We’re proud to kick off our four-part Modern Finance series with a conversation with Deluxe's Yogs Jayaprakasam, SVP, Chief Technology & Digital Officer.
AI isn't just changing treasury - it's completely reimagining how financial teams operate, transforming them from back-office functions into strategic business advisors.
Yogs Jayaprakasam reveals how treasury departments are leveraging artificial intelligence to solve age-old challenges in revolutionary ways. While treasury has always focused on managing liquidity, cash flow, risk, and compliance, today's AI-powered tools are eliminating manual workloads and enabling real-time, data-driven decision making that wasn't possible before.
The impact is already measurable. Document extraction accuracy has jumped from 65% to 95% using AI agents, dramatically reducing the need for manual matching and reconciliation. This frees treasury professionals to focus on strategic questions: What's driving cash flow delays? Why are dispute rates increasing? How can we optimize working capital? Real-world implementations at companies like McNaughton McKay demonstrate how these technologies improve not just receivables functionality but credit management as well.
What makes today's AI revolution different from previous technological advances? As Yogs explains, "AI is moving at a speed I haven't seen before in my 25+ year career." The convergence of big data, cloud computing, and specialized hardware has created perfect conditions for transformative change. Yet the fundamentals remain constant - treasury's core problems haven't changed, but how we solve them has.
For treasury leaders looking to navigate this transformation, Yogs recommends his "3×3" talent development framework, tailoring AI skills development across both proficiency levels and career stages. The message is clear: AI is here to stay, and the organizations that thrive will be those that embrace these technologies while maintaining unwavering focus on solving their customers' financial challenges.
To learn more visit: https://www.deluxe.com/receivables-management/cash-application/
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:01):
Welcome to this special four-part series titled
Modern Finance, sponsored byDeluxe, where we explore the
future of financial automation.
From treasury to accountspayable and receivable, we're
diving into how AI andintelligent automation are
transforming every corner offinance.
In each episode, you'll hearfrom leaders at Deluxe who are
(00:23):
driving innovation anddelivering real-world results.
Whether you're navigatingcompliance, fighting fraud or
connecting the financial dots,this series is packed with
insights you won't want to miss.
Let's get started.
Speaker 2 (00:37):
Hello everyone and welcome to the Leaders in
Payments podcast.
I'm your host, Greg Myers, andthis episode is part of our
four-part series titled ModernFinance, being brought to you by
Deluxe.
So today we have a very specialguest on the show, Yogs
Jayprakasam, who is the SVP,Chief Technology and Digital
Officer at Deluxe.
So, Yogs, thank you so much forbeing here and welcome to the
(00:58):
show.
Speaker 3 (00:58):
It's my pleasure to be here, Greg.
Thanks for having me.
Speaker 2 (01:01):
Absolutely.
Let's start off having you tella little bit about yourself,
maybe where you're from, whereyou grew up, where you currently
live, just a few things likethat.
Speaker 3 (01:09):
That's a fantastic question.
I'm originally from India.
I was born and raised here.
I did my master's in computerapplication.
Then I started my career with astartup company Now India has a
very thriving startup ecosystem, but back then, when I started,
I would say I'm one of theearliest cohorts in the startup
(01:31):
community.
Then I moved on to consulting.
I did consulting for severalyears, which led me to american
express where I spent 13 yearsof my career working in payments
across b2b b, b2c merchantservices network business, which
brought me to fantasticopportunity with Deluxe and the
current role that I'm in.
(01:51):
But coincidentally I happen tobe here in India as we speak now
.
Speaker 2 (01:56):
Okay, great.
Well, so for the audience whomay not know, can you tell us a
little bit about Deluxe?
Speaker 3 (02:02):
Absolutely.
Deluxe is a 110-year paymentsand data company with our
original roots in check printing, our founder, wh Hotkiss, who
borrowed $300 from hisfather-in-law.
So the story goes where hefounded a little company to
invent what is called, 110 yearsback, check and checkbook.
(02:25):
So he is the original inventorand we proudly call ourselves
the original payments companyBecause, after the currency,
check is the first alternativepayment form.
From there today we are about$2.2 million revenue company,
publicly traded.
Half of our revenue still comesfrom our original check and
(02:46):
printing business, as we call,and the rest of the revenue
comes from our payments and dataside of growth business, which
is composed of three growthbusinesses deluxe merchant
services, b2b payments, whichincludes treasury management,
receivables and paymentspayables automation solution,
and we also have a fastgrowinggrowing data-driven
(03:06):
marketing business.
We serve 4,000-plus financialinstitutions across the US and
North America and millions ofsmall business and mid-size
companies through our solutions.
Speaker 2 (03:19):
Okay, so tell us a little bit about your role today
.
Speaker 3 (03:22):
I think my role is a very exciting role.
I joined the company about threeyears back and the way our CEO
defined my role is for amanufacturing company, using
technology is more as a utilityto support the business, but
when it comes to payments anddata, as you know, technology is
the business.
(03:43):
So the products and servicesthat we build is all pretty much
technology, particularly inpayments.
All things that we do istechnology.
So my team's role is to helpthe company to leverage
technology from technology beingthe utility to truly turning
this into a technology-drivenproduct development.
So we have been modernizing ourtechnology ecosystems into more
(04:07):
of a platform-driven company.
We created more data, ai andAPI foundational platform
ecosystem for the company, which, by the way, won a CAO 100
award for 2024.
On top of that, we built a lotof innovative payment products,
ranging from digital wallets,our payment gateways, various
(04:30):
DeluxePay mobile.
Colleagues use services insidethe company, whether it is HR,
legal, finance department, sales, marketing.
We support every one of ourinternal colleagues as our
(05:00):
internal customers in the way weuse technology to help their
jobs to become easier and todeliver their services to our
customers.
