June 13, 2025 • 35 mins
With volatility at levels high and risks evolving faster than some team's action plans, investment professionals may need a new approach. In this episode of Trading Tomorrow - Navigating Trends in Capital Markets, host Jim Jockle is joined by Rajiv Bhat, CEO and Co-Founder of Martini.ai, to explore how AI-powered scenario modeling is transforming credit risk management. From real-time portfolio stress testing to the rise of agentic AI as your investment co-pilot, discover how to turn uncertainty into strategy and resilience.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Jim (00:06):
Welcome to Trading Tomorrow Navigating Trends in Capital
Markets the podcast where wedeep dive into technologies
reshaping the world of capitalmarkets.
I'm your host, jim Jockle, aveteran of the finance industry
with a passion for thecomplexities of financial
technologies and market trends.
In each episode, we'll explorethe cutting-edge trends, tools
and strategies driving today'sfinancial landscapes and paving
(00:29):
the way for the future.
With the finance industry at apivotal point, influenced by
groundbreaking innovations, it'smore crucial than ever to
understand how thesetechnological advancements
interact with market dynamics.
Interact with market dynamics.
(00:49):
Today's episode dives intourgent questions facing
investment professionals today.
How do you build a resilientportfolio where there's so much
instability in the world?
From interest rate shocks andgeopolitical tensions to the
ripple effects of economicpolicy shifts?
Today's market volatilitydemands the latest and greatest
tools.
Now the question is how can AIhelp?
Joining us to explore thequestion further is Rajiv Bhatt,
(01:12):
ceo and co-founder of MartiniAI, a leader in AI-driven credit
analytics.
Rajiv brings a unique blend oftechnical depth and
entrepreneurial experience tothe table.
He holds a PhD in theoreticalphysics, previously led data
science at AdTech, unicorn inMobi, and founded Y Combinator,
a startup acquired by Groupon.
(01:32):
At Martini AI, rajiv is helpingfinancial professionals rethink
portfolio credit riskmanagement through intelligent
automation, real-time creditsignals and scenario modeling
that captures theinterconnectedness of today's
macro forces.
Rajiv, thanks for being here.
Rajiv (01:48):
Thanks for having me here , jim, very excited, big fan of
your work and always excited tobe up close.
Jim (01:54):
Well, you know what?
Why don't we jump in?
We'll start with the basics.
So what exactly is AI-poweredscenario building and why is it
so important in today'sinvestment climate?
Rajiv (02:03):
The world's been changing much faster than it was before
and it's also much moreconnected than it ever was
before, and what that means foran investor is that cumulative
risks are now just getting moreand more pronounced.
So you need to be likesituations that you could never
have imagined can now suddenlyarise and can kind of run away
(02:25):
from you, and which is why weare very excited about how you
can grapple with that beast byusing AI.
So that's what the AI-poweredscenario builder is.
It's a way in which you canunderstand the fast-changing
world around you as it impactsyour portfolio.
So that's what we built.
So it's a very simple toolwhere you just type in what do
(02:48):
you think is going to happen, ormaybe the Panama Canal getting
a shutdown and you can see whathappens to your credit portfolio
.
It could be a portfolio ofthousands of companies and it'll
tell you what are the expectedimpacts, how does the expected
loss change, what is the valueat risk, which companies get
impacted and this is allquantitatively estimated and it
(03:10):
gives you that in seconds.
Jim (03:11):
And so how would you categorize the volatility we're
experiencing today versus pastcycles and, from your
perspective, what does thatreally mean for risk?
Rajiv (03:22):
management.
The volatility that we areseeing now is matching what we
were seeing during the GFCbefore.
It's very heightened because ofseveral things like geopolitics,
the tariff situation that'sgoing back and forth.
Folks are unable to figure outhow some of these fast-changing
events are going to impact theirhistorically slower-moving
(03:44):
portfolios, and I think a coupleof things that it's impacting
managers is.
One is in terms ofunderstanding what's the risk in
their portfolio.
Like, though the assets areilliquid, what happens to the
risk?
Because the risk is stillliquid, like you know, your
(04:05):
companies are still going to getin trouble, though you might
not be able to sell the debt.
So that's one way in which it'simpacting portfolio managers,
because it's gotten themthinking about how do I handle
the situation now.
Like you know, historically thekind of actions that they could
take were commensurate with thetimescale of, like, the
timescale of the risks that wereunfolding, but now what's
(04:27):
happened is the risks havegotten ahead of the timescales
of action, so folks have feweroptions to respond to the risks
that they have, and people arebeginning to think about how do
I operate in this new world, andthey're thinking of new tools,
new kinds of strategies, andthat's gotten everyone thinking
(04:47):
about it.
Jim (04:48):
So how can AI help managers turn that uncertainty into a
more strategic advantage?
Rajiv (04:54):
That's such a great question and I think that's a
question that's not asked enoughreally which is saying AI.
There's a fair bit ofcommentary on how AI can help
you save time, can make yourthings, can make work more
efficient, but the real questionpeople should be asking is how
can AI help me operate now?
(05:16):
And, like you rightly said, howcan AI take me to the next
level, Like how can give metools that are going to be
superior in the new world?
Two ways.
One is that AI is phenomenalfor operating with large volumes
of data, fast-moving data,making sense of what's happening
(05:39):
right now, for example, thetools that I was just mentioning
, the AI-powered scenariobuilder.
It does exactly that.
What it does is it looks at howevery company is situated in
the universe.
Every single company iscalibrated to different risk
factors.
It throws all of that togetherand it figures out OK, this is
what's going to happen to thisportfolio if this kind of event
manifests itself, and so that'ssomething that would take
(06:02):
analysts four weeks of time todo.
Now it takes a second, and sothat's the first thing, which is
AI helps you just get on top ofthe existing situation.
So now, if you're a portfoliomanager, the next thing you do
is well, that's great.
I can do this in a few seconds.
How about I do this for 100different scenarios and let's
see what happens to my portfolio, which companies are the most
(06:23):
vulnerable, which companies aregoing to thrive in these new
situations?
Let me get into that.
It makes for more well-roundeddecisioning there.
The second thing is to say canAI come up with better
strategies?
And that's the new, powerfulthing that's beginning to make
its way up, which is helpingusing AI to design policies, to
(06:47):
design strategies.
To say, in a world where howshould I design my portfolio,
how should I structure myportfolio?
What assets will make sense inthese scenarios and what should
me buy off ramps in case some ofthese situations don't pan out?
