the podcast and some of
my guests and listeners are corporate boarddirectors. So they're sitting on either advisory
(38:31):
boards or fiduciary corporate boards. And withall the hype around AI. it's not uncommon
for them to be asking, what are we doing, right?For existing companies, right? And I'd like
you to walk my listeners through while it'sin the, you know, in the imaginary realm,
(38:56):
what is it? I think any founder today that'sactually scaling, right? Has to have some AI
element. At least I've even heard you need tohave it. an AI officer in the company. So
what's your take on that? What would you respondto either to your board of advisors, your advisory
(39:17):
board, or your board of directors? So, andof course, a lot of it depends on the type
of company you are. Absolutely. Right. If you'remaking alternative material I-beams, for example,
for skyscraper construction, then AI, otherthan maybe in the design process of these specialized
(39:42):
materials, AI may not be as big a criticalfactor, although for invoice reconciliation
and distribution and scheduling and all that,AI could be a huge value to you if you don't
have super efficient systems already. For mosteverybody else though, if you have not embraced
(40:05):
the need to leverage AI and everything you'redoing, then you're way behind already. That
doesn't mean you have to be in a race to dothis. just, because I'm of the belief that
you have to slow down to speed up. But youdo need to make it a priority. And in a lot
(40:34):
of different ways. Number one is, The most obviousis workflow automation. You should be probably
tackling workflow automation as just a partof your constant improvement program to become
more efficient, whether it's with AI or not. But AI is particularly good at workflow automation
(40:55):
because it can tackle steps in that workflowthat couldn't be tackled without AI. So the
first thing the companies should be doing ifthey're not doing it is documenting all of
their processes, all of their tribal knowledgeinto playbooks. So when you have somebody who's
(41:17):
an expert in something in your company and they'rethe person who's the only one that knows how
to do it and so we can't live without them,that's a bottleneck for scaling. Because if
you bring somebody else in to expand their capacity,they're going to... put a big dependency on
that person with all the expertise, which isgoing to cause problems. So anybody in a
(41:42):
position like that should be documenting allof their procedures and protocols and especially
all the nuances and all the edge cases intoplaybooks. And there should be some centralized
playbook repository for the company. And thisbecomes part of your intellectual property
and part of your value if you ever you're tryingto raise money or you're trying to sell your
(42:04):
company. So it increases your value. So youdo that, then AI, you start to look at automating
those workflows because now they're documented.So now what can be automated in them from just
a workflow automation perspective. And thenhow much can you implement AI in there? Because
(42:25):
now AI can learn to make the same kinds of decisionsthat this person is making. And this is like
the low hanging fruit that I'm talking aboutright now. Right. Exactly. Right. Because the
bigger stuff is if we implement AI in here,what workflows would we totally throw away
and start from scratch? Because we can thinkof way more sophisticated ways of addressing
(42:50):
this now that we have intelligence involvedin all these steps. But that's later. worry
about that once you get your arms around implementingAI, automated workflows and then- So workflow
automation. So playbooks, workflows and AI inyour automated workflows. That's sort of the
stepped wise process. Excellent. You heard ithere on the founder sandbox. Thank you, David.
(43:20):
And if you're not sure how to do all that,ask AI, okay, here's my company. What should
I be focusing on if I wanna implement playbooks,workflow automation and AI? And AI will help
you figure this all out. Right. That's a jewelhere. So what'd you do? Chat GBT, co-pilot,
(43:42):
what's your complexity? Where would you go to?All right. Well, it just depends on the flavor
of the day. Right now. I was using chat GPTprimarily for this stuff just because it was
a first and I'm very comfortable with the apps.have them everywhere. And Claude's recently
(44:02):
come out with a new version and it's in someways I'm just finding the output way more organized
and smarter. And so I've been using Claude morein the last couple of weeks, but that'll change
in another week or two. Any one of them willdo a pretty decent job. I'm not using perplexity
(44:23):
because it's built on top of the other ones. But perplexity is a great tool if you're newer
with this because it makes some of the... It'sa little bit more accessible for somebody who
doesn't know how to use AI. Gemini is also really good, but that's more of a technical...
