February 11, 2025 • 48 mins
We have deployed AI agents in many industries, such as venture capital, insurance, education, healthcare, and sales; let us show you what we can do for you!
Schedule a demo & free proof-of-concept to discover the full potential of iClerk Agents.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Intro/Outro (00:10):
Welcome to Building the Future with your host, Kevin
Horrick. Get ready to explorethe cutting edge of innovation
and entrepreneurship. Fromstartups to industry giants,
we're bringing you thevisionaries shaping tomorrow.
With listeners in over 150countries, we're one of the
fastest growing shows worldwide.Join Kevin, a seasoned software
(00:31):
designer and tech expert, as wedive into the ideas and
inventions changing our world.
Building the future starts now.
Ad (00:42):
Building the future is brought to you by iClerk.ai. Are
you a small business owner, aprofessional, or company looking
to automate? Do you havedifficulty hiring staff to
perform mundane knowledge worklike reporting, form creation,
or information analysis?Leverage the power of AI and the
automation of agents even if youdon't have the time or
(01:04):
experience. Try us out withlittle risk and no upfront
investment.
It's time to work smarter, notharder. Visit iclerk.ai to learn
more and get started today.
Kevin Horek (01:18):
Welcome back to the show. Today, we have Tom and
Sean from I clerk. Guys, welcometo the show.
Tom Blair (01:24):
Thank you. Hi, Kevin.
Kevin Horek (01:25):
Yeah, I'm excited to have you on the show. We've
been working together for anumber of months now. Really
love the technology and theproduct. But maybe before we
dive into all that, let's get toknow each one of you a little
bit better. And, Tom, maybe yougo first, and then I'll let Sean
go.
Tom Blair (01:40):
Sure. I've had a longer journey than Sean. My AI
journey has been has spentdecades. I in the mid eighties,
I went to Eastman Kodak Company,and I was, exposed to AI at a
really early age working onsemantic image recognition using
(02:03):
neural networks. Kodak also paidfor my grad school where I did
the same thing at school's work,which is working on these neural
networks in order to build imagedatabase systems for Kodak's
photo CD program.
We ultimately built, KodakPicture Exchange, which was a
(02:24):
image database with searchcapability, pre Internet.
Fantastic. Very large databases,that you couldn't access.
Sean Feehan (02:34):
Nice.
Tom Blair (02:36):
So I spent most of my twenties at Kodak, and in grad
school. And then, once I gotout, I went to work at Sybase.
Sybase was an early relationaldatabase company. I was one of
the founding team that startedSybase Japan, Sybase Australia,
(02:57):
Sybase India, Sybase, China, andspent a number of years based in
Tokyo and then later based inHong Kong, building distributed
database systems. I then leftAsia and went, again with Sybase
to Europe where I deployed theirmobile their mobile and embedded
(03:19):
computing small footprintdatabases, embedded databases,
and different applications.
And so we rolled that out overthe course of a couple years. I
ended up at Economy in Cambridgeagain, doing the, you know, pre
transformer architectures. Evenduring the AI winter, you still
(03:42):
had a lot of technologies outthere, things like, Bayesian
inference, probability theorythat it could that could read
any language. It could dosemantic image recognition. It
could classify text.
A lot of the same character thatyou see now in with these
transformer technologies, theideas were there and the
(04:04):
applications were there. So allof that really influenced my
career and the product thatwe'll describe today, I Clerk.
Later, in the February, the midtwo thousands, I started, a
number of different, startups.So I built a weather company, at
(04:26):
Iteris, and we managed thenation's snow plows and, did
agriculture applications usingweather modeling, soil modeling,
how water and nutrients movethrough the soil, and ultimately
crop modeling so we could,optimize when to plant, when to
(04:47):
harvest, when to fertilize,things like that. Out of that
experience, we founded anothercompany called Performance
Livestock Analytics.
Also used AI to aggregatemultiple types of information,
things like, what you're feedingthem, what kind of nutrients,
what kind of pharma or,medications are going into that
(05:10):
feed, what's the weather, what'sthe cow's maintenance,
ultimately optimizing the entiresupply chain and, and care and
feeding of those crops. That waspurchased in February, '20 by
Zoetis, the large animal healthcompany. And that then led us
(05:33):
into, where we are today with ITwith, iCLERC and and what we'll
talk to you about today. So
Kevin Horek (05:40):
Sure. Sean, do you wanna give us a bit of
background in yourself?
Sean Feehan (05:43):
Yeah. For sure. So I don't know the AI pedigree
that, Tom has. So I I lean onhim for a lot of that stuff. But
my, background is mostly insoftware development and
specifically creating softwareapplications, and all of the
fads that go along with that.
So I graduated university in02/2004, the bachelor of science
(06:04):
in computing science, and thenwent to work for a consulting
company for five ish years anddid a lot of work for the
Canadian and Alberta governmentsin terms of making small
specialized applications forthem. So I made about 12 or so
applications in those fiveyears.
Tom Blair (06:21):
That's cool.
Sean Feehan (06:22):
Built them all with whatever stack has to be, you
know, whatever trendy stackthere was Yeah. Yeah. At the
time. So built a really, reallybroad base on not only how to
build software, but also how towork with people, clients, end
users to make that software, youknow, usable and and really
valuable. From there, I went andworked for the government for a
(06:44):
few years, and got my MBA at thesame time.
And that was really anexperience in learning, process
workflow and just having alonger vision than just getting
something out to the client, butactually working on stability
and longevity for applications.And then I jumped into the
startup world. So I got pulledinto a company called Drivewise,
(07:06):
which is, located in Numbington,Alberta. And we built a trucking
software application, which,ended up growing to about 200 or
so people and sold about a yearor I guess six months ago sold.
And that was a really, reallyexciting journey going from I
was one of the first, managementemployees there, and there's
(07:29):
about 10 people, and I helpedgrow it to about a 60 people
while I was there.
And so went through all of thatgrowth and, just really loved
the startup world. And, met areally interesting guy there,
Leo, who introduced me to Tom,and, here we are.
Kevin Horek (07:47):
No. That that's really cool. It's it's
interesting. I'm always curioushow people connect with their
kind of co founders and becauseyou hear from so many people,
it's really hard to actuallyfind a co founder, never mind
somebody that you can work withand all the other fun stuff that
that comes with. So I'm curious,how did you come up with the
idea for iCLERC?
Let's talk about kind of theoriginal and how it's kind of
(08:10):
evolved to what it is today, andthen we'll we'll dive a bit
deeper.
Tom Blair (08:14):
Yeah. Certainly, the the the critical aspect of
starting a company is findingthe right partners and
teammates. And that's a longjourney, and I'm thankful for
John walking into my lifebecause he's really transformed
where the company started towhere it is today. I had
(08:34):
mentioned that we I had my priortwo companies were in ag tech,
one with weather modeling andthe other with animal health
applications. And as really anintellectual exercise during
that time, we started looking atcommodity price discovery.
(08:57):
A lot of the our customers atthat time were small farmers, so
under 2,000 head of cattle orunder, you know, couple hundred
acres of cropland, and theydon't have the benefits of the
financial markets. They'reunable to hedge at those
quantities. And so they justbasically get their lunch eaten
in the commodities markets. Andthese are complex intersections
(09:20):
of different commodities. So toprice a cow, for instance, you
have, you have live cattle,grain, feeder cattle, ethanol,
fuel.
It's really the intersection offive different commodities, and
it's a difficult human problemto solve. If you look at the
charts on that over time, humansdon't do a very good job at that
(09:41):
intersection of complexcommodities. So the original
intent was to build theseautonomous agents that could
understand each individualcommodity and then interact
between them. And, and so we,followed that intellectual
exercise for a while withouteven having, customers
(10:03):
identified or a specific problemwas, like I said, more of an
intellectual exercise. And so,over time, we spoke with
literally hundreds and hundredsof customers cross sectors.
The AI evolution, the currentevolution phase, took place
during all this time, right, theadvent of LLMs and the current
(10:25):
fervor around AI. And so peopleare very receptive to the
conversations, but it's adifferent thing to actually find
applications that can be solvedwith these types of tools. And
so over the last thousand daysor so, we have developed and
talked to people and ultimatelydefined iClerk's value
(10:48):
proposition where we, don'tmodel different commodities
rather than you you model thingsthat are important for in in a
business setting. So if it's amedical doctor, it could be all
patient visits. If it's ateacher, it could be all student
homework and tests.
If it's, venture capitalsetting, it could be, generating
(11:11):
specific reports each quarter.And so it's taken a while for
the market to catch up with ourideas, the to for us to be able
to have a nomenclaturenomenclature to describe what we
do for a living, which isautomation of mundane and
repetitive tasks, being able toreduce the toil, the work that
(11:36):
nobody likes to do with thesetypes of AI agents.
Kevin Horek (11:41):
Sure. It's interesting that you you talk
about just doing a ton ofresearch and talking to people
to basically validate an idea. Ithink so many people in the
startup world think that youcome up with an idea, you launch
it, you get a million customers,you know, you get sold in six
months to a year, but it'sreally rarely like that at all,
if ever. It takes years ofgrinding and kind of iterating
(12:05):
on a product to get to, youknow, where you start getting
your first few customers. Butbut I'm curious, how did you get
your first few customers?
Because I think that's reallythe hard part.
Tom Blair (12:19):
Really word-of-mouth. It's it's it's good. There there
are some advantages to beingold. And it's like and Matt, you
know, and and I've worked allover the world. I had this
experience with databases, whichwere are horizontal
technologies.