Speaker 2 (05:08):
Okay, great.
Well, thanks for sharing allthat background information.
I think it's important to helplevel set.
So let's dive into the topic athand, which is one that's on
the top of mind for everyone inour industry and a lot of
industries across the world,right?
So we're going to talk about AIand how it's really modernizing
payments in the paymentsecosystem, with a little bit of
(05:29):
a special emphasis on treasury.
So, if you don't mind to levelset, can you maybe describe what
the typical treasury functionlooks like and maybe the kind of
size and kinds of companies andfinancial institutions that you
work with every day?
Speaker 3 (05:43):
I believe treasury.
I always think treasury as oneof the oldest functions in the
world of business in my mindbecause ever since the first
business would have been created, you probably created treasury
function first, because money isthe core foundation of how you
operate your business.
In my opinion, treasuryincludes the core function of
(06:05):
managing liquidity, cash flowmanagement, risk, compliance,
fraud all those foundationalcapabilities as well.
So the way money comes in andthe way money goes out, managing
all of that thing is whattreasury's function is.
But in the past I believe,treasury was considered more of
a back office function.
But more and more today, withthe technological innovation,
(06:26):
treasury is becoming more of astrategic advisor how quickly
you can settle the cash andpredict the cash flow and really
figure out which product isgenerating the expected revenue,
etc.
It all becomes part of thetreasury management function as
well.
So the type of customers that weserve we either serve through
(06:48):
4,000 financial institutionsthat we sell and they take our
capabilities and deliver totheir customers, or we also
deliver our treasury managementsolutions to our end customers
directly as well.
So typically the size that wework with would be a
medium-sized company with a $50million plus revenue range all
(07:10):
the way to a multi-billiondollar revenue company.
So what works nicely for thekind of capabilities that we
deliver is you have thousands ofinvoices that you are receiving
and your customers pay throughmultiple payment modalities, and
your payment also is associatedwith a lot of data, invoicing,
addendums and evidences.
(07:30):
So, finally, applying morecomplex business logic to settle
in your ERP systems and applyyour predictive cash flow
management and all of thosecapabilities is what potentially
becomes the best target market.
We always remind our customersthat treasury can be complex.
(07:51):
Complex is okay, complicated isnot, and we apply technology to
simplify the complex for ourcustomers.
Speaker 2 (07:59):
Okay, okay, great.
So, as you well know, andeveryone knows, that AI is
playing a bigger role intreasury operations today, so
maybe can you walk us throughhow technologies like AI, ml,
machine learning and agenticcommerce maybe, or agentic AI,
how that's sort of reshaping thetreasury operations today.
Speaker 3 (08:20):
Yeah, I think everyone in treasury would say
AA is not new for them, right?
Ever since the machine learningstatistical modeling started,
treasury definitely is one ofthe areas that has been using
and leveraging that technologyand capabilities.
So the way I look at thespectrum is treasury management,
and finance in particular, hasbeen leveraging machine learning
(08:43):
statistical modeling for quitesome time, and now with the
breakthrough innovation ofgenerative AI, which allows
tertiary management inparticular to leverage the
ability to reason with theinformation and allow for you to
have natural languageconversation with the machine,
(09:04):
makes things a lot easier.
So the way to think about thisis your statistical modeling
created more predictive patternrecognition in a particular
situation, whether it is cashflow, fraud detection and stuff
like that.
Now with the generative AI, youcan actually interact with the
system, asking questions likedefine my FX rate, breaches of
(09:29):
certain threshold and thingslike that.
But certainly the agentic AI isa paradigm shift.
I mean, generative AI is aparadigm shift on its own
measure, but it lowers, in myopinion, the barrier for anyone
to leverage this technology.
You don't need to be astatistician, you don't need to
be a data scientist.
You should be able to ask theright question Any good domain
(09:52):
expert in treasury managementcan ask a very good question to
get the right answers from themachine.
But agentic AI becomes a lotmore targeted in the sense that
it can act as your agentextension.
So you can apply more reasoninginto the capability.
You can let the machine definethe strategy and threshold.
(10:15):
So in a situation like yourEuro strategy, if it preaches 5%
deviation, execute my hedgingstrategy.
Now an agent can actuallydefine what that hedging
strategy looks like.
Earlier you might want yourexpert to define it.
Now your agent can actuallydefine it and even validate what
(10:36):
that hedging strategy lookslike.
So in the paradigm oftechnology, I think it is an
exciting time for domain expertsin treasury management
particularly.
All you've got to know is howto use the technology and ask
the right question that ispertinent in your domain.
Speaker 2 (10:54):
What kind of manual workloads are being eliminated
through AI, and how is thatenabling treasury teams to maybe
make more real-time data-drivendecisions?
Speaker 3 (11:05):
I think the immediate place that AI addresses a lot
for treasury management isautomating and improving the
quality of the documentextraction Because, as you can
imagine, in B2B handling whetherit is check images or invoices,
even addendum data extractionof the image quality could be
(11:28):
very tricky and with ourspecific agent capabilities that
we have developed, our accuracyrate for such image extraction
dramatically improved fromroughly about 65% before to
about 95%, so you can imagine a30% bump in accuracy Handling.
Image extraction and error ratereduction is immediate uplift
(11:53):
for treasury agents to now nothave to do matching,
reconciliation, exceptionhandling, to really start to
focus on your strategicdecisioning, senior leaders, on
what is driving your cash flowdelays, potentially what is
driving your increased disputerates and how do you address
(12:15):
those dispute rate challengesand things like that.
So that's where, immediately,we are seeing a lot of uplift.
But of course, we have a lot ofexciting new features that we
are in the development usingmore agentic capabilities as
well.
Speaker 2 (12:30):
Okay, Can you share maybe some real-world examples,
maybe of a company or a bankthat's successfully integrating
AI into their treasuryoperations?