(44:46):
And there's so many things you can do. withAI that you wouldn't even think about. And
I'll give you an example, more as a brain openingexercise for everybody than anything else.
Because this is something I did about sevenweeks ago. I, chat GPT had just come out
a week or two before with their vision capabilityin the mobile app. And for those of you who
(45:12):
don't know it, with chat GPT, there's a talkbutton. It's not the microphone. It's the
one that looks like a sound wave in the mobileapp. You tap that, and now you have a voice
conversation with chat, which I use this constantly.Even when I'm working with, I've got some
(45:35):
contractors at my house whose English isn'tvery good, so I ask it to do real-time translation
for me. And it does matter the language. AndI start talking, and it translates to their
language. And they respond in their languageand it translates to English and it's doing
it perfectly. And so I can have a very naturalconversation with anybody just holding my phone
(45:57):
up in front of them now. Right? But it hasthis vision capability where when you go into
that voice mode, you tap the camera next toit, and now it's looking out the front of your
screen while you're talking to it. And so I'llgive you a couple of examples where I've used
it six weeks ago and again, like weeks laterand I now used it many times like this. I
(46:24):
was in Lowe's, which is a store for home improvement. And for some project I was on, my wife calls
me and says, I need fertilizer for a hibiscus.And I say, well, what do I get? She says, anything
that says hibiscus on it, it'll be fine. I said,okay, fine. And if anybody that knows these
big box stores, there's like hundreds of bagsof fertilizer of different brands. And I couldn't
(46:49):
find one that said hibiscus. This is a typicalthing with my wife. Oh, just look for this.
And of course, there isn't that. So I askedChess GPT, okay, I'm in Lowe's and I'm looking
for a fertilizer for hibiscus. What would yousuggest? And it said, oh, there's a number
of brands that are high acid. And I said, we'llrecommend a brand. Tonal is a really good brand.
(47:14):
And I said, okay. So I'm looking and I can'tfind it. So I walked 30 feet back and I'm talking,
right? I'm having this, know, people are lookingat me like, what the hell is he doing? And
I walked 30 feet back because there's many,many shelves, you know, columns of shelves
with fertilizer. I walked back and I turnedon the vision and I say, okay, there's all
(47:39):
the fertilizers. And I'm moving my phone acrossall these shelves. say, do you see tonal here?
And it says, yes, look for the one in the redand white bag. And I see it on the shelf.
So I walk straight forward. see a red and whitebag. That's not tonal. said, this isn't it.
And she, cause it's a woman's voice that I have,she says, it's two shelves to the left, second
(48:04):
from the top. I walk over there and it's rightwhere she said it was. Crazy. And you're not
a beta user. So this is available today. Thisis available. It's been available for a couple
of months. And then My daughter-in-law askedme to get something from the pharmacy, from
CVS, another big box pharmacy store, right?And this is something I don't even know if
(48:27):
I'm in the right aisle because it's somethingI've never bought. So I ask it, I say, I'm
looking for this brand and I'm not sure ifI'm in the right aisle or not, but I'm going
to walk down the aisle and tell me if you seeit. As I'm walking down the aisle, holding
it straight forward so it can see both sides. And it says, well, Yes, I'm familiar with
the brand. You should look for it in a greenand white box. then she goes like this. Oh,
(48:51):
I see it. It's down there on the right on thebottom shelf. And I turn and I look and it's
right by my right foot. You heard it here. Thisis crazy. think it's a bit creepy. How many
times have you been looking for something ona shelf? You know, and you're like, oh, how
long, how many hours is this going to take meto spot it? Good internet connection and all
(49:14):
that. So, oh my goodness. It's creepy and it'swonderful. So same time. the same time. Yeah.
Yeah. For quality of life and even for, um,yeah. So That's a mind opening thing is all
the reason I bring that up. Excellent. Hey,let's go. Let's continue on in the founder
sandbox. I'd like to ask each of my guests to share with me. I'm all about working with
(49:40):
resilient, purpose driven and scalable companiesin the growth phase. So what does resilience
mean to you? You can either answer, you know,what's the first thing that comes out of your,
you cannot use chat, GBT. I'm not fancy. Nohands. No hands, and I don't have the voice
version going because you'd hear it. Podcastwe could do it. And we are real. We're not.