So you can work in anything fromretail to telecommunications to
(12:41):
transportation to weather, youknow. And and this AI these AI
agent technologies have the samecharacteristics.
Sean Feehan (12:48):
Right.
Tom Blair (12:48):
So, finding finding the low hanging fruit where
there's a discernible valueproposition. There's you know,
the customer can actually seethe automation. They can see the
cost savings. It used to costthem a hundred thousand dollars
a year to perform some sort ofrepetitive work, and now you can
do it for 10,000 or $20,000. So,the the the the act of finding
(13:17):
those solutions is hard becausewe're AI experts.
We are not subject matterexperts. We're not lawyers, then
we're not professionals. We'renot doctors. We're you know? And
so it's a partnership, and youhave to be able to listen.
The the first our firstcustomers were, literally
through relationships that wehad. I think Sean can explain
(13:41):
some of them in our ed in oureducation sector, but they're,
you know, friends or friends offriends or or two degrees of
separation or three degrees ofseparation. Soon, we'll be
meeting with Kevin Bacon.
Sean Feehan (13:56):
Yeah. I would say a lot of our journey is learning
how to speak about what ourvalue proposition is across
sectors and helping peopleunderstand, what value we're
bringing to their to theirorganization without actually
knowing how to solve theirspecific problem. We kind of ran
at a startup backwards. We had areally good technology, a really
good approach, but not an actualuse case for that technology.
(14:21):
And so to be able to communicateto somebody what we are able to
do for them and reach out andthen partner with them to to
figure out what their problem isand how we can help solve it and
identify problems that we'll dowell at.
I think that's probably whattook us the most time in terms
of how we approach, the market.And so some of the, a lot of
(14:44):
what we do, I think Tom said itearlier, is that we help take
away those mundane, repetitivetasks, those things that you
don't wanna do at work but haveto do. Yeah. Those are the
things that we attack andapproach, with partnerships to
help figure out exactly whatpeople need, and then providing
and and showing proof that we'reactually doing it and, we're
(15:06):
actually solving that problem.And then it frees them up from
half of their workday, sometimeseven three quarters of of a
workday, to be able to focus onthose, like, way more higher
value solutions or or a highervalue spent time.
Kevin Horek (15:20):
No. That that makes a lot of sense. So can you give
some examples or some advicearound actually showing that
value and getting them from justhaving a quick call to, you
know, converting to a customer?Because I think that's the real
hard part, right, is like,showing them that value,
especially early on.
Tom Blair (15:40):
One of the unique aspects of this these AI agent
technologies is that they don'tnecessarily replace existing
software systems. It's it's moreof a human labor automation. And
as such, it has a very low,footprint in an organization.
You can you can implement thesetechnologies and and achieve the
(16:04):
benefits of these technologieswithout doing any physical
integration or modifying yourexisting systems. It's really
working with the client tounderstand what they do every
day and then automating thatprocess.
It allows the customer then tofocus on the higher level
(16:24):
things, the analytics, ratherthan the groundwork of the toil.
Kevin Horek (16:29):
Sure. Can you give us some maybe examples of tasks
that we've automated goingforward? Because I think that's
where people will really graspkind of how people can leverage
these agents.
Tom Blair (16:42):
Sean, you wanna
Sean Feehan (16:43):
Yeah. Sure. Take a I'll take a run at that. So,
like Tom said, we're kinda crossindustry. We're we're taking
away that grinding work.
So a really good example is our,venture capital LPR agent, our
reporting agent. So essentially,for a venture capital fund, they
invest in a whole bunch ofcompanies. And quarterly,
(17:04):
typically, the cadence isquarterly. They have to report
out to all of their LPs,although the limited partners.
Here's how the fund's doing.
Here's what the companies aredoing. Here's the company's
financials at a high level. Allof those kind of things.
Traditionally, that's a personsending 25, 30 emails depending
on how many companies you'reyou're you're working with and
(17:26):
are invested in the fund,possibly more. And then getting
in all of the spreadsheets andinformation, all of the back and
forths, and then taking that infor each company, producing a
one page report.
So what we've done is, we'veautomated that full process. So
all you have to do is sign acompany up, and they can enter
in a survey. And the survey canbe one question or 50 questions
(17:49):
depending on what you requirefor your fund, and then upload
your financials. And that'sreally where the magic comes in
and what iClark can do, becauseas a startup, I don't have time
to produce financials in 20different formats for my 20
different investors. That's justasking a lot for, somebody who's
trying to get a business off theground and spend money wisely.
(18:12):
And so what we do is we take allsorts of different formats from
spreadsheets, PDFs, pictures ofspreadsheets from iPhones. You
know, we can even, even if it'sa recorded meeting, we can take
all of that information and turnit into a one page report on
here's the state of thisparticular company, in the fund.
(18:33):
And then from there, once wehave all of the company
information collated, we canthen work with those fund
managers to identify things thatare going on across their fund.
Kevin Horek (18:43):
Yeah. No. Well, and you're saving people tens of
hours, if not hundreds of hours,just doing that stuff. Like,
collecting all that stuff's anightmare. Nobody wants to do
that.
They wanna be, obviously,investing in new companies or
finding new opportunities oretcetera. Right? Like, they
wanna work on the fun stuff, notall the grunt work of chasing
down documents and people andpaperwork and etcetera,
etcetera.
Sean Feehan (19:03):
Absolutely. And for the for the example I gave, it's
about a hundred and sixty hoursevery quarter, which is one
month out of three where oneeight one, analyst has to gather
all of this data and producethese reports. Sure. That makes
sense.
Kevin Horek (19:18):
Do you wanna maybe give a couple other examples in
other markets of how people areleveraging agents?
Tom Blair (19:24):
Yeah. We can there's really at a high level, there's
three general categories ofagents that we found within
these businesses. One is exactlywhat Sean just talked about with
our our limited partnerreporting agent where it's a
very complex workflow. Itcontains numerical processing,
(19:45):
math, and language processing.You you you're doing your
financials and describing howthe company's doing.
You're doing that across a wideset of different companies, and
it takes time to collect theinformation that it you have
specific rules and ways ofhandling your financial data. So
(20:06):
all of that is, complex set oftasks, and then it generates the
reports and sends them to theLPs themselves. So that's a a
workflow agent, one thatemulates a complex human
workflow where they're movingdata from one system to another
and compiling it and using theirbrain and reasoning and then
(20:28):
outputting. The other two typesof agents that we've seen are
knowledge agents where youcompile a specific set of
information. One example thatwe've seen recently is for
professional servicesorganizations where over the
decades, they've done hundredsor thousands of different
professional servicesengagements, each of which has a
(20:49):
a report at the end.
And these reports are fordifferent types of customers and
sectors, but they all have somecommonality. And so this company
uses AI to build the knowledgebase of all their prior art. And
then when a new set of customerrequirements come in, they can
use that and interact andanalyze and search across their
(21:13):
entire content set using thatknowledge base. So knowledge
base is complex workflows. Andthen the third type of agent
that we see commonly isintegration agents where, a lot
of these examples that we'rethat we've used to date where
there's a human on one end and ahuman on the other.
There are many applicationswhere, it may come out of a
(21:35):
machine and go to a machine. Soan integration agent. Something
like taking, we've got oneexample where they're a a
manufacturing of lightingfixtures, and they have multiple
sites in in country and thenmultiple partners in other
countries. And so the monthly orquarterly act of going out and
(21:56):
finding out what was sold, alltheir inventory, all their
sales, all that information, andthen go, from an ERP system,
requirements planning system.And then they go and get all the
customer information from a CRMor a customer relationship
management system, and processand and analyze that information
(22:20):
to do things like, inventoryanalysis and and and and
prediction of inventory levels.
That's something that, you know,it would take a human days per
or or weeks per quarter, to movethat data, process that data,
(22:40):
and put that data back intoanother system. So that's really
the third type of agent is anintegration agent, and, with
examples of those types ofagents. Right.
Kevin Horek (22:51):
And then you're doing that lot real or real time
then too. Right? Like, it's notit's constantly happening. Like,
somebody doesn't have to do itevery week or every quarter or
every month or whatever. Like,that's happening in real time.
So you're just saving them like,you're making it even way more
efficient. Right?
Sean Feehan (23:08):
Yeah. Exactly. So, our agents, have triggers,
essentially. So that triggercould be a recorded meeting. We
have a Right.
One of our agents goes to all ofour internal meetings and
provides us meeting notes andsummaries, all of those kind of
things. And it it's trigger is arecorded meeting. So,
essentially a calendar invite,but we can we have agents that
(23:30):
are set for time periods, agentsthat are set for document
updates, agents that are set fordifferent document uploads. And
and so any of these processescan be automatically kicked off.
Tom Blair (23:43):
The the real value is is not in just the automation.
That's, you know, that'sprobably the most, obvious
benefit to these types ofsystems. It saves time. You can
measure how long it took thehuman to do, and now it doesn't.
Right?
And so but there's a number ofother benefits to these
(24:03):
technologies. I mentioned theway that it it basically acts as
an index across all of yourexisting data. And that
integration of different typesof data from voice to meetings
to text or if it's in a databaseor a CRM system or if it's a
phone call. All of that captureand processing is very expensive
(24:26):
for humans to do and very good.It's a good application for AIs.
And so that it the in theability to integrate into
existing systems, allows you to,have a very low touch
integration. Right? You don'tneed to change too much stuff.
But the real value we provide isin the guardrails. Nobody trusts
(24:47):
AI.