Just being able to have a morethorough plan which is in far
(07:13):
more fine structure than everbefore?
I think that's what AI lets youdo.
Jim (07:17):
I've heard excuse me, I've heard stories where sometimes
you need to blindly trust the AI.
Right when you're asking forscenarios, especially around
portfolio management, perhapsthe AI is finding different
patterns and things of thatnature and that, to the extent
you're asking these tools tomake things or the logic human,
(07:40):
understandable, sometimes thatdeprecates performance.
Have you found that to be true?
Oh, it's such a great question.
Rajiv (07:48):
If you the temptation is there, just blindly trust AI and
let it just follow the overlordand do what it's asking you to
do, and to your point.
Yes, if you do that, you willrun into some of the blind spots
that AI will have right, likethere are, but the reason is
(08:10):
basically this right, which isthat we live in a very high
dimensional world, likeeverything that we do has so
many aspects to it and it's veryfine textured and there's only
so much information you cancommunicate in a prompt.
And so an AI is going to makesome assumptions, and if you
blindly trust it, you'reassuming that its assumptions
(08:31):
are the assumptions that youhave, which is not often the
case, and so, unless you aredoing a process where you're
double-checking you'revalidating what its output you
will quickly run into asituation where what the AI is
thinking in its head isdifferent from what you have in
mind.
So you're validating what itsoutput.
You will quickly run into asituation where what the AI is
thinking in its head isdifferent from what you have in
mind and it's not going to lineup well, and to that extent, the
(08:56):
way we think about it is wethink of the tools that we are
building, particularly thescenario build, as well as the
Ironman suit.
It's not Ironman, so it'sbasically letting you do so many
things, even in our scenariobuilder.
For example, if you have ascenario where invades Taiwan,
if our AI comes back and says,you know, I think the S&P is
(09:19):
kind of going to crash about 10%because of the pressure of the
semiconductor industry, it letsthe user exercise editorial
oversight, come and move theslider down to say 80%, if
that's what they think is goingto happen, and then look at the
scenario.
So our design principle thereis have the output of AI.
(09:40):
It's always going to beverifiable and changeable so
that you can do your work fast.
But it doesn't mean that youlet go of the wheel and let AI
do the driving.
Jim (09:53):
So what does scenario building look like in practice?
Can you walk us through anexample, sure?
Rajiv (09:59):
In our case, the way we think about credit portfolios is
in three ways.
There's companies, portfoliosand sectors, and our scenario
builder operates when you have aportfolio loaded, so a
portfolio with essentially alist of companies and with some
percentage allocations to eachof them, and this is where the
(10:21):
magic starts.
So once you have the portfolioloaded up, building the scenario
for us is just typing it in, soit's like you just type in the
scenario that you think you wantto test your portfolio against.
So it could be something likeyou know what happens if tariffs
(10:42):
on imported metals goes up 100%.
Or you can say there's somegeopolitical conflict in Europe,
or you could talk about a localelection and see what happens.
So then what?
The uh, the ai, takes over fromthere, where what it does is.
It takes that scenario andbreaks it down and it says, okay
, let me see in the past what'shappened with this.
(11:03):
Like if we're going back to ourexample of .
It says, okay, let me look atsome of the bigger factors this
the s&p.
There's a price of oil thisdollars, this dollars.
There's inflation what happensto each of these typical insert
scenarios?
And then it figures out, comesback with reasonably good
estimates on how each of thosewould move.
(11:24):
And that's going back to theprevious question.
That's where the first bit ofeditorial oversight comes in.
The user can then kind ofeyeball those and say, okay,
this is in the ballpark or couldget better or worse, and make
adjustments if required.
And then what it does is ittakes that and then it applies
(11:47):
it to all the companies in theportfolio through our knowledge
graph.
So the way our system operatesis that we have we cover three
and a half million companies.
Each of those companies areembedded in a knowledge graph.
The knowledge graph isterabytes of data.
Every company is connected toover 100 other companies along
(12:08):
multiple dimensions things likeyou know, transactions, supply
chain news, investors, thingslike transactions, supply chain
news, investors.
And then we also pull in about10,000 to 20,000 risk reference
points every day.
So these could be things likeprices of bonds, prices of loans
, bankruptcies, stock prices andthe movement of those, and
(12:31):
portfolio valuations and thingslike that.
And so because of that, we get areal-time view on risk in the
market and how it pertains tocompanies in the knowledge graph
and over the past few years,like we've seen how the spreads
and probabilities of default ofcompanies in the graph has
(12:51):
responded to all those differentevents and all those different
market events.
So because of that now forevery single company in that
universe 3.5 million companieswe have a quantitative
understanding of how it moves todifferent fluctuations in the
environment.
And so now when we have thistop-level situation and it's
(13:12):
broken down into all of thesedifferent factors, we can
trickle that down to eachcompany and kind of figure out
how each company will respond tothat.
And then we roll that up at theportfolio level and figure out
what happens to each portfolioresponse to this.
All of this happens in likeseconds but makes for a very
exciting kind of use case whereyou can actually play out
(13:37):
different scenarios, kind ofdesign, and shape your portfolio
accordingly.
Jim (13:41):
Yeah, you often talk about building resilience into
investment strategies.
Rajiv (13:49):
What does that look like, and how does AI enable that?
In my mind, resilience in aninvestment strategy is its
ability to withstand shocks, andwhat level of shocks can it
withstand?
And, of course, all portfoliosare subject to some level of
weakness in the face of someshocks, but, in general, your
portfolio is better if the kindof shocks that can withstand are
(14:14):
higher, and so then that's whatI mean by resilience.
So now, if you can design yourportfolio such that you take run
stories from last three monthsthrough your portfolio and you
see what happens to which arethe weakest companies in that
portfolio, like which companiesare consistently showing up as
being impacted by these stories,then, as a portfolio manager,
(14:39):
like a simple step for you wouldbe to just drop those companies
and then automatically makeyour portfolio a little better.
The other way that you couldmake your portfolio strong, of
course, is through classichedging mechanisms, so you could
design your portfolio such thatand that's something that we
provide at Martinet as well, forlike you can see how exposed
(15:00):
your portfolio is along multipledimensions.
The utterly truth is like allof us end up becoming factor
investors, whether we like it ornot, and so you can kind of
identify which kind of factorsyou're beginning to lean towards
and compensate for that.
That makes for greater reliancetoo.