(50:04):
Yeah, we are real. We're not. So I think that's,I don't think that's a difficult question to
answer. Resilience means opportunity. So nomatter what happens, even if it seems terrible,
what opportunity does that create? Excellent.If you ask that. keep reframing everything
from that perspective, it creates resilience.Right. Thank you. What about purpose-driven?
(50:33):
Purpose-driven means having a clear long-termpath and goal and asking yourself if the
things you're doing keep you on purpose to that.
(50:56):
Scalable. What's scalable mean for you? Scalablefor me means eliminating tribal knowledge
or not eliminating it, but documenting tribalknowledge. First of all, figuring out how
(51:18):
you generate revenue and then how you expandyour ability to generate revenue, which means
growing your growing your team, growing yourcapacity and identifying the bottlenecks and
focusing all your energy on the bottlenecks.And usually the bottlenecks have to do with
tribal knowledge or with lack of workflow automation.Wow, you know, it's easier said than done though,
(51:45):
that tribal knowledge, it is resistant, right?Oh yeah, because it's career, what's the
word I'm trying to think of? It keeps youin your job forever if you're the only one
that knows how to do the thing. Absolutely.That's for another podcast, David. My final
(52:05):
question today is, did you have fun in theFounder Sandbox? Oh, yes. I had a lot of fun.
Thanks. That's a great question too. Thank you,Brenda. Did you have fun? Did you? I had had
fun. And particularly in this last part, right?Cause we're talking about some heavy duty,
(52:26):
you know, uses of, um, agentic AI, right. Andscalable, you know, LTV, CAC and all that.
And then we get to hear these real life, youknow, kind of creepy, um, uh, uses of, um,
on our phones today with, um, with AI, whichis, which is quite amazing. But I also know
(52:48):
that in your world of techies, your team, whichis distributed, have a lot of fun events too.
So you probably- have one more thing on thewhole scalable thing. You have to be compassionately
ruthless or ruthlessly compassionate, howeveryou want to say it. Okay. So that the people,
(53:09):
every, and the ruthless is anything that's goingto get in the way of you growing your company,
which benefits everybody in the company. itneeds to be addressed in a ruthless way. But
if you build a culture of ruthlessly compassionate,then all the people that work for you feel
that same level of ruthlessness to protect thecompany and make it grow. And you practice
(53:36):
what you preach, I suspect, at Techies. Yes.Yes. It took me a while, but if we accidentally
hire the wrong person, either because we madea mistake in the process or they faked us out
and we recognize they're not smart enough. Literally,that's usually the problem. They're not smart
enough to carry their weight. We fire them immediately.We don't try to bring them along because you
(54:01):
can't improve somebody's IQ. You can improveany other aspect, but their IQ is their IQ.
And that will be a bottleneck forever. inour team and it'll require other people to
carry that person. And it sends the wrong messageto the team that I don't value them enough
to make sure that we only surround them withpeople that are going to inspire them and help
(54:26):
them grow. Excellent. And I suspect they arenot fungible by AI, your employees, not techies.
I mean, we've gotten better and better. at notmaking those mistakes over the years. So that
doesn't typically happen. takes us, we're muchmore careful about how we hire. AI gives us
(54:50):
the ability to recruit faster, more broadly, along with workflow automation. But what
I mean by real, this is the compassionate. Oncemy team understood this, now they embody that
and they will get rid of somebody if they madea mistake. I don't have to force the issue
ever anymore because they recognize how much,important it is to protect their teams. So
(55:14):
to my listeners, if you liked this episode todaywith the CEO and founder of Techies, sign up
for the monthly release of founders, businessowners, corporate directors, and professional
service providers who provide their examplesof how they're building companies or consulting
with companies to make them more resilient,scalable, and purpose-driven. to make profits
(55:40):
for good. Signing off for today. See you nextmonth in the Founder Sandbox. Thank you.
Popular Podcasts
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
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.