If the people always want to beable to see where the decision
came from or where, the datacame from. And so that ability
to emulate human reasoningrequires guardrails things that
can ensure that when datachanges you can track back what
(25:10):
was changed if you are lookingat a number in a spreadsheet,
you need to understand what wasthe source of that. So that
grounding or that traceabilityis really important for the
application of these types of AIsystems.
Kevin Horek (25:27):
Yeah. No. That that makes sense. Can you maybe talk
about where you wanna take this,or where do you see these AI
agents going? Because they'rekind of in the headlines right
now.
Sean Feehan (25:41):
You want I could talk Yeah. Briefly about, yeah,
our our road map internally, andthen I think maybe pass it over
to Tom to, sort of talk aboutgenerally what he sees
happening. Our our main road mapso right now, it's AI process
automation is what we're reallyfocused on. We're doing those
grindy, painful processes. Thereally interesting thing about
(26:04):
this is is as you're processingthat data and as your agent gets
more access to that data, youalso have the ability to chat
with your agent and askquestions about those processes.
Kevin Horek (26:13):
Which is cool. Yeah.
Sean Feehan (26:15):
It's not just the process that you're getting.
You're you're also getting fullinteraction with that data. And
so an agent essentially, has ahorizon, the data horizon, how
far it can see, what data it hasaccess to, and then it can reach
out, pull edit pull informationfrom that data and do actions on
it. Our our AI processautomation is fairly linear and,
(26:40):
you know, get the document,process the data, spit out the
results, email them out, createa PDF, those kind of things. And
we wanna do that so that,organizations don't have to, sit
and type in, like, WICHEAT GPTevery time an a document gets
updated.
And so what we're what we'rereally looking for is to move
(27:01):
away from a small c copilotwhere you're always driving to
an autopilot where your agentcan automate those processes
without you there. You just sitthere and train your agent. And
then once your agent gets all ofthis data, you can now ask it
questions. What do you knowabout this? What trends have you
noticed here?
I you know, what is my so for areally good example of that is
(27:24):
all of, like I said, all of ourorg our meetings get, recorded,
by an iClark agent. So at theend of the week, I go and ask
it, what was everybody up tothis week? And it can go in and
say, here's each person. Here'sall of their updates. Here's
what they were doing.
And all of that information isavailable. Where we're moving to
now is being able to chat withan agent about a process, and
(27:47):
then it will provide thatprocess to, after, you know,
interacting with that agent andit'll say, here's what I think
the process should be step one,step two, step three, step four,
step five. Once you'recomfortable with that process,
you then save it, assign it atrigger, and then your process
is now automated. So you can dothat without a developer at all
in the loop. Yeah, that's cool.
(28:10):
Yeah. And iClick, that's wherewe're going. So I'll just, Tom,
how about the whole industry?
Tom Blair (28:16):
So, you know, I mentioned we've we've done
hundreds or thousands ofdifferent phone conversations
across sector and globally, andwhat and that's but that act of
discovery has been has reallyformed where we're bringing our
company. The the the the thefirst thing they teach you when
(28:37):
you're starting a company isthat you have to find a specific
problem and focus. It's focus isthe operative word. And that's
almost counter to where I foundsuccess in my career. Early, it
was with digital imaging, whichis really a horizontal
technology and databases, whichis a really horizontal
(28:59):
technology.
And we see the samecharacteristics in AI. So, what
we've built here is somethingthat goes against that type of
focus wisdom, and we built across sector, cross border
platform because people usemodels for different problems.
They use, you know, you you'lluse OCR models or language
(29:24):
models or math models or, youknow, any combination of models
based on cost or performance orjurisdiction where you're
operating in the world. And sowe built a solution that really
matches that how businesses haveevolved. Nobody is just on
Microsoft or just on Yeah.
You know, on Google or, youknow, it's a heterogeneous
(29:47):
environment. And we've builtsomething that adapts to the
requirements of those thosecompanies. And I think that's
what you'll see is for this totruly be adopted cross sector,
cross border globally and andaffect all workflows like like
everyone's predicting, then youneed these types of systems that
(30:09):
can, bridge that gap betweencurrent manually operated SAS
systems and these fullyautomated systems that everyone
is envisioning for the future.So we're in this transitory
phase where, you know, it it ittakes what we have and converts
it. And and so over the nextcouple years, five years, say,
(30:30):
for instance, all workflows willbenefit from these type of AI
processes.
Sean Feehan (30:38):
I think you said something really interesting
there in the middle, Tom, justabout how we're at iCleark, we
don't ask people to adopt awhole new platform or learn a
whole new platform or get ridof, you know, their current
workflow. So we plug into that,and we allow businesses to just
supercharge their currentworkflows as opposed to asking
them for a huge processmanagement and change management
(31:01):
Right. Sort of exercise.
Tom Blair (31:02):
And to the cost of the humans. Right? That's about
what these things cost to bringin the market. So no integration
cost. Usually, you can implementthese things in a hundred man
hours.
So, you know, give or take amonth of time, it doesn't affect
your existing infrastructure. Sothere's very little training
that's required for this ordowntime required for
(31:24):
integration. And, you know, it'sit's of great benefit. It saves
time immediately, and it's it'sit, you know, it it it, shows
you the great benefit that itbrings, almost immediately.
Kevin Horek (31:39):
The other thing that you mentioned that I
thought was interesting that Iwanna get dive a little bit
deeper on is if you obviously,if you read all the startup
books or whatever you feedonline, it's you basically build
a product and then for a marketand try to get everybody into it
as quick as possible. But theinteresting thing about you
you've done it kind ofbackwards. How has that worked
(31:59):
for you? And give us some adviceon kind of doing that. Because
when you're probably gettingadvice building a company,
people are like, you gotta stopdoing it that way.
Tom Blair (32:10):
Yeah. We're a team of hockey players. I I tell you
that, and we forecheck a lot.So, gee, our mantra is shots on
net. And at this stage of thegame where the the industry has
been searching for solutions forthis technology, that's been
where we've where we've been atfor the last twenty four, thirty
(32:30):
six months.
Okay. There's really a lot ofchange that's going on right
now. People are you you have tosell less. People are actually
there's a lot lot of demand.This is demand generated rather
than, you know, intentgenerated.
And so that's the transitionthat's taken place. The vertical
AI, which addresses AI specificfor different sectors, is where
(32:54):
we found found ourselves and andand, you know, and building
these types of specific AI forspecific solutions is where we
finally found ourselves. Right?So I think you end up there. But
but it you know, there's there'snothing against building to an
idea rather than solving aproblem.
(33:16):
And and I've I've proven thatthroughout my career. So, it's
not for everyone, but you fishyyou know, it this type of
technology calls for that typeof breadth.
Sean Feehan (33:29):
Anything to add to that? Yeah. Go ahead. Yeah. I
was just gonna say it'sdefinitely a longer road, and,
have and longer costs more.
But if you have a good enoughidea and a good enough approach,
I think you'll eventually getthere. And, one of the key
things that we're doing is weare building a aging operating
system, and that is all of ourtechnology. But when we go out
(33:52):
to companies, especially afterwe've found a really solid
partner in a vertical, then westart to understand that
vertical a little more, start tounderstand a little bit more of
that vertical's need, and thenapply that same agent over and
over and over again in thatvertical. And that's where we've
been finding a lot of ourtraction and success is through
really valuable first touchpartners and then their networks
(34:17):
and and word-of-mouth from thereto be able to repeat that agent,
and, address a very specificproblem in that particular
vertical.
Kevin Horek (34:28):
It's also interesting because then you're
not pigeonholing yourself intoone vertical, right, where your
startup becomes way morevaluable because you're not just
pigeonholed to one industry. Andif something happens, like, I
don't know, like Apple releasessomething in that vertical
that's identical and gives itaway for free, hypothetically,
(34:49):
it doesn't wreck the wholecompany. Right? Like, you've
you've seen that kind of happen.You're like future proofing the
company.
Right? Do you agree or thoughtsaround that?
Sean Feehan (34:59):
Yeah. I like to think of our company as, three
small companies or 10 smallcompanies in one giant trench
coat. Okay. So all stacked ontop of each other.
Tom Blair (35:08):
Well, I think we are focused. We're just focused on
AI, and that's why partners areso important is because all that
subject matter expertise,whether you're a lawyer or a
doctor or a teacher or anaccountant, they don't expect us
to know their business. Right?Right. And so that that
interaction and thatcommunication is essential.
(35:30):
Right? Because, you know, whatwe're trying to do is the power
of AI lies it's in its abilityto make us more human. Right?
And allowing us to focus on whatmatters. So if you can take out
the toil of the actual LPreporting and focus on the
analysis of your portfolio,right, that that that's
(35:52):
beneficial to us.
Right? It's not replacing yourjob. It's augmenting what's
going on. Right? And and so I Ireally think that these types of
systems will affect virtuallyevery industry out there.
There's a huge opportunity.
Kevin Horek (36:08):
Yeah. Interesting. So you've both been through
building multiple start ups.What advice do you give to
people either before they'restarting to actually make the
leap? And what advice do yougive to people that are, you
know, kind of in the trenchesbuilding these things?
Tom Blair (36:26):
I you have to swallow your pride and dial for dollars.
I it's it's a hard thing to do,but I have called everyone I've
met in the last thirty fiveyears. I I just it's it's you
have have to be obsessive, andand and persist. This startup
we're working on right now, theoriginal ideas I wrote down in
(36:48):
in our early patents seven yearsago. Uh-huh.