And the third way in which youcan kind of identify which kind
of factors you're beginning tolean towards and compensate for
that, that makes for greaterreliance too.
And the third way in which youcan beat reliance is this thing
that I keep talking about, whichis that just because your
(15:23):
assets are illiquid does notmean your risks are illiquid.
Your risks are just as liquid asthe stock market.
And so just being able to and ifyou try and like I've heard
often heard that why would Icare about the risk in my
private credit portfolio?
Because I can't sell it anyway,because you know have to mark
(15:46):
the market and you can hold ontoit for longer and my response
to that always is look, in aprivate credit scenario, if
you're going to be compensatingfor risk or compensating for
(16:11):
some change in your portfoliobased on actuals, it's way too
late.
Like if a company is alreadybankrupt and you're trying to
offload their debt, you're notgoing to have much luck.
On the other hand, if you areable to be on top of it on a
daily basis and make smalladjustments here and there
change the size of portfolio orchange the drawdown limits, add
in a few more covenants, youcould just pick the phone up,
(16:33):
call the borrower and ask himwhat's up.
Doing small things like thatcan suddenly change the shape
and nature of your portfolio andso, just understanding that
because the risks are liquid,you can adopt more frequent
interventions to compensate forthat.
Jim (16:51):
So what's staying on private credit?
Obviously it is booming rightnow.
You know it's the next horizon,I think, although we've seen
some shocks amid the trade warsand tariffs at this point in
time.
But I guess the question is andone of the concerns I've always
had is data availability.
You know, finding data onprivate companies can be very,
(17:14):
very challenging.
So you know, how do you manageprivate credit within your
system, and is there a role formore agentic AI to help improve
transparency in that space?
Rajiv (17:29):
Absolutely such a great question.
So I think that's been one ofthe most exciting propositions,
value propositions for Martini,which is our ability to provide
coverage for private companies.
So, of the three and a halfbillion companies we cover,
almost all of them are private,and the way we've gotten around
that is by stepping back alittle bit and saying, hey look,
(17:51):
in this day and age, everymeaningful company has to have a
significant presence, digitalpresence, and it's not just a
website, but it's just data setswhich are out there.
There's structured data setsabout, like supply chain,
employees leaving joining.
There's structured data setsaround, like you know traffic,
investments, transactions, newsand it's an elf, it's an
(18:14):
enormous treasure trove outthere.
And so then, the question thatwe wanted to answer was saying,
given that there is so muchstructured and unstructured data
out there it could be filings,it could be data sets, it could
be news what's a good way to setit up so that you can consume
everything that's out there?
It could be partial, noisy,incomplete, it could be
(18:39):
difficult to parse, and how doyou take all of that
unstructured data or structureddata, incomplete data, and
combine it with the fact thatthere is so much risk signal
available in the market by wayof all the bond prices Like if a
bond trades at 110 instead of100, it's telling you something
(19:00):
about that company.
It's also telling you somethinglike companies like it.
It's also telling you somethingabout industry in general.
How do you combine all of thatinformation and make sense of it
all in a real-time way?
And that's where I think someof our backgrounds helped a lot.
So me and my co-founder bothhave a huge background, like
(19:22):
very long experience withmachine learning and large data
sets.
Co-founder Rohit, master's fromStanford, phd from MIT, worked
as a PhD at the AI lab at MIT,won his work on knowledge graphs
, won the Test of Time Award.
Then he ran a hedge fund withthe co-founders of Akamai, doing
trading up to $3 million a day.
So significant experienceputting money where his elbows
(19:44):
are.
My background is I'm aphysicist.
I did my PhD in quantummechanics.
I'm privileged to work with acouple of Nobel laureates.
I've been working on thestartup side with big data
companies.
I'm privileged to work with acouple of Nobel laureates.
I've been working on thestartup side with big data
companies.
In my last gig I was working atIntercon.
We were processing petabytes ofdata, doing 600 billion
predictions a day, 12 billionauctions a day.
(20:05):
So we brought a lot of that in,which is basically big data,
knowledge graphs, ai, and sothen what we did is first we
built a huge knowledge graph ofcompanies connected to every
other company.
The second, we pulled in everysingle risk market or market
risk metric out there.
Then we did a little bit ofwork, which is pretty tough and
(20:30):
heavy, which is taking each ofthe signals and calibrating it
to probabilities of default andrisk of each of these companies.
And then we have these graphneural networks running on these
, on the graph, propagatingthose risk information.
So it's almost like saying youhave the surface of companies.
If there's a disturbancesomewhere, how does it propagate
?
Which companies get impacted?
(20:51):
And what we found is this we'vebeen very excited about the
results.
Like you know, all of thisstuff can be back-tested and
we've been loving theperformance of it so far.
We find over 80% of defaultshappened, the worst 20% of our
predictions and we find thatwe're able to catch signals way
ahead.
Like in a back-test with a bank, we found we were, on average,
(21:13):
seven months ahead of theirfirst sign of non-accrual.
So we've been excited about howwell it works.
And I think it works well for afew simple reasons.
One is the quantity that we'repredicting is probability of
default, which is just like isthe company getting better or
worse?
And that is a nice,well-defined physical quantity
(21:36):
to predict.
We're not trying to predictalpha, we're not trying to
predict stock price.
It's just saying is thiscompany getting better or worse?
The second thing, the reason itworks is no, it's very tough
for companies to escape thepressures impacting their sector
.
Like if Carvana is strugglingto sell cars online, it's very
(21:56):
unlikely that CarMax andCarGurus are also not feeling it
, and I think that's part of thereason.
That's something that you'llnever catch in financials.
And the other reason, a coupleof reasons we've been excited
about.
Zendaya approaches One to yourpoint about data being difficult
for private companies.
(22:18):
I think most people are talkingabout financials and over time,
what we've realized isfinancials themselves, while
they're important, don't aredifficult, because one they can
be pretty late.
It can be up to six monthsbefore you get financials for a
company quarterly financials.
Second, every term in afinancial report is defined
(22:40):
differently.
Every term might be somethingfor one company, it might be
something else for anothercompany.
They're not standardized, andso for you to understand what
each one means is a fair bit ofwork, and if you have a
portfolio of thousands ofcompanies, it's going to be
pretty difficult.
And the third reason is everyfinancial can come in at a
different point in time.
So if you're trying to sit downon a Monday morning and saying,
(23:01):
hey, how's my portfolio risklooking at, you'll probably look
at it and be like, hey, youknow what?