I know we've said it's sort ofalways been there, but I think
it's gotten to the point wherepeople are using it better
differently, solving uniqueproblems.
So maybe share some real-worldexamples, and maybe something in
(12:52):
cash management, forecasting ora risk mitigation area.
Speaker 3 (12:57):
That's a great question, greg.
I see not just with our clients.
There are many companies thatare taking bold actions.
Even banks are known forleveraging newer technologies
very risk averse, particularlyin the banking domain, but now I
see a lot of headlines wherethe banks are taking bold
actions, partnering withcompanies like OpenAI, using a
(13:19):
lot of AI capabilities.
So our clients, particularly.
You can go look up on ourwebsite.
We recently posted a case studywith one of our direct clients,
mcnaughton McKay, who publishedleveraging agentic capabilities
and how they improved theircash management applications,
reducing header handling, andparticularly how it not only
(13:42):
improved the receivablesfunctionality but also credit
management side of things aswell.
We also recently published, inpartnership with one of our
banking customers, comerica, howthey are leveraging receivables
capabilities in partnershipwith one of our banking
customers, comerica, how theyare leveraging receivables
capabilities in partnership withus, and how things are
improving too.
As we speak, we are working onour next level, cash application
(14:03):
module.
Some of our primary customersare working in alpha mode.
More to come on that.
I'm waiting to get thoseresults out soon as well.
Speaker 2 (14:13):
Yeah, and I think if you keep up with AI and you read
about it, especially inregulated industries like
banking, governance, compliance,risk mitigation are huge
factors.
So, as more and more financialinstitutions are adopting AI,
what are some of the bestpractices that you see or that
you help your clients with whenyou know it comes to that kind
(14:36):
of governance and compliance?
Especially, you know around thewhole new AI thing, the agentic
, maybe even the.
You know the tools that youbring to the table.
You know what are some of thosebest practices.
Speaker 3 (14:48):
That's a wonderful question, because it's very easy
to jump onto any shiny objectand not forgetting the
foundational discipline that weneed to have.
We always tell that ourapproach to AI is more
responsible adoption of AI,which, in my opinion, goes back
to starting with the permissibleuse of the data.
(15:08):
So whose data are you handlingand what permissions do you have
to use that data?
So it goes back to the basicfoundational 101 of data
governance principles, so thatwe are very particular about.
So we have obligations to ourcustomers and their customers
about what we can and cannot dowith our data.
(15:30):
So the approach that we take iswhere we have permissible
rights to use the data, then weuse the data to specifically
train the model only for the useof that particular customer.
So our approach is not touniversally take the data and
train across the customer side,and we continue to elevate that
(15:51):
conversation with our customersas well.
We have a dedicated advisoryboard where we periodically go
back and talk about our approachto AI and how we are
maintaining the data governance,and also sometimes we push the
envelope a little bit more onhow we may want to rethink the
usage of the data.
(16:11):
The idea is we create a safespace for the customer community
to feel comfortable with theway that we are agreeing to use
the data.
That way, when we provide thevalue back to our customers, we
all know that we still did rightby our customers.
Our approach is not just dowhat law mandates.
It is what is right by ourcustomers.
(16:32):
Our approach is not just dowhat law mandates.
It is what is right by ourcustomers, and that's our
principle.
So in order to do that, we workclosely with the client
advisory board.
We also internally have AIgovernance team, which includes,
by the way, many of my peergroups, so we constantly review
the approach to AI and datagovernance together as well.
Speaker 2 (16:54):
Okay, great.
So obviously you know there's alot of excitement around AI and
what's possible and what's next.
But from your perspective Imean, you're living, breathing
this every day, both what Ithink is interesting internally
within your company, but alsoexternally within you,
externally within your clientset.
So, from your opinion, what'sreal, what's next and what
(17:18):
should treasury operations andtreasury leaders really be
prioritizing right now?
Speaker 3 (17:23):
So that's a fantastic question, and I personally have
been passionate about not onlylearning AI but seriously
thinking about what's happeningin AI and this important
question around who's movingwho's cheese, in other words,
where the value is moving.
So I fundamentally believe thatAI dramatically moved, much
(17:48):
faster than any other technologyI witnessed.
In my 25-plus yeartechnological career, I
witnessed multipletransformations.
I have been part of helpingmany companies drive those
transformations too.
Ai is moving in a speed that Ihave not seen before.
The reason for that is AI isusing many of the technological
(18:10):
advancements that already tookplace.
Whether it is big datatransformation, whether it is
cloud transformation, or noweven the chip-level hardware
transformation.
It's all now colliding togetherat a point where AI is
potentially becoming the user inthat particular area.
So AI is real and it isfundamentally going to change,
(18:32):
in my opinion, every layer ofthe software delivery that we
have known.
That's why many companies andcountries are actually investing
in a lot of power plants,because AI requires a lot of
computation, so you need a lotmore power than what we can
generate today.
So it starts with power, thenit moves up to the
(18:52):
infrastructure.
So there are a lot of newinfrastructure being developed
by many countries.
In many of our United Statesbig states as well whether take
Texas as an example,indianapolis, arizona there are
new chip makers coming.
Then the value moves up frominfrastructure and chip to how
you are building the softwareand then to applications like
(19:14):
threshold management.
So the point that I'm trying tomake for the threshold
management leaders is you mightfeel that it's not moving fast
because you are not seeing it inyour area immediately, right
now, but I would like you tothink that it's fundamentally
changing everything.
So it will feel like it iscoming slow and then it will
(19:34):
feel it is sudden.
So the way to keep up with thisis, in my opinion, each of us
should be investing and learningabout AI and how to use AI.
That's number one, because whenyou start to use AI, you will
start to realize how toreimagine the way that you are
solving problems.