Right? And, you know, we builtfor three years on a vision, and
not everyone can afford to dothat from a time or money
perspective. It's not optimal.But, if you have a vision, the
message is you have to persistand sometimes wait for the
market to catch up to you.Right?
Kevin Horek (37:10):
Yeah. That's good advice. Sean?
Sean Feehan (37:13):
You know what? I'll I'll get some advice from the
technology side that I'vefigured out. And, for anybody
who's sort of leading thetechnology branch of these
startup companies is reallyunderstand the business,
understand the flex points inthe business and where it is
most likely to change, anddesign your application to allow
that flexibility, because itthings will change inevitably.
(37:36):
And so if you've built somethingthat is not flexible in the same
places, you're gonna run into awhole lot of pain and problems
in rebuilding. Whereas if youbuild an application with a bit
of foresight, that can change,depending on how the company is
built, it's much easier.
Kevin Horek (37:53):
Oh, I % agree. It it it's a bit mind boggling how
many companies don't do thatfrom a tech side of things. But
it's also I find like a lot ofpeople can't see that, right, or
like figure it out. And like, Iguess a simple example is what
I've seen recently was somebodysomebody implemented just Google
(38:15):
sign on and picked a technologythat's all they could implement.
And then another client came tothem and said, oh, we need to
sign in with Office three sixty.
And they had to rewrite thewhole thing. And to me, I was
like, well, that's a no brainer.But I've seen that happen time
and time again. So how do youmotivate or get your team to
think about kind of that kind ofcrude example of it across the
(38:39):
entire platform? Becausesometimes it adds extra work,
timelines, etcetera.
Sean Feehan (38:46):
Yeah. That's that's a fine touch, and that's
learning a lot of hard lessonsto get there, to be perfectly
honest. But I think the mostimportant part is that your
development team needs to befully invested in the business.
They need to understand what'sgoing on. They need to
understand what's gonna happenor at at at best how you can
kind of how you're foreseeingthe future and some of the
(39:08):
possible changes.
And the more you can get yourtechnology team to understand
the business, the better yourplatform will be.
Kevin Horek (39:15):
And and so how have you in the past worked with the
entire team to understand thebusiness? Because, like, is it
meetings? Like, what is thewhat's some advice to actually
make that happen?
Sean Feehan (39:27):
Yeah. I think Tom's really good at that, to be
honest. We do a weekly all handsmeeting where Tom's just
explaining here's some of thethings we're thinking about,
some of the things that go onour industry. Here's clients in
who we're working with. And alot of that just sort of sharing
that information, and andexposing everyone in the
organization to that informationis really valuable.
Tom Blair (39:49):
That's essential when you run a global distributed
team where we're inherently disdistributed. I'm in Southern
California, Sean's in Edmonton,and people are spread, you know,
from Asia to Europe on our teameven though we only have a dozen
people. And so thatcommunication and and the
camaraderie and the communitythat you build is as essential
(40:11):
as with any other business. Andit's taken us a while to do
that. You know, building a teamis an evolution, especially as
you're going through potentiallybrand changes or market changes
and all the things that you, youknow, the transitions that you
go through as a startup.
Right? And and so if you have acore set of people that believe
(40:33):
in the overall vision, then,midstream changes don't really
affect the team as much. It'sthat it's really important to
have communication.
Kevin Horek (40:43):
Makes sense. Advice for managing a remote team
because a lot of the bigcompanies are making everybody
go back to an office, and a lotof people aren't happy about
that.
Tom Blair (40:55):
I I have a a strong philosophy about this. I, you
know, I've worked in 38countries, and and, I hire the
type of of people that areindependent. They're smart.
There's not a lot of risk whenyou hire somebody that is a
world traveler and educated anda specialist in what you're
looking for. So it's right.
(41:17):
I will see where we go as wegrow. I've never had a fully,
fully remote large company.
Sean Feehan (41:26):
Okay.
Tom Blair (41:27):
We'll we'll see where this goes. Right now, I'm I'm
pretty confident that, our worldheadquarters will remain in our
houses.
Kevin Horek (41:35):
Fair enough. Sean, you were gonna add something?
Sean Feehan (41:38):
Nope. I think talk good with that. Okay.
Kevin Horek (41:40):
No. I think yeah. It's it's interesting. It like,
I think the weird thing I'vealways found is somebody that's
a creative person is it's likeyou want me to be creative nine
to five, Monday to Friday in anoffice. And then if you put in a
dress code where I have to belike, wear something I'm not
comfortable in, it's like, okay,you're like shooting yourself in
the foot in my opinion anyway.
And so it's just an interestingI never got it until, like,
(42:01):
COVID hit, and then everybodyseemed to get it or had to get
it.
Tom Blair (42:04):
But we take it a step farther. We try to hire,
engineers in the North. So ifyou're in, like, Canada or
Norway or, you know, Georgia orwherever, because it's dark most
of the year and And cold. Gotnothing else to do. So Yeah.
Kevin Horek (42:22):
When you can't go outside, it's the you're just
totally
Tom Blair (42:25):
remote work. It's remote work from the tundra.
Kevin Horek (42:28):
That's right. There you go. That's that's the word
to live by there. Any any otherkind of thoughts around kind of
AI right now? There's a lot ofhype around it.
There's a lot of, like,overpromise, underdeliver stuff.
Like, what are your guys'thoughts on kind of where we're
actually at with it? Because Ithink there's so much
(42:48):
misinformation out there, andthere's a lot of the
misinformation is coming fromthe people actually building it,
and it's driving me absolutelycrazy.
Tom Blair (42:58):
I have a rough time doing this. I got my degree in
AI in the eighties. And, youknow, what's happened since then
was the Internet and 1,000,000times computational improvement.
And so Okay. That was evensitting there at Kodak and
knowing about Moore's law and,you know, trying to solve the
(43:18):
problem in real time, it's I Ican't even imagine what thirty
years from now will look like.
Sean Feehan (43:25):
Makes a sense.
Tom Blair (43:25):
And so we really do have a scope of five years
ahead, you know, two yearsahead, and even that gets very
foggy. What we do know is thatthere's a huge market out there
that needs to take advantage ofthese technologies and and but
but yet are unable to do sobecause of capital constraints
or human resource constraints.And that's where we come in.
Sean Feehan (43:48):
No. That makes sense.
Kevin Horek (43:49):
Sean, anything to add to that?
Sean Feehan (43:52):
Yeah, I think part of what we've found success in
is the show, don't tellmethodology. So to actually
demonstrate here's exactly howwe'll help you solve your
problems and get people, getpeople's feedback on it. Let
them go through the process, andreally understand not just the
inputs and outputs, but how thewhole process works together.
(44:14):
And once we're able to showthat's when you get belief. And,
and so that's our entireapproach with our partners is is
to show them here's exactly howit works.
And and the more you candemonstrate, the more there is
belief that it will actually dothe things that are promised.
Kevin Horek (44:30):
Yeah. No. I think that's really good. Do you wanna
talk a bit about the businessmodel? Because you have a
different approach than, Ithink, traditional companies
ever mind the startup take.
Tom Blair (44:40):
Sean, do you wanna go over that?
Sean Feehan (44:43):
Yeah. Like I said earlier, we're building
technology wise, we're buildinga platform. But what we're doing
is we're working with partnersto solve actual tangible
problems that they're dealingwith. So we're kind of a weird
hybrid consulting workflow andand platform automation. And
(45:03):
then, I guess not really an LLMcompany because we don't build
our own LLMs, but what we do iswe call pick best of breed.
And we, allow companies not tohave to become experts in AI or
in which LLM is, de rigueurright now. And and so part of
(45:23):
part of our methodology is to bethat consulting partner to help
them solve those reputableproblems. And then part of it is
to, essentially productize thoseagents that we build and then
move into that individualmarket. And I think the really
neat thing about this model isthat, given enough the support
(45:45):
constraints, we can go intofive, ten, or 15 different
verticals, and attack problemsthat may be smaller than, are
able to support a regular, a oneoff company in that vertical.
Yeah.
Go
Tom Blair (46:01):
ahead. What we've really found effective, in our
go to market strategy is to,alleviate people's concerns
about accuracy, privacy, andsecurity. And so the way we
prove that to them is byrunning, a thirty day pilot
(46:23):
where we work with them. Youknow, there's a human process
and they're interacting withspreadsheets or PDFs or
databases. Whatever that humanprocess is, we work with them
and effectively emulate that.
So at the end of the thirtydays, they can see what the
human cranked output is, andthey can see what the automated
agent output is. And once theysee that and see that it's
(46:45):
secure, private, accurate,traceable, explainable, then
it's much easier for them tosign up. So we do a thirty day
free proof of concept. We provethat we're gonna do it, and
then, that converts into asubscription agreement. You
typically, it's for one tenth ofthe cost of what the human, is
(47:06):
performing enough for.
So if, you know, you have a afull time professional at a
hundred thousand dollars a year,and they're performing, you
know, x, y, and z set of work,and we can automate that. You
can usually do that for 10 to$20 thousand or $2,000 a month.
Kevin Horek (47:23):
Very cool. Well, we're kinda coming to the end of
the show. So is there any otherkind of advice or takeaways that
you wanna leave the listenerwith, and then, let you guys get
on with your day?
Tom Blair (47:35):
Iclark dot ai. We're love to have conversations, and,
we're learning along witheveryone. So please contact us,
and we would love to discover
Sean Feehan (47:46):
together. Cool. Thanks, guys. K. Bye.
Thanks, Kevin. Awesome.