I have data for only like 25%of my companies and I don't have
data for the rest.
I need to call them up and pull.
That.
It's pretty heavy work and withour systems, you can actually
just, on a daily basis, monitoryour entire portfolio.
Be on top of that and the goahead I'll just make one last
(23:24):
one.
Sorry, go ahead, jim, pleaseyeah.
Jim (23:26):
No, no yeah.
Rajiv (23:27):
I was just going to say that the one thing that we're
struck is that every six to 12months, what we're seeing is
most companies are impacted bysome unexpected shock to their
margin structure or revenuestructure because something like
you know, raw metal pricesgoing up or tariffs or some
competitor coming and eatingtheir lunch and if, like we find
(23:50):
, the traditional methods aretoo slow to catch those things.
Jim (23:54):
I guess you know one of the things that just kind of popped
into my mind and you know Ilook at my iPhone, you know, and
you know it's going to tell meit's going to rain here in.
You know the.
You know it's going to tell meit's going to rain here and you
know the next 25 minutes.
But obviously that's astatistical kind of analysis
versus a longer radarperspective.
You know what is the timehorizon and or is there accuracy
(24:17):
degradation over time horizonswithin within the AI, or is it?
You know, like how?
How is the AI?
How is the AI looking at theuniverse?
Is it a smaller window?
Is it longer?
Where is the accuracy?
Rajiv (24:32):
That's such a great question and the way we do it is
at the risk of maybe givingaway too much, but the way we do
it is we actually look atall-time horizons.
So in all debt instruments youhave an yield curve and you have
like risk at pricing at like athree-month horizon or a
one-month horizon.
You're like, if you look at thetreasury yields, you have
(24:53):
yields which are like for a10-year treasury bond or a
one-year treasury bond.
It's telling you differentthings about what the market
thinks is the risk at thosehorizons.
So we actually pull that in foras many debt instruments that
we can get and we fit everythingto that.
So for any company that wecover, we actually provide our
sense for probability of defaultat like a three-month horizon
(25:16):
or a one-year horizon or afive-year horizon, and we fit
everything to that.
And so what that lets you do isanswer this question in
different ways.
So if you say, hey, give it tome at a one-year horizon, which
is what typically everybodylooks at, we can do that.
And so what that lets you do isanswer this question in
different ways.
So if you say, hey, give it tome at a one-year horizon, which
is what typically everybodylooks at, we can do that.
But if you want to look at alonger horizon, it can do that.
So, yeah, so, and it actuallyhelps, because once you are
(25:40):
willing to look at all-timehorizons, you get a lot more
signals.
If you try to constrainyourself just to data points at
a three-month or a one-yearhorizon, it's not enough.
Jim (25:55):
You know.
One of the things that you'remaking me think about is you
know you have so much marketobservable opinion coming out of
the market, whether it's bondspreads or credit default swaps,
et cetera, and you have so muchdata and information being
produced by organizations.
What is the future role of therating agencies given this new
AI-related world?
Rajiv (26:14):
Absolutely, jim.
I think what's taken everyoneby surprise is the rate and
speed with which AI can nowprocess unstructured data and
illiquid data and sparse data,which is exactly what this
entire rating industry workswith.
And from the rating agency'sperspective, I think it's
(26:37):
inevitable.
I feel because in our signalswe see that we are frequently
six to eight months ahead of anyrating changes, because just
the markets know, and just beingable to pick that up and
propagate it itself alreadymakes it very useful.
Uh, from a to your questionabout what's the role of rating
agencies going forward, I thinkit's rating agencies owe it to
(26:59):
make the latest and mostimportant ratings available to
the users or anyone who needs it.
In fact, we are beginning tothink a little more broadly and
we feel maybe we should beavailable to everybody, anyone
who is operating in the space.
Like you know, earlier data wasfree.
Now, maybe, going forward,maybe insight will be free or
(27:23):
opinions will be free, and so inthat world world, how does
things change?
Like uh, and I think, uh, if Iwere.
Uh.
So, in fact, we don't think ofourselves.
We don't think of ourselves asa rating agency.
We think of ourselves as acredit interpolation company,
and what we're saying is, giveneverything that we know right
now, what's the best estimate ofthe risk?
(27:44):
And that's the question we'retrying to answer.
We are not trying to say, hey,this is our process and this is
how we rate this company.
Instead, we're just saying,given everything that's
available, what's the best wecan do?
And for a rating agency, Ithink this is still valuable.
I think it would be especiallyfor some of the bigger investors
and bigger investments.
(28:07):
I think some of the processesthat the rating agencies have
for due diligence andunderwriting will still be very
valuable, but I think solutionslike Martini will be far more
powerful for two big parts ofthe credit workflow.
One, pre-trade due diligence,when you have hundreds of deals
are coming in and you're tryingto figure out which ones should
(28:29):
you go after.
That's one place where we thinkwe'll have big impact.
The other part that I'm franklyvery excited about is portfolio
monitoring and creditsurveillance.
Like once you have portfolios,you could just be a S&p 500
company doing business withmaybe 10 000 companies and
you're worried about youraccounts, receivable risk and
(28:52):
you need to monitor risk on adaily basis, and I think that's
where a solution like martin youit's super powerful because it
lets you do things like uh, howdo you monitor your portfolio on
a daily basis, how do you seewhat happens to your business in
the face of inevitable bigevents happening?
Or things like, how do youensure your portfolio on a daily
(29:14):
basis?
So those are kind of big usecases that I think Martini
unlocks, which is going beyondthe world of just labeling
companies for risk but sayinghow do I operate in this world?
And in fact, a couple of yearsago when I started, there used
to be a slide I used to show allthe time, which was?
You remember that incident ofthe ship getting stuck in the
(29:35):
swiss canal I think it's theever given, or something like
that and like the ship swervedjust a little bit, got stuck in
the sand and then all theseships got backed up behind it.
And then price of oil nextthing, you know, the price of
oil was shooting up just because, like, the ship had drifted
like a few feet and this is justlike a fantastic example of the
(29:55):
butterfly effect and I don'tknow how many companies got
impacted then.
And uh, you know, like you know, it would be difficult to
understand what's happening.
And what we are excited aboutnow is like you can just type
out those situations in ourscenario builder, and it'll tell
you what happens, whichcompanies get impacted, and so
(30:15):
it's just very exciting to beable to see that come to life
and being able to have tools tooperate in such a world.