So that's number one.
(19:55):
So number two is I think it'sgoing to affect various career
levels differently.
So if you were hiring anentry-level treasury analyst
before, so now the entry-levelanalyst will come with AI
technology learning already.
So the way you can help them tocatch up on your soft skills,
(20:17):
how to approach decision-making,exception handling inside a
company, they can become veryproductive faster.
But if you are intimidated as amid-career leader because they
speak a lot about AI and you arenot, then that's going to
create friction as well.
So the leaders should one,learn about AI.
Two, try to combine theentry-level folks and their AI
(20:40):
skills with your expertise andsoft skills.
And three, facilitate moreconversation with your
organizations on how toreimagine the problems that you
are already solving.
So, bottom line, myrecommendation is AI is already
here.
You may not feel the change fastenough, but starting to
reimagine the way that you aresolving solutions will
(21:01):
tremendously help.
You will not know how toreimagine unless you really
start to use this technology.
For example, I witnessed a lotof people using ChatGPT and they
think that they are using AI,but many folks are using ChatGPT
the exact same way that you'reusing Google search.
That's not what it is meant tobe.
(21:23):
You've got to know how toprompt the machine learning
model sorry, the large languagemodel.
So the way to get the answer isabout the way you ask the
question and that is not likesearching in Google.
So, learning about promptengineering, learning about how
to prompt it the right way, it'sall going to be critical.
But on top of that, alsofiguring out what's your
(21:46):
intellectual property for thecompany that you can't leave it
out on the internet.
So working with your technologypartners, or even partners like
Dilex, to figure out how do youprotect your intellectual
property with the architecturalapproach, such as retrieval,
augmented generative, whichprotects your IP while also
(22:09):
taking advantage of the power ofthe large language model.
So this combination is what Iwould recommend for the leaders
to consider going forward.
Speaker 2 (22:17):
Okay, Okay.
Well, Yogs, this has been agreat discussion.
I mean, there's so much here.
I feel like we're still in theearly days, right, we're maybe
in the top of the first inning,just getting started for a
baseball analogy.
But maybe if you could try tosummarize this conversation and
provide treasury leaders maybewith just one key insight or one
(22:38):
piece of advice I know it'll bereally hard to boil it down to
just one, but if you can try toboil it down to maybe just one
piece of advice, what would thatbe?
Speaker 3 (22:45):
My one piece of advice would be that AA is here.
It is going to fundamentallychange everything, but what it
will not change is your focus onthe customer and the customer's
problem that you solve Intreasury.
Your problems are cash flowmanagement, free cash liquidity
(23:07):
management, risk and compliancemanagement.
Those problems will always bethere.
So how do you now reimaginethose problems through the lens
of ai and how do you solve itdifferently is what I would like
you to take away, but you canonly do it if you believe how
this technology is fundamentallychanging.
(23:28):
That requires self-learning andcontinuous learning so you can
practically apply the technologyto solve.
So the bottom line, yourproblem is not changing.
Your focus on customer is notchanging.
Hang on to what is not changingand apply the technology on
those problems.
Speaker 2 (23:47):
Okay, I think that's a great way to wrap up the show,
but I do want to open it up,see if there's anything else you
wanted to cover, any topics wemay have missed, before we wrap
up the show.
Speaker 3 (23:56):
So in treasury management.
I think we talked a lot aboutthe domain of treasury
management and how it is goingto evolve with AI.
So I would like the treasuryleaders to think that this is a
great opportunity for each ofyou to become a strategic
advisor for your business, frombeing a back office leaders
(24:18):
providing some reports andliquid positions and things like
that.
But the way you can approachthis problem is also through the
lens of talent development inyour tertiary management domain.
So inside Dilex and alsooutside, I continue to preach
talent approach, what I callthree by three and I would like
(24:42):
to conclude with that thoughtfrom talent development
standpoint as well.
So what I mean by three bythree is when you learn any new
technology, you will always gointo three phases, whether it is
a beginner, intermediate andadvanced.
But actually, when it comes toAI, I would like us to also
think about based on whichcareer level that you are in.
Your beginner, intermediate andadvanced are going to be
(25:05):
totally different.
So that is what the next threelayers are in my opinion.
So if you are entry-level,intermediate career or an
executive level, so that threecombined with this three, which
is beginner, intermediate andadvanced level, forms your
three-by-three metrics.
What I mean by that is, as abeginner learning AI, prompt
(25:29):
engineering is something we allhave to learn, but an
entry-level person prompting thelarge language model would be
different than a seniorexecutive or a CEO.
So I would like the CEOs to bethinking about how do I prompt
large language model to evendevelop a point of view for a
new strategy?
(25:50):
For example, stablecoin is a bigtopic now with the Genius Act
and we are all in payments world.
So even developing what is itgoing to do the stablecoin is
going to do inside payments.
You don't have to wait for anexpert to come and tell you a
30-page report.
You actually can go prompt tochat GPT or Perplexity or Gemini
(26:12):
whichever is the large languagemodel medium that you use, if
you're prompted to say, okay,use jobs to be done framework or
pick any strategic frameworkand say, give me a report on
stablecoin and its impact on mybusiness, so it will provide a
preliminary strategic point ofview.
(26:34):
That's fascinating for a CISO toquickly understand what's going
to happen with the emergingstablecoin.
So I just wanted to give thatone example to say that if each
company should be thinking aboutcreating a 3x3 talent strategy
on how you are going to educateyour organization on AI
development, and that goes withthe Y-axis entry-level,
(26:59):
mid-career and executive leveland X-axis beginner,
intermediate and advanced level,and that's exactly what we are
doing inside Deluxe as well.
So, again, to conclude, takeany domain, but, since we are
talking about, treasury is anamazing place to be in in the
age of AI but take that, learnfor yourself.