Intro/Outro (47:57):
For more episodes, you can find us on all major
podcasting platforms or visitIambuildingthefuture.com. Join
us next time and keep buildingthe future.
Welcome to Building the Future with your host, Kevin
Horrick. Get ready to explorethe cutting edge of innovation
and entrepreneurship. Fromstartups to industry giants,
we're bringing you thevisionaries shaping tomorrow.
With listeners in over 150countries, we're one of the
fastest growing shows worldwide.Join Kevin, a seasoned software
(00:31):
designer and tech expert, as wedive into the ideas and
inventions changing our world.
Building the future starts now.
Ad (00:42):
Building the future is brought to you by iClerk.ai. Are
you a small business owner, aprofessional, or company looking
to automate? Do you havedifficulty hiring staff to
perform mundane knowledge worklike reporting, form creation,
or information analysis?Leverage the power of AI and the
automation of agents even if youdon't have the time or
(01:04):
experience. Try us out withlittle risk and no upfront
investment.
It's time to work smarter, notharder. Visit iclerk.ai to learn
more and get started today.
Kevin Horek (01:18):
Welcome back to the show. Today, we have Tom and
Sean from I clerk. Guys, welcometo the show.
Tom Blair (01:24):
Thank you. Hi, Kevin.
Kevin Horek (01:25):
Yeah, I'm excited to have you on the show. We've
been working together for anumber of months now. Really
love the technology and theproduct. But maybe before we
dive into all that, let's get toknow each one of you a little
bit better. And, Tom, maybe yougo first, and then I'll let Sean
go.
Tom Blair (01:40):
Sure. I've had a longer journey than Sean. My AI
journey has been has spentdecades. I in the mid eighties,
I went to Eastman Kodak Company,and I was, exposed to AI at a
really early age working onsemantic image recognition using
(02:03):
neural networks. Kodak also paidfor my grad school where I did
the same thing at school's work,which is working on these neural
networks in order to build imagedatabase systems for Kodak's
photo CD program.
We ultimately built, KodakPicture Exchange, which was a
(02:24):
image database with searchcapability, pre Internet.
Fantastic. Very large databases,that you couldn't access.
Sean Feehan (02:34):
Nice.
Tom Blair (02:36):
So I spent most of my twenties at Kodak, and in grad
school. And then, once I gotout, I went to work at Sybase.
Sybase was an early relationaldatabase company. I was one of
the founding team that startedSybase Japan, Sybase Australia,
(02:57):
Sybase India, Sybase, China, andspent a number of years based in
Tokyo and then later based inHong Kong, building distributed
database systems. I then leftAsia and went, again with Sybase
to Europe where I deployed theirmobile their mobile and embedded
(03:19):
computing small footprintdatabases, embedded databases,
and different applications.
And so we rolled that out overthe course of a couple years. I
ended up at Economy in Cambridgeagain, doing the, you know, pre
transformer architectures. Evenduring the AI winter, you still
(03:42):
had a lot of technologies outthere, things like, Bayesian
inference, probability theorythat it could that could read
any language. It could dosemantic image recognition. It
could classify text.
A lot of the same character thatyou see now in with these
transformer technologies, theideas were there and the
(04:04):
applications were there. So allof that really influenced my
career and the product thatwe'll describe today, I Clerk.
Later, in the February, the midtwo thousands, I started, a
number of different, startups.So I built a weather company, at
(04:26):
Iteris, and we managed thenation's snow plows and, did
agriculture applications usingweather modeling, soil modeling,
how water and nutrients movethrough the soil, and ultimately
crop modeling so we could,optimize when to plant, when to
(04:47):
harvest, when to fertilize,things like that. Out of that
experience, we founded anothercompany called Performance
Livestock Analytics.
Also used AI to aggregatemultiple types of information,
things like, what you're feedingthem, what kind of nutrients,
what kind of pharma or,medications are going into that
(05:10):
feed, what's the weather, what'sthe cow's maintenance,
ultimately optimizing the entiresupply chain and, and care and
feeding of those crops. That waspurchased in February, '20 by
Zoetis, the large animal healthcompany. And that then led us
(05:33):
into, where we are today with ITwith, iCLERC and and what we'll
talk to you about today. So
Kevin Horek (05:40):
Sure. Sean, do you wanna give us a bit of
background in yourself?
Sean Feehan (05:43):
Yeah. For sure. So I don't know the AI pedigree
that, Tom has. So I I lean onhim for a lot of that stuff. But
my, background is mostly insoftware development and
specifically creating softwareapplications, and all of the
fads that go along with that.
So I graduated university in02/2004, the bachelor of science
(06:04):
in computing science, and thenwent to work for a consulting
company for five ish years anddid a lot of work for the
Canadian and Alberta governmentsin terms of making small
specialized applications forthem. So I made about 12 or so
applications in those fiveyears.
Tom Blair (06:21):
That's cool.
Sean Feehan (06:22):
Built them all with whatever stack has to be, you
know, whatever trendy stackthere was Yeah. Yeah. At the
time. So built a really, reallybroad base on not only how to
build software, but also how towork with people, clients, end
users to make that software, youknow, usable and and really
valuable. From there, I went andworked for the government for a
(06:44):
few years, and got my MBA at thesame time.
And that was really anexperience in learning, process
workflow and just having alonger vision than just getting
something out to the client, butactually working on stability
and longevity for applications.And then I jumped into the
startup world. So I got pulledinto a company called Drivewise,
(07:06):
which is, located in Numbington,Alberta. And we built a trucking
software application, which,ended up growing to about 200 or
so people and sold about a yearor I guess six months ago sold.
And that was a really, reallyexciting journey going from I
was one of the first, managementemployees there, and there's
(07:29):
about 10 people, and I helpedgrow it to about a 60 people
while I was there.
And so went through all of thatgrowth and, just really loved
the startup world. And, met areally interesting guy there,
Leo, who introduced me to Tom,and, here we are.
Kevin Horek (07:47):
No. That that's really cool. It's it's
interesting. I'm always curioushow people connect with their
kind of co founders and becauseyou hear from so many people,
it's really hard to actuallyfind a co founder, never mind
somebody that you can work withand all the other fun stuff that
that comes with. So I'm curious,how did you come up with the
idea for iCLERC?
Let's talk about kind of theoriginal and how it's kind of
(08:10):
evolved to what it is today, andthen we'll we'll dive a bit
deeper.
Tom Blair (08:14):
Yeah. Certainly, the the the critical aspect of
starting a company is findingthe right partners and
teammates. And that's a longjourney, and I'm thankful for
John walking into my lifebecause he's really transformed
where the company started towhere it is today. I had
(08:34):
mentioned that we I had my priortwo companies were in ag tech,
one with weather modeling andthe other with animal health
applications. And as really anintellectual exercise during
that time, we started looking atcommodity price discovery.
(08:57):
A lot of the our customers atthat time were small farmers, so
under 2,000 head of cattle orunder, you know, couple hundred
acres of cropland, and theydon't have the benefits of the
financial markets. They'reunable to hedge at those
quantities. And so they justbasically get their lunch eaten
in the commodities markets. Andthese are complex intersections
(09:20):
of different commodities. So toprice a cow, for instance, you
have, you have live cattle,grain, feeder cattle, ethanol,
fuel.
It's really the intersection offive different commodities, and
it's a difficult human problemto solve. If you look at the
charts on that over time, humansdon't do a very good job at that
(09:41):
intersection of complexcommodities. So the original
intent was to build theseautonomous agents that could
understand each individualcommodity and then interact
between them. And, and so we,followed that intellectual
exercise for a while withouteven having, customers
(10:03):
identified or a specific problemwas, like I said, more of an
intellectual exercise. And so,over time, we spoke with
literally hundreds and hundredsof customers cross sectors.
The AI evolution, the currentevolution phase, took place
during all this time, right, theadvent of LLMs and the current
(10:25):
fervor around AI. And so peopleare very receptive to the
conversations, but it's adifferent thing to actually find
applications that can be solvedwith these types of tools. And
so over the last thousand daysor so, we have developed and
talked to people and ultimatelydefined iClerk's value
(10:48):
proposition where we, don'tmodel different commodities
rather than you you model thingsthat are important for in in a
business setting. So if it's amedical doctor, it could be all
patient visits. If it's ateacher, it could be all student
homework and tests.
If it's, venture capitalsetting, it could be, generating
(11:11):
specific reports each quarter.And so it's taken a while for
the market to catch up with ourideas, the to for us to be able
to have a nomenclaturenomenclature to describe what we
do for a living, which isautomation of mundane and
repetitive tasks, being able toreduce the toil, the work that
(11:36):
nobody likes to do with thesetypes of AI agents.
Kevin Horek (11:41):
Sure. It's interesting that you you talk
about just doing a ton ofresearch and talking to people
to basically validate an idea. Ithink so many people in the
startup world think that youcome up with an idea, you launch
it, you get a million customers,you know, you get sold in six
months to a year, but it'sreally rarely like that at all,
if ever. It takes years ofgrinding and kind of iterating
(12:05):
on a product to get to, youknow, where you start getting
your first few customers. Butbut I'm curious, how did you get
your first few customers?
Because I think that's reallythe hard part.
Tom Blair (12:19):
Really word-of-mouth. It's it's it's good. There there
are some advantages to beingold. And it's like and Matt, you
know, and and I've worked allover the world. I had this
experience with databases, whichwere are horizontal
technologies.
So you can work in anything fromretail to telecommunications to
(12:41):
transportation to weather, youknow. And and this AI these AI
agent technologies have the samecharacteristics.