Jim (30:22):
And what advice do you have for firms that are just
beginning to explore AI-poweredscenario modeling?
Rajiv (30:27):
The first thing is just getting familiar with the tools.
The second thing is building avocabulary.
I feel like a vocabulary doesnot necessarily exist for
understanding what kind ofscenario should you worry about?
Because I know that theregulators require banks to do
some of this.
They have stress tests andthose stress tests are mostly
(30:48):
around interest rate hikes ormodeling like a past scenario
saying hey, if the GFC happensagain, what happens to your
portfolio?
Do you have enough capital tocover that?
But I think scenarios can be somuch broader, risks can be so
much deeper, and just being ableto build a set of scenarios and
(31:08):
responses and action plansaround each of those will just
make firms much more resilient,and I think so.
That would be my first piece ofadvice, which is just saying
what does your playbook looklike?
What risks are you set up for,what can you handle, what can't
you handle and which of theseare consequential.
And the rate at which thesescenarios are unfolding now is
(31:32):
mind-boggling.
And like I was in New York afew weeks ago, the week of the
tariffs, every single day thetariffs changed and like for the
traders on the floor, I wastalking to someone at a japanese
bank and she just threw herhands up in the air saying, hey,
I just don't know how torespond to any of this because,
(31:53):
uh, things are changing wayfaster than I can, uh, uh,
anticipate and or I can plan for.
So, given that these, now thesescenarios have just become, you
know, on a daily or hourlybasis, they're unfolding at that
rate it's just so much moreimportant to have a good handle
(32:14):
on what kind of scenarios couldcome up.
Two is what's your responseplans?
And three is what kind oflevers do you actually have?
Do you have enough levers?
And I think that's the thirdpart.
The third part is where I thinka lot of users will end up
introspecting a lot, becausewhat good is being able to
(32:37):
understand scenarios if youcan't act on it and just being
able to build the levers tooperate in this?
Jim (32:44):
that's the part they need to focus on so unfortunately
we've made it to the lastquestion of the podcast.
We call it the trend drop.
Uh, it's like a desert islandquestion.
And if you can only watch ortrack one trend in ai and
investing, what would that be?
Rajiv (32:58):
so well, it's the obvious one and but it's uh obvious and
very powerful it is the trackof, like you know, agentic AI in
helping making your decisioningand understanding what's
happening and shaping yourportfolio and your decisions.
And the way I think of it ishow can agentic AI be a co-pilot
for you?
(33:18):
So you're already seeing thisin code, in coding.
You have companies like Cursorwho act like engineers for you.
You can just sit and talk to itall day long and it generates
phenomenal pieces of code thatyou can use to get.
That does work for you.
So what's the analogy in finance?
So it's going to be anassistant which sits with you
(33:40):
and helps you think through allthese huge data sets, helps you
think through all of thescenarios, helps you think
through decisioning and where itwill go from.
There is, once you have thedecisioning, the next thing is
the actions, and so AgentDK aremaking actions for you, like
(34:01):
deciding limits, decidinginvestment decisions and then
eventually shaping policies.
So then you could be aninvestment manager who's
spending time thinking aboutwhat should my portfolio kind of
roughly look like to meet theinvestment goals of my LPs and
then working with your agent tocraft that policy and execute on
(34:22):
that policy.
So I think that is becomingreal much, much faster than
people anticipate, and I wouldkeep an eye open for that.
Jim (34:31):
Well, rajiv, I want to thank you so much for your time,
your insights and theinformation you shared with all
of us.
Thank you so much.
Rajiv (34:38):
Thanks, Jim.
All the questions are fantastic.
I absolutely love being on here.
Absolutely love being on here.
Jim (34:49):
Thanks so much for listening to today's episode and
if you're enjoying TradingTomorrow, navigating trends in
capital markets, be sure to like, subscribe and share and we'll
see you next time you.
Welcome to Trading Tomorrow Navigating Trends in Capital
Markets the podcast where wedeep dive into technologies
reshaping the world of capitalmarkets.
I'm your host, jim Jockle, aveteran of the finance industry
with a passion for thecomplexities of financial
technologies and market trends.
In each episode, we'll explorethe cutting-edge trends, tools
and strategies driving today'sfinancial landscapes and paving
(00:29):
the way for the future.
With the finance industry at apivotal point, influenced by
groundbreaking innovations, it'smore crucial than ever to
understand how thesetechnological advancements
interact with market dynamics.
Interact with market dynamics.
(00:49):
Today's episode dives intourgent questions facing
investment professionals today.
How do you build a resilientportfolio where there's so much
instability in the world?
From interest rate shocks andgeopolitical tensions to the
ripple effects of economicpolicy shifts?
Today's market volatilitydemands the latest and greatest
tools.
Now the question is how can AIhelp?
Joining us to explore thequestion further is Rajiv Bhatt,
(01:12):
ceo and co-founder of MartiniAI, a leader in AI-driven credit
analytics.
Rajiv brings a unique blend oftechnical depth and
entrepreneurial experience tothe table.
He holds a PhD in theoreticalphysics, previously led data
science at AdTech, unicorn inMobi, and founded Y Combinator,
a startup acquired by Groupon.
(01:32):
At Martini AI, rajiv is helpingfinancial professionals rethink
portfolio credit riskmanagement through intelligent
automation, real-time creditsignals and scenario modeling
that captures theinterconnectedness of today's
macro forces.
Rajiv, thanks for being here.
Rajiv (01:48):
Thanks for having me here , jim, very excited, big fan of
your work and always excited tobe up close.
Jim (01:54):
Well, you know what?
Why don't we jump in?
We'll start with the basics.
So what exactly is AI-poweredscenario building and why is it
so important in today'sinvestment climate?
Rajiv (02:03):
The world's been changing much faster than it was before
and it's also much moreconnected than it ever was
before, and what that means foran investor is that cumulative
risks are now just getting moreand more pronounced.
So you need to be likesituations that you could never
have imagined can now suddenlyarise and can kind of run away
(02:25):
from you, and which is why weare very excited about how you
can grapple with that beast byusing AI.
So that's what the AI-poweredscenario builder is.
It's a way in which you canunderstand the fast-changing
world around you as it impactsyour portfolio.
So that's what we built.
So it's a very simple toolwhere you just type in what do
(02:48):
you think is going to happen, ormaybe the Panama Canal getting
a shutdown and you can see whathappens to your credit portfolio
.