(27:21):
But also apply the three bythree talent strategy, which
will help you transform the waythat you run your company and
your business.
Speaker 2 (27:30):
Well, yogs, I think that's an incredible way to wrap
up the show.
So thank you so much for beinghere today sharing all your
great insights.
I really appreciate your time.
Speaker 3 (27:38):
Thank you so much for this great opportunity.
Greg, have a wonderful day.
Speaker 2 (27:41):
You too, and to the rest of you out there, you
listeners out there.
Thank you so much for your timeas well, and until the next
story.
Speaker 1 (27:50):
Thank you for listening to today's episode.
If you'd like more informationon the transformative potential
of AI and automation in modernfinance, please visit
wwwdeluxecom.
Slash receivables hyphenmanagement.
Slash cash hyphen application.
Welcome to this special four-part series titled
Modern Finance, sponsored byDeluxe, where we explore the
future of financial automation.
From treasury to accountspayable and receivable, we're
diving into how AI andintelligent automation are
transforming every corner offinance.
In each episode, you'll hearfrom leaders at Deluxe who are
(00:23):
driving innovation anddelivering real-world results.
Whether you're navigatingcompliance, fighting fraud or
connecting the financial dots,this series is packed with
insights you won't want to miss.
Let's get started.
Speaker 2 (00:37):
Hello everyone and welcome to the Leaders in
Payments podcast.
I'm your host, Greg Myers, andthis episode is part of our
four-part series titled ModernFinance, being brought to you by
Deluxe.
So today we have a very specialguest on the show, Yogs
Jayprakasam, who is the SVP,Chief Technology and Digital
Officer at Deluxe.
So, Yogs, thank you so much forbeing here and welcome to the
(00:58):
show.
Speaker 3 (00:58):
It's my pleasure to be here, Greg.
Thanks for having me.
Speaker 2 (01:01):
Absolutely.
Let's start off having you tella little bit about yourself,
maybe where you're from, whereyou grew up, where you currently
live, just a few things likethat.
Speaker 3 (01:09):
That's a fantastic question.
I'm originally from India.
I was born and raised here.
I did my master's in computerapplication.
Then I started my career with astartup company Now India has a
very thriving startup ecosystem, but back then, when I started,
I would say I'm one of theearliest cohorts in the startup
(01:31):
community.
Then I moved on to consulting.
I did consulting for severalyears, which led me to american
express where I spent 13 yearsof my career working in payments
across b2b b, b2c merchantservices network business, which
brought me to fantasticopportunity with Deluxe and the
current role that I'm in.
(01:51):
But coincidentally I happen tobe here in India as we speak now
.
Speaker 2 (01:56):
Okay, great.
Well, so for the audience whomay not know, can you tell us a
little bit about Deluxe?
Speaker 3 (02:02):
Absolutely.
Deluxe is a 110-year paymentsand data company with our
original roots in check printing, our founder, wh Hotkiss, who
borrowed $300 from hisfather-in-law.
So the story goes where hefounded a little company to
invent what is called, 110 yearsback, check and checkbook.
(02:25):
So he is the original inventorand we proudly call ourselves
the original payments companyBecause, after the currency,
check is the first alternativepayment form.
From there today we are about$2.2 million revenue company,
publicly traded.
Half of our revenue still comesfrom our original check and
(02:46):
printing business, as we call,and the rest of the revenue
comes from our payments and dataside of growth business, which
is composed of three growthbusinesses deluxe merchant
services, b2b payments, whichincludes treasury management,
receivables and paymentspayables automation solution,
and we also have a fastgrowinggrowing data-driven
(03:06):
marketing business.
We serve 4,000-plus financialinstitutions across the US and
North America and millions ofsmall business and mid-size
companies through our solutions.
Speaker 2 (03:19):
Okay, so tell us a little bit about your role today
.
Speaker 3 (03:22):
I think my role is a very exciting role.
I joined the company about threeyears back and the way our CEO
defined my role is for amanufacturing company, using
technology is more as a utilityto support the business, but
when it comes to payments anddata, as you know, technology is
the business.
(03:43):
So the products and servicesthat we build is all pretty much
technology, particularly inpayments.
All things that we do istechnology.
So my team's role is to helpthe company to leverage
technology from technology beingthe utility to truly turning
this into a technology-drivenproduct development.
So we have been modernizing ourtechnology ecosystems into more
(04:07):
of a platform-driven company.
We created more data, ai andAPI foundational platform
ecosystem for the company, which, by the way, won a CAO 100
award for 2024.
On top of that, we built a lotof innovative payment products,
ranging from digital wallets,our payment gateways, various
(04:30):
DeluxePay mobile.
Colleagues use services insidethe company, whether it is HR,
legal, finance department, sales, marketing.
We support every one of ourinternal colleagues as our
(05:00):
internal customers in the way weuse technology to help their
jobs to become easier and todeliver their services to our
customers.
Speaker 2 (05:08):
Okay, great.
Well, thanks for sharing allthat background information.
I think it's important to helplevel set.
So let's dive into the topic athand, which is one that's on
the top of mind for everyone inour industry and a lot of
industries across the world,right?
So we're going to talk about AIand how it's really modernizing
payments in the paymentsecosystem, with a little bit of
(05:29):
a special emphasis on treasury.
So, if you don't mind to levelset, can you maybe describe what
the typical treasury functionlooks like and maybe the kind of
size and kinds of companies andfinancial institutions that you
work with every day?
Speaker 3 (05:43):
I believe treasury.
I always think treasury as oneof the oldest functions in the
world of business in my mindbecause ever since the first
business would have been created, you probably created treasury
function first, because money isthe core foundation of how you
operate your business.