Sean Feehan (12:48):
Right.
Tom Blair (12:48):
So, finding finding the low hanging fruit where
there's a discernible valueproposition. There's you know,
the customer can actually seethe automation. They can see the
cost savings. It used to costthem a hundred thousand dollars
a year to perform some sort ofrepetitive work, and now you can
do it for 10,000 or $20,000. So,the the the the act of finding
(13:17):
those solutions is hard becausewe're AI experts.
We are not subject matterexperts. We're not lawyers, then
we're not professionals. We'renot doctors. We're you know? And
so it's a partnership, and youhave to be able to listen.
The the first our firstcustomers were, literally
through relationships that wehad. I think Sean can explain
(13:41):
some of them in our ed in oureducation sector, but they're,
you know, friends or friends offriends or or two degrees of
separation or three degrees ofseparation. Soon, we'll be
meeting with Kevin Bacon.
Sean Feehan (13:56):
Yeah. I would say a lot of our journey is learning
how to speak about what ourvalue proposition is across
sectors and helping peopleunderstand, what value we're
bringing to their to theirorganization without actually
knowing how to solve theirspecific problem. We kind of ran
at a startup backwards. We had areally good technology, a really
good approach, but not an actualuse case for that technology.
(14:21):
And so to be able to communicateto somebody what we are able to
do for them and reach out andthen partner with them to to
figure out what their problem isand how we can help solve it and
identify problems that we'll dowell at.
I think that's probably whattook us the most time in terms
of how we approach, the market.And so some of the, a lot of
(14:44):
what we do, I think Tom said itearlier, is that we help take
away those mundane, repetitivetasks, those things that you
don't wanna do at work but haveto do. Yeah. Those are the
things that we attack andapproach, with partnerships to
help figure out exactly whatpeople need, and then providing
and and showing proof that we'reactually doing it and, we're
(15:06):
actually solving that problem.And then it frees them up from
half of their workday, sometimeseven three quarters of of a
workday, to be able to focus onthose, like, way more higher
value solutions or or a highervalue spent time.
Kevin Horek (15:20):
No. That that makes a lot of sense. So can you give
some examples or some advicearound actually showing that
value and getting them from justhaving a quick call to, you
know, converting to a customer?Because I think that's the real
hard part, right, is like,showing them that value,
especially early on.
Tom Blair (15:40):
One of the unique aspects of this these AI agent
technologies is that they don'tnecessarily replace existing
software systems. It's it's moreof a human labor automation. And
as such, it has a very low,footprint in an organization.
You can you can implement thesetechnologies and and achieve the
(16:04):
benefits of these technologieswithout doing any physical
integration or modifying yourexisting systems. It's really
working with the client tounderstand what they do every
day and then automating thatprocess.
It allows the customer then tofocus on the higher level
(16:24):
things, the analytics, ratherthan the groundwork of the toil.
Kevin Horek (16:29):
Sure. Can you give us some maybe examples of tasks
that we've automated goingforward? Because I think that's
where people will really graspkind of how people can leverage
these agents.
Tom Blair (16:42):
Sean, you wanna
Sean Feehan (16:43):
Yeah. Sure. Take a I'll take a run at that. So,
like Tom said, we're kinda crossindustry. We're we're taking
away that grinding work.
So a really good example is our,venture capital LPR agent, our
reporting agent. So essentially,for a venture capital fund, they
invest in a whole bunch ofcompanies. And quarterly,
(17:04):
typically, the cadence isquarterly. They have to report
out to all of their LPs,although the limited partners.
Here's how the fund's doing.
Here's what the companies aredoing. Here's the company's
financials at a high level. Allof those kind of things.
Traditionally, that's a personsending 25, 30 emails depending
on how many companies you'reyou're you're working with and
(17:26):
are invested in the fund,possibly more. And then getting
in all of the spreadsheets andinformation, all of the back and
forths, and then taking that infor each company, producing a
one page report.
So what we've done is, we'veautomated that full process. So
all you have to do is sign acompany up, and they can enter
in a survey. And the survey canbe one question or 50 questions
(17:49):
depending on what you requirefor your fund, and then upload
your financials. And that'sreally where the magic comes in
and what iClark can do, becauseas a startup, I don't have time
to produce financials in 20different formats for my 20
different investors. That's justasking a lot for, somebody who's
trying to get a business off theground and spend money wisely.
(18:12):
And so what we do is we take allsorts of different formats from
spreadsheets, PDFs, pictures ofspreadsheets from iPhones. You
know, we can even, even if it'sa recorded meeting, we can take
all of that information and turnit into a one page report on
here's the state of thisparticular company, in the fund.
(18:33):
And then from there, once wehave all of the company
information collated, we canthen work with those fund
managers to identify things thatare going on across their fund.
Kevin Horek (18:43):
Yeah. No. Well, and you're saving people tens of
hours, if not hundreds of hours,just doing that stuff. Like,
collecting all that stuff's anightmare. Nobody wants to do
that.
They wanna be, obviously,investing in new companies or
finding new opportunities oretcetera. Right? Like, they
wanna work on the fun stuff, notall the grunt work of chasing
down documents and people andpaperwork and etcetera,
etcetera.
Sean Feehan (19:03):
Absolutely. And for the for the example I gave, it's
about a hundred and sixty hoursevery quarter, which is one
month out of three where oneeight one, analyst has to gather
all of this data and producethese reports. Sure. That makes
sense.
Kevin Horek (19:18):
Do you wanna maybe give a couple other examples in
other markets of how people areleveraging agents?
Tom Blair (19:24):
Yeah. We can there's really at a high level, there's
three general categories ofagents that we found within
these businesses. One is exactlywhat Sean just talked about with
our our limited partnerreporting agent where it's a
very complex workflow. Itcontains numerical processing,
(19:45):
math, and language processing.You you you're doing your
financials and describing howthe company's doing.
You're doing that across a wideset of different companies, and
it takes time to collect theinformation that it you have
specific rules and ways ofhandling your financial data. So
(20:06):
all of that is, complex set oftasks, and then it generates the
reports and sends them to theLPs themselves. So that's a a
workflow agent, one thatemulates a complex human
workflow where they're movingdata from one system to another
and compiling it and using theirbrain and reasoning and then
(20:28):
outputting. The other two typesof agents that we've seen are
knowledge agents where youcompile a specific set of
information. One example thatwe've seen recently is for
professional servicesorganizations where over the
decades, they've done hundredsor thousands of different
professional servicesengagements, each of which has a
(20:49):
a report at the end.
And these reports are fordifferent types of customers and
sectors, but they all have somecommonality. And so this company
uses AI to build the knowledgebase of all their prior art. And
then when a new set of customerrequirements come in, they can
use that and interact andanalyze and search across their
(21:13):
entire content set using thatknowledge base. So knowledge
base is complex workflows. Andthen the third type of agent
that we see commonly isintegration agents where, a lot
of these examples that we'rethat we've used to date where
there's a human on one end and ahuman on the other.
There are many applicationswhere, it may come out of a
(21:35):
machine and go to a machine. Soan integration agent. Something
like taking, we've got oneexample where they're a a
manufacturing of lightingfixtures, and they have multiple
sites in in country and thenmultiple partners in other
countries. And so the monthly orquarterly act of going out and
(21:56):
finding out what was sold, alltheir inventory, all their
sales, all that information, andthen go, from an ERP system,
requirements planning system.And then they go and get all the
customer information from a CRMor a customer relationship
management system, and processand and analyze that information
(22:20):
to do things like, inventoryanalysis and and and and
prediction of inventory levels.
That's something that, you know,it would take a human days per
or or weeks per quarter, to movethat data, process that data,
(22:40):
and put that data back intoanother system. So that's really
the third type of agent is anintegration agent, and, with
examples of those types ofagents. Right.
Kevin Horek (22:51):
And then you're doing that lot real or real time
then too. Right? Like, it's notit's constantly happening. Like,
somebody doesn't have to do itevery week or every quarter or
every month or whatever. Like,that's happening in real time.
So you're just saving them like,you're making it even way more
efficient. Right?
Sean Feehan (23:08):
Yeah. Exactly. So, our agents, have triggers,
essentially. So that triggercould be a recorded meeting. We
have a Right.
One of our agents goes to all ofour internal meetings and
provides us meeting notes andsummaries, all of those kind of
things. And it it's trigger is arecorded meeting. So,
essentially a calendar invite,but we can we have agents that
(23:30):
are set for time periods, agentsthat are set for document
updates, agents that are set fordifferent document uploads. And
and so any of these processescan be automatically kicked off.
Tom Blair (23:43):
The the real value is is not in just the automation.
That's, you know, that'sprobably the most, obvious
benefit to these types ofsystems. It saves time. You can
measure how long it took thehuman to do, and now it doesn't.
Right?
And so but there's a number ofother benefits to these
(24:03):
technologies. I mentioned theway that it it basically acts as
an index across all of yourexisting data. And that
integration of different typesof data from voice to meetings
to text or if it's in a databaseor a CRM system or if it's a
phone call. All of that captureand processing is very expensive
(24:26):
for humans to do and very good.It's a good application for AIs.
And so that it the in theability to integrate into
existing systems, allows you to,have a very low touch
integration. Right? You don'tneed to change too much stuff.
But the real value we provide isin the guardrails. Nobody trusts
(24:47):
AI.