It could be a portfolio ofthousands of companies and it'll
tell you what are the expectedimpacts, how does the expected
loss change, what is the valueat risk, which companies get
impacted and this is allquantitatively estimated and it
(03:10):
gives you that in seconds.
Jim (03:11):
And so how would you categorize the volatility we're
experiencing today versus pastcycles and, from your
perspective, what does thatreally mean for risk?
Rajiv (03:22):
management.
The volatility that we areseeing now is matching what we
were seeing during the GFCbefore.
It's very heightened because ofseveral things like geopolitics,
the tariff situation that'sgoing back and forth.
Folks are unable to figure outhow some of these fast-changing
events are going to impact theirhistorically slower-moving
(03:44):
portfolios, and I think a coupleof things that it's impacting
managers is.
One is in terms ofunderstanding what's the risk in
their portfolio.
Like, though the assets areilliquid, what happens to the
risk?
Because the risk is stillliquid, like you know, your
(04:05):
companies are still going to getin trouble, though you might
not be able to sell the debt.
So that's one way in which it'simpacting portfolio managers,
because it's gotten themthinking about how do I handle
the situation now.
Like you know, historically thekind of actions that they could
take were commensurate with thetimescale of, like, the
timescale of the risks that wereunfolding, but now what's
(04:27):
happened is the risks havegotten ahead of the timescales
of action, so folks have feweroptions to respond to the risks
that they have, and people arebeginning to think about how do
I operate in this new world, andthey're thinking of new tools,
new kinds of strategies, andthat's gotten everyone thinking
(04:47):
about it.
Jim (04:48):
So how can AI help managers turn that uncertainty into a
more strategic advantage?
Rajiv (04:54):
That's such a great question and I think that's a
question that's not asked enoughreally which is saying AI.
There's a fair bit ofcommentary on how AI can help
you save time, can make yourthings, can make work more
efficient, but the real questionpeople should be asking is how
can AI help me operate now?
(05:16):
And, like you rightly said, howcan AI take me to the next
level, Like how can give metools that are going to be
superior in the new world?
Two ways.
One is that AI is phenomenalfor operating with large volumes
of data, fast-moving data,making sense of what's happening
(05:39):
right now, for example, thetools that I was just mentioning
, the AI-powered scenariobuilder.
It does exactly that.
What it does is it looks at howevery company is situated in
the universe.
Every single company iscalibrated to different risk
factors.
It throws all of that togetherand it figures out OK, this is
what's going to happen to thisportfolio if this kind of event
manifests itself, and so that'ssomething that would take
(06:02):
analysts four weeks of time todo.
Now it takes a second, and sothat's the first thing, which is
AI helps you just get on top ofthe existing situation.
So now, if you're a portfoliomanager, the next thing you do
is well, that's great.
I can do this in a few seconds.
How about I do this for 100different scenarios and let's
see what happens to my portfolio, which companies are the most
(06:23):
vulnerable, which companies aregoing to thrive in these new
situations?
Let me get into that.
It makes for more well-roundeddecisioning there.
The second thing is to say canAI come up with better
strategies?
And that's the new, powerfulthing that's beginning to make
its way up, which is helpingusing AI to design policies, to
(06:47):
design strategies.
To say, in a world where howshould I design my portfolio,
how should I structure myportfolio?
What assets will make sense inthese scenarios and what should
me buy off ramps in case some ofthese situations don't pan out?
Just being able to have a morethorough plan which is in far
(07:13):
more fine structure than everbefore?
I think that's what AI lets youdo.
Jim (07:17):
I've heard excuse me, I've heard stories where sometimes
you need to blindly trust the AI.
Right when you're asking forscenarios, especially around
portfolio management, perhapsthe AI is finding different
patterns and things of thatnature and that, to the extent
you're asking these tools tomake things or the logic human,
(07:40):
understandable, sometimes thatdeprecates performance.
Have you found that to be true?
Oh, it's such a great question.
Rajiv (07:48):
If you the temptation is there, just blindly trust AI and
let it just follow the overlordand do what it's asking you to
do, and to your point.
Yes, if you do that, you willrun into some of the blind spots
that AI will have right, likethere are, but the reason is
(08:10):
basically this right, which isthat we live in a very high
dimensional world, likeeverything that we do has so
many aspects to it and it's veryfine textured and there's only
so much information you cancommunicate in a prompt.
And so an AI is going to makesome assumptions, and if you
blindly trust it, you'reassuming that its assumptions
(08:31):
are the assumptions that youhave, which is not often the
case, and so, unless you aredoing a process where you're
double-checking you'revalidating what its output you
will quickly run into asituation where what the AI is
thinking in its head isdifferent from what you have in
mind.
So you're validating what itsoutput.
You will quickly run into asituation where what the AI is
thinking in its head isdifferent from what you have in
mind and it's not going to lineup well, and to that extent, the
(08:56):
way we think about it is wethink of the tools that we are
building, particularly thescenario build, as well as the
Ironman suit.
It's not Ironman, so it'sbasically letting you do so many
things, even in our scenariobuilder.
For example, if you have ascenario where invades Taiwan,
if our AI comes back and says,you know, I think the S&P is
(09:19):
kind of going to crash about 10%because of the pressure of the
semiconductor industry, it letsthe user exercise editorial
oversight, come and move theslider down to say 80%, if
that's what they think is goingto happen, and then look at the
scenario.
So our design principle thereis have the output of AI.
(09:40):
It's always going to beverifiable and changeable so
that you can do your work fast.
But it doesn't mean that youlet go of the wheel and let AI
do the driving.
Jim (09:53):
So what does scenario building look like in practice?
Can you walk us through anexample, sure?
Rajiv (09:59):
In our case, the way we think about credit portfolios is
in three ways.
There's companies, portfoliosand sectors, and our scenario
builder operates when you have aportfolio loaded, so a
portfolio with essentially alist of companies and with some
percentage allocations to eachof them, and this is where the
(10:21):
magic starts.
So once you have the portfolioloaded up, building the scenario
for us is just typing it in, soit's like you just type in the
scenario that you think you wantto test your portfolio against.
So it could be something likeyou know what happens if tariffs
(10:42):
on imported metals goes up 100%.
Or you can say there's somegeopolitical conflict in Europe,
or you could talk about a localelection and see what happens.
So then what?
The uh, the ai, takes over fromthere, where what it does is.
It takes that scenario andbreaks it down and it says, okay
, let me see in the past what'shappened with this.
(11:03):
Like if we're going back to ourexample of .