In my opinion, treasuryincludes the core function of
(06:05):
managing liquidity, cash flowmanagement, risk, compliance,
fraud all those foundationalcapabilities as well.
So the way money comes in andthe way money goes out, managing
all of that thing is whattreasury's function is.
But in the past I believe,treasury was considered more of
a back office function.
But more and more today, withthe technological innovation,
(06:26):
treasury is becoming more of astrategic advisor how quickly
you can settle the cash andpredict the cash flow and really
figure out which product isgenerating the expected revenue,
etc.
It all becomes part of thetreasury management function as
well.
So the type of customers that weserve we either serve through
(06:48):
4,000 financial institutionsthat we sell and they take our
capabilities and deliver totheir customers, or we also
deliver our treasury managementsolutions to our end customers
directly as well.
So typically the size that wework with would be a
medium-sized company with a $50million plus revenue range all
(07:10):
the way to a multi-billiondollar revenue company.
So what works nicely for thekind of capabilities that we
deliver is you have thousands ofinvoices that you are receiving
and your customers pay throughmultiple payment modalities, and
your payment also is associatedwith a lot of data, invoicing,
addendums and evidences.
(07:30):
So, finally, applying morecomplex business logic to settle
in your ERP systems and applyyour predictive cash flow
management and all of thosecapabilities is what potentially
becomes the best target market.
We always remind our customersthat treasury can be complex.
(07:51):
Complex is okay, complicated isnot, and we apply technology to
simplify the complex for ourcustomers.
Speaker 2 (07:59):
Okay, okay, great.
So, as you well know, andeveryone knows, that AI is
playing a bigger role intreasury operations today, so
maybe can you walk us throughhow technologies like AI, ml,
machine learning and agenticcommerce maybe, or agentic AI,
how that's sort of reshaping thetreasury operations today.
Speaker 3 (08:20):
Yeah, I think everyone in treasury would say
AA is not new for them, right?
Ever since the machine learningstatistical modeling started,
treasury definitely is one ofthe areas that has been using
and leveraging that technologyand capabilities.
So the way I look at thespectrum is treasury management,
and finance in particular, hasbeen leveraging machine learning
(08:43):
statistical modeling for quitesome time, and now with the
breakthrough innovation ofgenerative AI, which allows
tertiary management inparticular to leverage the
ability to reason with theinformation and allow for you to
have natural languageconversation with the machine,
(09:04):
makes things a lot easier.
So the way to think about thisis your statistical modeling
created more predictive patternrecognition in a particular
situation, whether it is cashflow, fraud detection and stuff
like that.
Now with the generative AI, youcan actually interact with the
system, asking questions likedefine my FX rate, breaches of
(09:29):
certain threshold and thingslike that.
But certainly the agentic AI isa paradigm shift.
I mean, generative AI is aparadigm shift on its own
measure, but it lowers, in myopinion, the barrier for anyone
to leverage this technology.
You don't need to be astatistician, you don't need to
be a data scientist.
You should be able to ask theright question Any good domain
(09:52):
expert in treasury managementcan ask a very good question to
get the right answers from themachine.
But agentic AI becomes a lotmore targeted in the sense that
it can act as your agentextension.
So you can apply more reasoninginto the capability.
You can let the machine definethe strategy and threshold.
(10:15):
So in a situation like yourEuro strategy, if it preaches 5%
deviation, execute my hedgingstrategy.
Now an agent can actuallydefine what that hedging
strategy looks like.
Earlier you might want yourexpert to define it.
Now your agent can actuallydefine it and even validate what
(10:36):
that hedging strategy lookslike.
So in the paradigm oftechnology, I think it is an
exciting time for domain expertsin treasury management
particularly.
All you've got to know is howto use the technology and ask
the right question that ispertinent in your domain.
Speaker 2 (10:54):
What kind of manual workloads are being eliminated
through AI, and how is thatenabling treasury teams to maybe
make more real-time data-drivendecisions?
Speaker 3 (11:05):
I think the immediate place that AI addresses a lot
for treasury management isautomating and improving the
quality of the documentextraction Because, as you can
imagine, in B2B handling whetherit is check images or invoices,
even addendum data extractionof the image quality could be
(11:28):
very tricky and with ourspecific agent capabilities that
we have developed, our accuracyrate for such image extraction
dramatically improved fromroughly about 65% before to
about 95%, so you can imagine a30% bump in accuracy Handling.
Image extraction and error ratereduction is immediate uplift
(11:53):
for treasury agents to now nothave to do matching,
reconciliation, exceptionhandling, to really start to
focus on your strategicdecisioning, senior leaders, on
what is driving your cash flowdelays, potentially what is
driving your increased disputerates and how do you address
(12:15):
those dispute rate challengesand things like that.
So that's where, immediately,we are seeing a lot of uplift.
But of course, we have a lot ofexciting new features that we
are in the development usingmore agentic capabilities as
well.
Speaker 2 (12:30):
Okay, Can you share maybe some real-world examples,
maybe of a company or a bankthat's successfully integrating
AI into their treasuryoperations?
I know we've said it's sort ofalways been there, but I think
it's gotten to the point wherepeople are using it better
differently, solving uniqueproblems.
So maybe share some real-worldexamples, and maybe something in
(12:52):
cash management, forecasting ora risk mitigation area.
Speaker 3 (12:57):
That's a great question, greg.
I see not just with our clients.
There are many companies thatare taking bold actions.
Even banks are known forleveraging newer technologies
very risk averse, particularlyin the banking domain, but now I
see a lot of headlines wherethe banks are taking bold
actions, partnering withcompanies like OpenAI, using a
(13:19):
lot of AI capabilities.
So our clients, particularly.
You can go look up on ourwebsite.