If the people always want to beable to see where the decision
came from or where, the datacame from. And so that ability
to emulate human reasoningrequires guardrails things that
can ensure that when datachanges you can track back what
(25:10):
was changed if you are lookingat a number in a spreadsheet,
you need to understand what wasthe source of that. So that
grounding or that traceabilityis really important for the
application of these types of AIsystems.
Kevin Horek (25:27):
Yeah. No. That that makes sense. Can you maybe talk
about where you wanna take this,or where do you see these AI
agents going? Because they'rekind of in the headlines right
now.
Sean Feehan (25:41):
You want I could talk Yeah. Briefly about, yeah,
our our road map internally, andthen I think maybe pass it over
to Tom to, sort of talk aboutgenerally what he sees
happening. Our our main road mapso right now, it's AI process
automation is what we're reallyfocused on. We're doing those
grindy, painful processes. Thereally interesting thing about
(26:04):
this is is as you're processingthat data and as your agent gets
more access to that data, youalso have the ability to chat
with your agent and askquestions about those processes.
Kevin Horek (26:13):
Which is cool. Yeah.
Sean Feehan (26:15):
It's not just the process that you're getting.
You're you're also getting fullinteraction with that data. And
so an agent essentially, has ahorizon, the data horizon, how
far it can see, what data it hasaccess to, and then it can reach
out, pull edit pull informationfrom that data and do actions on
it. Our our AI processautomation is fairly linear and,
(26:40):
you know, get the document,process the data, spit out the
results, email them out, createa PDF, those kind of things. And
we wanna do that so that,organizations don't have to, sit
and type in, like, WICHEAT GPTevery time an a document gets
updated.
And so what we're what we'rereally looking for is to move
(27:01):
away from a small c copilotwhere you're always driving to
an autopilot where your agentcan automate those processes
without you there. You just sitthere and train your agent. And
then once your agent gets all ofthis data, you can now ask it
questions. What do you knowabout this? What trends have you
noticed here?
I you know, what is my so for areally good example of that is
(27:24):
all of, like I said, all of ourorg our meetings get, recorded,
by an iClark agent. So at theend of the week, I go and ask
it, what was everybody up tothis week? And it can go in and
say, here's each person. Here'sall of their updates. Here's
what they were doing.
And all of that information isavailable. Where we're moving to
now is being able to chat withan agent about a process, and
(27:47):
then it will provide thatprocess to, after, you know,
interacting with that agent andit'll say, here's what I think
the process should be step one,step two, step three, step four,
step five. Once you'recomfortable with that process,
you then save it, assign it atrigger, and then your process
is now automated. So you can dothat without a developer at all
in the loop. Yeah, that's cool.
(28:10):
Yeah. And iClick, that's wherewe're going. So I'll just, Tom,
how about the whole industry?
Tom Blair (28:16):
So, you know, I mentioned we've we've done
hundreds or thousands ofdifferent phone conversations
across sector and globally, andwhat and that's but that act of
discovery has been has reallyformed where we're bringing our
company. The the the the thefirst thing they teach you when
(28:37):
you're starting a company isthat you have to find a specific
problem and focus. It's focus isthe operative word. And that's
almost counter to where I foundsuccess in my career. Early, it
was with digital imaging, whichis really a horizontal
technology and databases, whichis a really horizontal
(28:59):
technology.
And we see the samecharacteristics in AI. So, what
we've built here is somethingthat goes against that type of
focus wisdom, and we built across sector, cross border
platform because people usemodels for different problems.
They use, you know, you you'lluse OCR models or language
(29:24):
models or math models or, youknow, any combination of models
based on cost or performance orjurisdiction where you're
operating in the world. And sowe built a solution that really
matches that how businesses haveevolved. Nobody is just on
Microsoft or just on Yeah.
You know, on Google or, youknow, it's a heterogeneous
(29:47):
environment. And we've builtsomething that adapts to the
requirements of those thosecompanies. And I think that's
what you'll see is for this totruly be adopted cross sector,
cross border globally and andaffect all workflows like like
everyone's predicting, then youneed these types of systems that
(30:09):
can, bridge that gap betweencurrent manually operated SAS
systems and these fullyautomated systems that everyone
is envisioning for the future.So we're in this transitory
phase where, you know, it it ittakes what we have and converts
it. And and so over the nextcouple years, five years, say,
(30:30):
for instance, all workflows willbenefit from these type of AI
processes.
Sean Feehan (30:38):
I think you said something really interesting
there in the middle, Tom, justabout how we're at iCleark, we
don't ask people to adopt awhole new platform or learn a
whole new platform or get ridof, you know, their current
workflow. So we plug into that,and we allow businesses to just
supercharge their currentworkflows as opposed to asking
them for a huge processmanagement and change management
(31:01):
Right. Sort of exercise.
Tom Blair (31:02):
And to the cost of the humans. Right? That's about
what these things cost to bringin the market. So no integration
cost. Usually, you can implementthese things in a hundred man
hours.
So, you know, give or take amonth of time, it doesn't affect
your existing infrastructure. Sothere's very little training
that's required for this ordowntime required for
(31:24):
integration. And, you know, it'sit's of great benefit. It saves
time immediately, and it's it'sit, you know, it it it, shows
you the great benefit that itbrings, almost immediately.
Kevin Horek (31:39):
The other thing that you mentioned that I
thought was interesting that Iwanna get dive a little bit
deeper on is if you obviously,if you read all the startup
books or whatever you feedonline, it's you basically build
a product and then for a marketand try to get everybody into it
as quick as possible. But theinteresting thing about you
you've done it kind ofbackwards. How has that worked
(31:59):
for you? And give us some adviceon kind of doing that. Because
when you're probably gettingadvice building a company,
people are like, you gotta stopdoing it that way.
Tom Blair (32:10):
Yeah. We're a team of hockey players. I I tell you
that, and we forecheck a lot.So, gee, our mantra is shots on
net. And at this stage of thegame where the the industry has
been searching for solutions forthis technology, that's been
where we've where we've been atfor the last twenty four, thirty
(32:30):
six months.
Okay. There's really a lot ofchange that's going on right
now. People are you you have tosell less. People are actually
there's a lot lot of demand.This is demand generated rather
than, you know, intentgenerated.
And so that's the transitionthat's taken place. The vertical
AI, which addresses AI specificfor different sectors, is where
(32:54):
we found found ourselves and andand, you know, and building
these types of specific AI forspecific solutions is where we
finally found ourselves. Right?So I think you end up there. But
but it you know, there's there'snothing against building to an
idea rather than solving aproblem.
(33:16):
And and I've I've proven thatthroughout my career. So, it's
not for everyone, but you fishyyou know, it this type of
technology calls for that typeof breadth.
Sean Feehan (33:29):
Anything to add to that? Yeah. Go ahead. Yeah. I
was just gonna say it'sdefinitely a longer road, and,
have and longer costs more.
But if you have a good enoughidea and a good enough approach,
I think you'll eventually getthere. And, one of the key
things that we're doing is weare building a aging operating
system, and that is all of ourtechnology. But when we go out
(33:52):
to companies, especially afterwe've found a really solid
partner in a vertical, then westart to understand that
vertical a little more, start tounderstand a little bit more of
that vertical's need, and thenapply that same agent over and
over and over again in thatvertical. And that's where we've
been finding a lot of ourtraction and success is through
really valuable first touchpartners and then their networks
(34:17):
and and word-of-mouth from thereto be able to repeat that agent,
and, address a very specificproblem in that particular
vertical.
Kevin Horek (34:28):
It's also interesting because then you're
not pigeonholing yourself intoone vertical, right, where your
startup becomes way morevaluable because you're not just
pigeonholed to one industry. Andif something happens, like, I
don't know, like Apple releasessomething in that vertical
that's identical and gives itaway for free, hypothetically,
(34:49):
it doesn't wreck the wholecompany. Right? Like, you've
you've seen that kind of happen.You're like future proofing the
company.
Right? Do you agree or thoughtsaround that?
Sean Feehan (34:59):
Yeah. I like to think of our company as, three
small companies or 10 smallcompanies in one giant trench
coat. Okay. So all stacked ontop of each other.
Tom Blair (35:08):
Well, I think we are focused. We're just focused on
AI, and that's why partners areso important is because all that
subject matter expertise,whether you're a lawyer or a
doctor or a teacher or anaccountant, they don't expect us
to know their business. Right?Right. And so that that
interaction and thatcommunication is essential.
(35:30):
Right? Because, you know, whatwe're trying to do is the power
of AI lies it's in its abilityto make us more human. Right?
And allowing us to focus on whatmatters. So if you can take out
the toil of the actual LPreporting and focus on the
analysis of your portfolio,right, that that that's
(35:52):
beneficial to us.
Right? It's not replacing yourjob. It's augmenting what's
going on. Right? And and so I Ireally think that these types of
systems will affect virtuallyevery industry out there.
There's a huge opportunity.
Kevin Horek (36:08):
Yeah. Interesting. So you've both been through
building multiple start ups.What advice do you give to
people either before they'restarting to actually make the
leap? And what advice do yougive to people that are, you
know, kind of in the trenchesbuilding these things?
Tom Blair (36:26):
I you have to swallow your pride and dial for dollars.
I it's it's a hard thing to do,but I have called everyone I've
met in the last thirty fiveyears. I I just it's it's you
have have to be obsessive, andand and persist. This startup
we're working on right now, theoriginal ideas I wrote down in
(36:48):
in our early patents seven yearsago. Uh-huh.
Right? And, you know, we builtfor three years on a vision, and
not everyone can afford to dothat from a time or money
perspective. It's not optimal.But, if you have a vision, the
message is you have to persistand sometimes wait for the
market to catch up to you.Right?