It says, okay, let me look atsome of the bigger factors this
the s&p.
There's a price of oil thisdollars, this dollars.
There's inflation what happensto each of these typical insert
scenarios?
And then it figures out, comesback with reasonably good
estimates on how each of thosewould move.
(11:24):
And that's going back to theprevious question.
That's where the first bit ofeditorial oversight comes in.
The user can then kind ofeyeball those and say, okay,
this is in the ballpark or couldget better or worse, and make
adjustments if required.
And then what it does is ittakes that and then it applies
(11:47):
it to all the companies in theportfolio through our knowledge
graph.
So the way our system operatesis that we have we cover three
and a half million companies.
Each of those companies areembedded in a knowledge graph.
The knowledge graph isterabytes of data.
Every company is connected toover 100 other companies along
(12:08):
multiple dimensions things likeyou know, transactions, supply
chain news, investors, thingslike transactions, supply chain
news, investors.
And then we also pull in about10,000 to 20,000 risk reference
points every day.
So these could be things likeprices of bonds, prices of loans
, bankruptcies, stock prices andthe movement of those, and
(12:31):
portfolio valuations and thingslike that.
And so because of that, we get areal-time view on risk in the
market and how it pertains tocompanies in the knowledge graph
and over the past few years,like we've seen how the spreads
and probabilities of default ofcompanies in the graph has
(12:51):
responded to all those differentevents and all those different
market events.
So because of that now forevery single company in that
universe 3.5 million companieswe have a quantitative
understanding of how it moves todifferent fluctuations in the
environment.
And so now when we have thistop-level situation and it's
(13:12):
broken down into all of thesedifferent factors, we can
trickle that down to eachcompany and kind of figure out
how each company will respond tothat.
And then we roll that up at theportfolio level and figure out
what happens to each portfolioresponse to this.
All of this happens in likeseconds but makes for a very
exciting kind of use case whereyou can actually play out
(13:37):
different scenarios, kind ofdesign, and shape your portfolio
accordingly.
Jim (13:41):
Yeah, you often talk about building resilience into
investment strategies.
Rajiv (13:49):
What does that look like, and how does AI enable that?
In my mind, resilience in aninvestment strategy is its
ability to withstand shocks, andwhat level of shocks can it
withstand?
And, of course, all portfoliosare subject to some level of
weakness in the face of someshocks, but, in general, your
portfolio is better if the kindof shocks that can withstand are
(14:14):
higher, and so then that's whatI mean by resilience.
So now, if you can design yourportfolio such that you take run
stories from last three monthsthrough your portfolio and you
see what happens to which arethe weakest companies in that
portfolio, like which companiesare consistently showing up as
being impacted by these stories,then, as a portfolio manager,
(14:39):
like a simple step for you wouldbe to just drop those companies
and then automatically makeyour portfolio a little better.
The other way that you couldmake your portfolio strong, of
course, is through classichedging mechanisms, so you could
design your portfolio such thatand that's something that we
provide at Martinet as well, forlike you can see how exposed
(15:00):
your portfolio is along multipledimensions.
The utterly truth is like allof us end up becoming factor
investors, whether we like it ornot, and so you can kind of
identify which kind of factorsyou're beginning to lean towards
and compensate for that.
That makes for greater reliancetoo.
And the third way in which youcan kind of identify which kind
of factors you're beginning tolean towards and compensate for
that, that makes for greaterreliance too.
And the third way in which youcan beat reliance is this thing
that I keep talking about, whichis that just because your
(15:23):
assets are illiquid does notmean your risks are illiquid.
Your risks are just as liquid asthe stock market.
And so just being able to and ifyou try and like I've heard
often heard that why would Icare about the risk in my
private credit portfolio?
Because I can't sell it anyway,because you know have to mark
(15:46):
the market and you can hold ontoit for longer and my response
to that always is look, in aprivate credit scenario, if
you're going to be compensatingfor risk or compensating for
(16:11):
some change in your portfoliobased on actuals, it's way too
late.
Like if a company is alreadybankrupt and you're trying to
offload their debt, you're notgoing to have much luck.
On the other hand, if you areable to be on top of it on a
daily basis and make smalladjustments here and there
change the size of portfolio orchange the drawdown limits, add
in a few more covenants, youcould just pick the phone up,
(16:33):
call the borrower and ask himwhat's up.
Doing small things like thatcan suddenly change the shape
and nature of your portfolio andso, just understanding that
because the risks are liquid,you can adopt more frequent
interventions to compensate forthat.
Jim (16:51):
So what's staying on private credit?
Obviously it is booming rightnow.
You know it's the next horizon,I think, although we've seen
some shocks amid the trade warsand tariffs at this point in
time.
But I guess the question is andone of the concerns I've always
had is data availability.
You know, finding data onprivate companies can be very,
(17:14):
very challenging.
So you know, how do you manageprivate credit within your
system, and is there a role formore agentic AI to help improve
transparency in that space?
Rajiv (17:29):
Absolutely such a great question.
So I think that's been one ofthe most exciting propositions,
value propositions for Martini,which is our ability to provide
coverage for private companies.
So, of the three and a halfbillion companies we cover,
almost all of them are private,and the way we've gotten around
that is by stepping back alittle bit and saying, hey look,
(17:51):
in this day and age, everymeaningful company has to have a
significant presence, digitalpresence, and it's not just a
website, but it's just data setswhich are out there.
There's structured data setsabout, like supply chain,
employees leaving joining.
There's structured data setsaround, like you know traffic,
investments, transactions, newsand it's an elf, it's an
(18:14):
enormous treasure trove outthere.
And so then, the question thatwe wanted to answer was saying,
given that there is so muchstructured and unstructured data
out there it could be filings,it could be data sets, it could
be news what's a good way to setit up so that you can consume
everything that's out there?
It could be partial, noisy,incomplete, it could be
(18:39):
difficult to parse, and how doyou take all of that
unstructured data or structureddata, incomplete data, and
combine it with the fact thatthere is so much risk signal
available in the market by wayof all the bond prices Like if a
bond trades at 110 instead of100, it's telling you something
(19:00):
about that company.
It's also telling you somethinglike companies like it.
It's also telling you somethingabout industry in general.
How do you combine all of thatinformation and make sense of it
all in a real-time way?
And that's where I think someof our backgrounds helped a lot.
So me and my co-founder bothhave a huge background, like
(19:22):
very long experience withmachine learning and large data
sets.