We recently posted a case studywith one of our direct clients,
mcnaughton McKay, who publishedleveraging agentic capabilities
and how they improved theircash management applications,
reducing header handling, andparticularly how it not only
(13:42):
improved the receivablesfunctionality but also credit
management side of things aswell.
We also recently published, inpartnership with one of our
banking customers, comerica, howthey are leveraging receivables
capabilities in partnershipwith one of our banking
customers, comerica, how theyare leveraging receivables
capabilities in partnership withus, and how things are
improving too.
As we speak, we are working onour next level, cash application
(14:03):
module.
Some of our primary customersare working in alpha mode.
More to come on that.
I'm waiting to get thoseresults out soon as well.
Speaker 2 (14:13):
Yeah, and I think if you keep up with AI and you read
about it, especially inregulated industries like
banking, governance, compliance,risk mitigation are huge
factors.
So, as more and more financialinstitutions are adopting AI,
what are some of the bestpractices that you see or that
you help your clients with whenyou know it comes to that kind
(14:36):
of governance and compliance?
Especially, you know around thewhole new AI thing, the agentic
, maybe even the.
You know the tools that youbring to the table.
You know what are some of thosebest practices.
Speaker 3 (14:48):
That's a wonderful question, because it's very easy
to jump onto any shiny objectand not forgetting the
foundational discipline that weneed to have.
We always tell that ourapproach to AI is more
responsible adoption of AI,which, in my opinion, goes back
to starting with the permissibleuse of the data.
(15:08):
So whose data are you handlingand what permissions do you have
to use that data?
So it goes back to the basicfoundational 101 of data
governance principles, so thatwe are very particular about.
So we have obligations to ourcustomers and their customers
about what we can and cannot dowith our data.
(15:30):
So the approach that we take iswhere we have permissible
rights to use the data, then weuse the data to specifically
train the model only for the useof that particular customer.
So our approach is not touniversally take the data and
train across the customer side,and we continue to elevate that
(15:51):
conversation with our customersas well.
We have a dedicated advisoryboard where we periodically go
back and talk about our approachto AI and how we are
maintaining the data governance,and also sometimes we push the
envelope a little bit more onhow we may want to rethink the
usage of the data.
(16:11):
The idea is we create a safespace for the customer community
to feel comfortable with theway that we are agreeing to use
the data.
That way, when we provide thevalue back to our customers, we
all know that we still did rightby our customers.
Our approach is not just dowhat law mandates.
It is what is right by ourcustomers.
(16:32):
Our approach is not just dowhat law mandates.
It is what is right by ourcustomers, and that's our
principle.
So in order to do that, we workclosely with the client
advisory board.
We also internally have AIgovernance team, which includes,
by the way, many of my peergroups, so we constantly review
the approach to AI and datagovernance together as well.
Speaker 2 (16:54):
Okay, great.
So obviously you know there's alot of excitement around AI and
what's possible and what's next.
But from your perspective Imean, you're living, breathing
this every day, both what Ithink is interesting internally
within your company, but alsoexternally within you,
externally within your clientset.
So, from your opinion, what'sreal, what's next and what
(17:18):
should treasury operations andtreasury leaders really be
prioritizing right now?
Speaker 3 (17:23):
So that's a fantastic question, and I personally have
been passionate about not onlylearning AI but seriously
thinking about what's happeningin AI and this important
question around who's movingwho's cheese, in other words,
where the value is moving.
So I fundamentally believe thatAI dramatically moved, much
(17:48):
faster than any other technologyI witnessed.
In my 25-plus yeartechnological career, I
witnessed multipletransformations.
I have been part of helpingmany companies drive those
transformations too.
Ai is moving in a speed that Ihave not seen before.
The reason for that is AI isusing many of the technological
(18:10):
advancements that already tookplace.
Whether it is big datatransformation, whether it is
cloud transformation, or noweven the chip-level hardware
transformation.
It's all now colliding togetherat a point where AI is
potentially becoming the user inthat particular area.
So AI is real and it isfundamentally going to change,
(18:32):
in my opinion, every layer ofthe software delivery that we
have known.
That's why many companies andcountries are actually investing
in a lot of power plants,because AI requires a lot of
computation, so you need a lotmore power than what we can
generate today.
So it starts with power, thenit moves up to the
(18:52):
infrastructure.
So there are a lot of newinfrastructure being developed
by many countries.
In many of our United Statesbig states as well whether take
Texas as an example,indianapolis, arizona there are
new chip makers coming.
Then the value moves up frominfrastructure and chip to how
you are building the softwareand then to applications like
(19:14):
threshold management.
So the point that I'm trying tomake for the threshold
management leaders is you mightfeel that it's not moving fast
because you are not seeing it inyour area immediately, right
now, but I would like you tothink that it's fundamentally
changing everything.
So it will feel like it iscoming slow and then it will
(19:34):
feel it is sudden.
So the way to keep up with thisis, in my opinion, each of us
should be investing and learningabout AI and how to use AI.
That's number one, because whenyou start to use AI, you will
start to realize how toreimagine the way that you are
solving problems.
So that's number one.
(19:55):
So number two is I think it'sgoing to affect various career
levels differently.
So if you were hiring anentry-level treasury analyst
before, so now the entry-levelanalyst will come with AI
technology learning already.
So the way you can help them tocatch up on your soft skills,
(20:17):
how to approach decision-making,exception handling inside a
company, they can become veryproductive faster.
But if you are intimidated as amid-career leader because they
speak a lot about AI and you arenot, then that's going to
create friction as well.
So the leaders should one,learn about AI.
Two, try to combine theentry-level folks and their AI
(20:40):
skills with your expertise andsoft skills.
And three, facilitate moreconversation with your
organizations on how toreimagine the problems that you
are already solving.
So, bottom line, myrecommendation is AI is already
here.