Kevin Horek (37:10):
Yeah. That's good advice. Sean?
Sean Feehan (37:13):
You know what? I'll I'll get some advice from the
technology side that I'vefigured out. And, for anybody
who's sort of leading thetechnology branch of these
startup companies is reallyunderstand the business,
understand the flex points inthe business and where it is
most likely to change, anddesign your application to allow
that flexibility, because itthings will change inevitably.
(37:36):
And so if you've built somethingthat is not flexible in the same
places, you're gonna run into awhole lot of pain and problems
in rebuilding. Whereas if youbuild an application with a bit
of foresight, that can change,depending on how the company is
built, it's much easier.
Kevin Horek (37:53):
Oh, I % agree. It it it's a bit mind boggling how
many companies don't do thatfrom a tech side of things. But
it's also I find like a lot ofpeople can't see that, right, or
like figure it out. And like, Iguess a simple example is what
I've seen recently was somebodysomebody implemented just Google
(38:15):
sign on and picked a technologythat's all they could implement.
And then another client came tothem and said, oh, we need to
sign in with Office three sixty.
And they had to rewrite thewhole thing. And to me, I was
like, well, that's a no brainer.But I've seen that happen time
and time again. So how do youmotivate or get your team to
think about kind of that kind ofcrude example of it across the
(38:39):
entire platform? Becausesometimes it adds extra work,
timelines, etcetera.
Sean Feehan (38:46):
Yeah. That's that's a fine touch, and that's
learning a lot of hard lessonsto get there, to be perfectly
honest. But I think the mostimportant part is that your
development team needs to befully invested in the business.
They need to understand what'sgoing on. They need to
understand what's gonna happenor at at at best how you can
kind of how you're foreseeingthe future and some of the
(39:08):
possible changes.
And the more you can get yourtechnology team to understand
the business, the better yourplatform will be.
Kevin Horek (39:15):
And and so how have you in the past worked with the
entire team to understand thebusiness? Because, like, is it
meetings? Like, what is thewhat's some advice to actually
make that happen?
Sean Feehan (39:27):
Yeah. I think Tom's really good at that, to be
honest. We do a weekly all handsmeeting where Tom's just
explaining here's some of thethings we're thinking about,
some of the things that go onour industry. Here's clients in
who we're working with. And alot of that just sort of sharing
that information, and andexposing everyone in the
organization to that informationis really valuable.
Tom Blair (39:49):
That's essential when you run a global distributed
team where we're inherently disdistributed. I'm in Southern
California, Sean's in Edmonton,and people are spread, you know,
from Asia to Europe on our teameven though we only have a dozen
people. And so thatcommunication and and the
camaraderie and the communitythat you build is as essential
(40:11):
as with any other business. Andit's taken us a while to do
that. You know, building a teamis an evolution, especially as
you're going through potentiallybrand changes or market changes
and all the things that you, youknow, the transitions that you
go through as a startup.
Right? And and so if you have acore set of people that believe
(40:33):
in the overall vision, then,midstream changes don't really
affect the team as much. It'sthat it's really important to
have communication.
Kevin Horek (40:43):
Makes sense. Advice for managing a remote team
because a lot of the bigcompanies are making everybody
go back to an office, and a lotof people aren't happy about
that.
Tom Blair (40:55):
I I have a a strong philosophy about this. I, you
know, I've worked in 38countries, and and, I hire the
type of of people that areindependent. They're smart.
There's not a lot of risk whenyou hire somebody that is a
world traveler and educated anda specialist in what you're
looking for. So it's right.
(41:17):
I will see where we go as wegrow. I've never had a fully,
fully remote large company.
Sean Feehan (41:26):
Okay.
Tom Blair (41:27):
We'll we'll see where this goes. Right now, I'm I'm
pretty confident that, our worldheadquarters will remain in our
houses.
Kevin Horek (41:35):
Fair enough. Sean, you were gonna add something?
Sean Feehan (41:38):
Nope. I think talk good with that. Okay.
Kevin Horek (41:40):
No. I think yeah. It's it's interesting. It like,
I think the weird thing I'vealways found is somebody that's
a creative person is it's likeyou want me to be creative nine
to five, Monday to Friday in anoffice. And then if you put in a
dress code where I have to belike, wear something I'm not
comfortable in, it's like, okay,you're like shooting yourself in
the foot in my opinion anyway.
And so it's just an interestingI never got it until, like,
(42:01):
COVID hit, and then everybodyseemed to get it or had to get
it.
Tom Blair (42:04):
But we take it a step farther. We try to hire,
engineers in the North. So ifyou're in, like, Canada or
Norway or, you know, Georgia orwherever, because it's dark most
of the year and And cold. Gotnothing else to do. So Yeah.
Kevin Horek (42:22):
When you can't go outside, it's the you're just
totally
Tom Blair (42:25):
remote work. It's remote work from the tundra.
Kevin Horek (42:28):
That's right. There you go. That's that's the word
to live by there. Any any otherkind of thoughts around kind of
AI right now? There's a lot ofhype around it.
There's a lot of, like,overpromise, underdeliver stuff.
Like, what are your guys'thoughts on kind of where we're
actually at with it? Because Ithink there's so much
(42:48):
misinformation out there, andthere's a lot of the
misinformation is coming fromthe people actually building it,
and it's driving me absolutelycrazy.
Tom Blair (42:58):
I have a rough time doing this. I got my degree in
AI in the eighties. And, youknow, what's happened since then
was the Internet and 1,000,000times computational improvement.
And so Okay. That was evensitting there at Kodak and
knowing about Moore's law and,you know, trying to solve the
(43:18):
problem in real time, it's I Ican't even imagine what thirty
years from now will look like.
Sean Feehan (43:25):
Makes a sense.
Tom Blair (43:25):
And so we really do have a scope of five years
ahead, you know, two yearsahead, and even that gets very
foggy. What we do know is thatthere's a huge market out there
that needs to take advantage ofthese technologies and and but
but yet are unable to do sobecause of capital constraints
or human resource constraints.And that's where we come in.
Sean Feehan (43:48):
No. That makes sense.
Kevin Horek (43:49):
Sean, anything to add to that?
Sean Feehan (43:52):
Yeah, I think part of what we've found success in
is the show, don't tellmethodology. So to actually
demonstrate here's exactly howwe'll help you solve your
problems and get people, getpeople's feedback on it. Let
them go through the process, andreally understand not just the
inputs and outputs, but how thewhole process works together.
(44:14):
And once we're able to showthat's when you get belief. And,
and so that's our entireapproach with our partners is is
to show them here's exactly howit works.
And and the more you candemonstrate, the more there is
belief that it will actually dothe things that are promised.
Kevin Horek (44:30):
Yeah. No. I think that's really good. Do you wanna
talk a bit about the businessmodel? Because you have a
different approach than, Ithink, traditional companies
ever mind the startup take.
Tom Blair (44:40):
Sean, do you wanna go over that?
Sean Feehan (44:43):
Yeah. Like I said earlier, we're building
technology wise, we're buildinga platform. But what we're doing
is we're working with partnersto solve actual tangible
problems that they're dealingwith. So we're kind of a weird
hybrid consulting workflow andand platform automation. And
(45:03):
then, I guess not really an LLMcompany because we don't build
our own LLMs, but what we do iswe call pick best of breed.
And we, allow companies not tohave to become experts in AI or
in which LLM is, de rigueurright now. And and so part of
(45:23):
part of our methodology is to bethat consulting partner to help
them solve those reputableproblems. And then part of it is
to, essentially productize thoseagents that we build and then
move into that individualmarket. And I think the really
neat thing about this model isthat, given enough the support
(45:45):
constraints, we can go intofive, ten, or 15 different
verticals, and attack problemsthat may be smaller than, are
able to support a regular, a oneoff company in that vertical.
Yeah.
Go
Tom Blair (46:01):
ahead. What we've really found effective, in our
go to market strategy is to,alleviate people's concerns
about accuracy, privacy, andsecurity. And so the way we
prove that to them is byrunning, a thirty day pilot
(46:23):
where we work with them. Youknow, there's a human process
and they're interacting withspreadsheets or PDFs or
databases. Whatever that humanprocess is, we work with them
and effectively emulate that.
So at the end of the thirtydays, they can see what the
human cranked output is, andthey can see what the automated
agent output is. And once theysee that and see that it's
(46:45):
secure, private, accurate,traceable, explainable, then
it's much easier for them tosign up. So we do a thirty day
free proof of concept. We provethat we're gonna do it, and
then, that converts into asubscription agreement. You
typically, it's for one tenth ofthe cost of what the human, is
(47:06):
performing enough for.
So if, you know, you have a afull time professional at a
hundred thousand dollars a year,and they're performing, you
know, x, y, and z set of work,and we can automate that. You
can usually do that for 10 to$20 thousand or $2,000 a month.
Kevin Horek (47:23):
Very cool. Well, we're kinda coming to the end of
the show. So is there any otherkind of advice or takeaways that
you wanna leave the listenerwith, and then, let you guys get
on with your day?
Tom Blair (47:35):
Iclark dot ai. We're love to have conversations, and,
we're learning along witheveryone. So please contact us,
and we would love to discover
Sean Feehan (47:46):
together. Cool. Thanks, guys. K. Bye.
Thanks, Kevin. Awesome.
Intro/Outro (47:57):
For more episodes, you can find us on all major
podcasting platforms or visitIambuildingthefuture.com. Join
us next time and keep buildingthe future.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com