You may not feel the change fastenough, but starting to
reimagine the way that you aresolving solutions will
(21:01):
tremendously help.
You will not know how toreimagine unless you really
start to use this technology.
For example, I witnessed a lotof people using ChatGPT and they
think that they are using AI,but many folks are using ChatGPT
the exact same way that you'reusing Google search.
That's not what it is meant tobe.
(21:23):
You've got to know how toprompt the machine learning
model sorry, the large languagemodel.
So the way to get the answer isabout the way you ask the
question and that is not likesearching in Google.
So, learning about promptengineering, learning about how
to prompt it the right way, it'sall going to be critical.
But on top of that, alsofiguring out what's your
(21:46):
intellectual property for thecompany that you can't leave it
out on the internet.
So working with your technologypartners, or even partners like
Dilex, to figure out how do youprotect your intellectual
property with the architecturalapproach, such as retrieval,
augmented generative, whichprotects your IP while also
(22:09):
taking advantage of the power ofthe large language model.
So this combination is what Iwould recommend for the leaders
to consider going forward.
Speaker 2 (22:17):
Okay, Okay.
Well, Yogs, this has been agreat discussion.
I mean, there's so much here.
I feel like we're still in theearly days, right, we're maybe
in the top of the first inning,just getting started for a
baseball analogy.
But maybe if you could try tosummarize this conversation and
provide treasury leaders maybewith just one key insight or one
(22:38):
piece of advice I know it'll bereally hard to boil it down to
just one, but if you can try toboil it down to maybe just one
piece of advice, what would thatbe?
Speaker 3 (22:45):
My one piece of advice would be that AA is here.
It is going to fundamentallychange everything, but what it
will not change is your focus onthe customer and the customer's
problem that you solve Intreasury.
Your problems are cash flowmanagement, free cash liquidity
(23:07):
management, risk and compliancemanagement.
Those problems will always bethere.
So how do you now reimaginethose problems through the lens
of ai and how do you solve itdifferently is what I would like
you to take away, but you canonly do it if you believe how
this technology is fundamentallychanging.
(23:28):
That requires self-learning andcontinuous learning so you can
practically apply the technologyto solve.
So the bottom line, yourproblem is not changing.
Your focus on customer is notchanging.
Hang on to what is not changingand apply the technology on
those problems.
Speaker 2 (23:47):
Okay, I think that's a great way to wrap up the show,
but I do want to open it up,see if there's anything else you
wanted to cover, any topics wemay have missed, before we wrap
up the show.
Speaker 3 (23:56):
So in treasury management.
I think we talked a lot aboutthe domain of treasury
management and how it is goingto evolve with AI.
So I would like the treasuryleaders to think that this is a
great opportunity for each ofyou to become a strategic
advisor for your business, frombeing a back office leaders
(24:18):
providing some reports andliquid positions and things like
that.
But the way you can approachthis problem is also through the
lens of talent development inyour tertiary management domain.
So inside Dilex and alsooutside, I continue to preach
talent approach, what I callthree by three and I would like
(24:42):
to conclude with that thoughtfrom talent development
standpoint as well.
So what I mean by three bythree is when you learn any new
technology, you will always gointo three phases, whether it is
a beginner, intermediate andadvanced.
But actually, when it comes toAI, I would like us to also
think about based on whichcareer level that you are in.
Your beginner, intermediate andadvanced are going to be
(25:05):
totally different.
So that is what the next threelayers are in my opinion.
So if you are entry-level,intermediate career or an
executive level, so that threecombined with this three, which
is beginner, intermediate andadvanced level, forms your
three-by-three metrics.
What I mean by that is, as abeginner learning AI, prompt
(25:29):
engineering is something we allhave to learn, but an
entry-level person prompting thelarge language model would be
different than a seniorexecutive or a CEO.
So I would like the CEOs to bethinking about how do I prompt
large language model to evendevelop a point of view for a
new strategy?
(25:50):
For example, stablecoin is a bigtopic now with the Genius Act
and we are all in payments world.
So even developing what is itgoing to do the stablecoin is
going to do inside payments.
You don't have to wait for anexpert to come and tell you a
30-page report.
You actually can go prompt tochat GPT or Perplexity or Gemini
(26:12):
whichever is the large languagemodel medium that you use, if
you're prompted to say, okay,use jobs to be done framework or
pick any strategic frameworkand say, give me a report on
stablecoin and its impact on mybusiness, so it will provide a
preliminary strategic point ofview.
(26:34):
That's fascinating for a CISO toquickly understand what's going
to happen with the emergingstablecoin.
So I just wanted to give thatone example to say that if each
company should be thinking aboutcreating a 3x3 talent strategy
on how you are going to educateyour organization on AI
development, and that goes withthe Y-axis entry-level,
(26:59):
mid-career and executive leveland X-axis beginner,
intermediate and advanced level,and that's exactly what we are
doing inside Deluxe as well.
So, again, to conclude, takeany domain, but, since we are
talking about, treasury is anamazing place to be in in the
age of AI but take that, learnfor yourself.
(27:21):
But also apply the three bythree talent strategy, which
will help you transform the waythat you run your company and
your business.
Speaker 2 (27:30):
Well, yogs, I think that's an incredible way to wrap
up the show.
So thank you so much for beinghere today sharing all your
great insights.
I really appreciate your time.
Speaker 3 (27:38):
Thank you so much for this great opportunity.
Greg, have a wonderful day.
Speaker 2 (27:41):
You too, and to the rest of you out there, you
listeners out there.
Thank you so much for your timeas well, and until the next
story.
Speaker 1 (27:50):
Thank you for listening to today's episode.
If you'd like more informationon the transformative potential
of AI and automation in modernfinance, please visit
wwwdeluxecom.
Slash receivables hyphenmanagement.
Slash cash hyphen application.
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.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.