Co-founder Rohit, master's fromStanford, phd from MIT, worked
as a PhD at the AI lab at MIT,won his work on knowledge graphs
, won the Test of Time Award.
Then he ran a hedge fund withthe co-founders of Akamai, doing
trading up to $3 million a day.
So significant experienceputting money where his elbows
(19:44):
are.
My background is I'm aphysicist.
I did my PhD in quantummechanics.
I'm privileged to work with acouple of Nobel laureates.
I've been working on thestartup side with big data
companies.
I'm privileged to work with acouple of Nobel laureates.
I've been working on thestartup side with big data
companies.
In my last gig I was working atIntercon.
We were processing petabytes ofdata, doing 600 billion
predictions a day, 12 billionauctions a day.
(20:05):
So we brought a lot of that in,which is basically big data,
knowledge graphs, ai, and sothen what we did is first we
built a huge knowledge graph ofcompanies connected to every
other company.
The second, we pulled in everysingle risk market or market
risk metric out there.
Then we did a little bit ofwork, which is pretty tough and
(20:30):
heavy, which is taking each ofthe signals and calibrating it
to probabilities of default andrisk of each of these companies.
And then we have these graphneural networks running on these
, on the graph, propagatingthose risk information.
So it's almost like saying youhave the surface of companies.
If there's a disturbancesomewhere, how does it propagate
?
Which companies get impacted?
(20:51):
And what we found is this we'vebeen very excited about the
results.
Like you know, all of thisstuff can be back-tested and
we've been loving theperformance of it so far.
We find over 80% of defaultshappened, the worst 20% of our
predictions and we find thatwe're able to catch signals way
ahead.
Like in a back-test with a bank, we found we were, on average,
(21:13):
seven months ahead of theirfirst sign of non-accrual.
So we've been excited about howwell it works.
And I think it works well for afew simple reasons.
One is the quantity that we'repredicting is probability of
default, which is just like isthe company getting better or
worse?
And that is a nice,well-defined physical quantity
(21:36):
to predict.
We're not trying to predictalpha, we're not trying to
predict stock price.
It's just saying is thiscompany getting better or worse?
The second thing, the reason itworks is no, it's very tough
for companies to escape thepressures impacting their sector
.
Like if Carvana is strugglingto sell cars online, it's very
(21:56):
unlikely that CarMax andCarGurus are also not feeling it
, and I think that's part of thereason.
That's something that you'llnever catch in financials.
And the other reason, a coupleof reasons we've been excited
about.
Zendaya approaches One to yourpoint about data being difficult
for private companies.
(22:18):
I think most people are talkingabout financials and over time,
what we've realized isfinancials themselves, while
they're important, don't aredifficult, because one they can
be pretty late.
It can be up to six monthsbefore you get financials for a
company quarterly financials.
Second, every term in afinancial report is defined
(22:40):
differently.
Every term might be somethingfor one company, it might be
something else for anothercompany.
They're not standardized, andso for you to understand what
each one means is a fair bit ofwork, and if you have a
portfolio of thousands ofcompanies, it's going to be
pretty difficult.
And the third reason is everyfinancial can come in at a
different point in time.
So if you're trying to sit downon a Monday morning and saying,
(23:01):
hey, how's my portfolio risklooking at, you'll probably look
at it and be like, hey, youknow what?
I have data for only like 25%of my companies and I don't have
data for the rest.
I need to call them up and pull.
That.
It's pretty heavy work and withour systems, you can actually
just, on a daily basis, monitoryour entire portfolio.
Be on top of that and the goahead I'll just make one last
(23:24):
one.
Sorry, go ahead, jim, pleaseyeah.
Jim (23:26):
No, no yeah.
Rajiv (23:27):
I was just going to say that the one thing that we're
struck is that every six to 12months, what we're seeing is
most companies are impacted bysome unexpected shock to their
margin structure or revenuestructure because something like
you know, raw metal pricesgoing up or tariffs or some
competitor coming and eatingtheir lunch and if, like we find
(23:50):
, the traditional methods aretoo slow to catch those things.
Jim (23:54):
I guess you know one of the things that just kind of popped
into my mind and you know Ilook at my iPhone, you know, and
you know it's going to tell meit's going to rain here in.
You know the.
You know it's going to tell meit's going to rain here and you
know the next 25 minutes.
But obviously that's astatistical kind of analysis
versus a longer radarperspective.
You know what is the timehorizon and or is there accuracy
(24:17):
degradation over time horizonswithin within the AI, or is it?
You know, like how?
How is the AI?
How is the AI looking at theuniverse?
Is it a smaller window?
Is it longer?
Where is the accuracy?
Rajiv (24:32):
That's such a great question and the way we do it is
at the risk of maybe givingaway too much, but the way we do
it is we actually look atall-time horizons.
So in all debt instruments youhave an yield curve and you have
like risk at pricing at like athree-month horizon or a
one-month horizon.
You're like, if you look at thetreasury yields, you have
(24:53):
yields which are like for a10-year treasury bond or a
one-year treasury bond.
It's telling you differentthings about what the market
thinks is the risk at thosehorizons.
So we actually pull that in foras many debt instruments that
we can get and we fit everythingto that.
So for any company that wecover, we actually provide our
sense for probability of defaultat like a three-month horizon
(25:16):
or a one-year horizon or afive-year horizon, and we fit
everything to that.
And so what that lets you do isanswer this question in
different ways.
So if you say, hey, give it tome at a one-year horizon, which
is what typically everybodylooks at, we can do that.
And so what that lets you do isanswer this question in
different ways.
So if you say, hey, give it tome at a one-year horizon, which
is what typically everybodylooks at, we can do that.
But if you want to look at alonger horizon, it can do that.
So, yeah, so, and it actuallyhelps, because once you are
(25:40):
willing to look at all-timehorizons, you get a lot more
signals.
If you try to constrainyourself just to data points at
a three-month or a one-yearhorizon, it's not enough.
Jim (25:55):
You know.
One of the things that you'remaking me think about is you
know you have so much marketobservable opinion coming out of
the market, whether it's bondspreads or credit default swaps,
et cetera, and you have so muchdata and information being
produced by organizations.
What is the future role of therating agencies given this new
AI-related world?
Rajiv (26:14):
Absolutely, jim.
I think what's taken everyoneby surprise is the rate and
speed with which AI can nowprocess unstructured data and
illiquid data and sparse data,which is exactly what this
entire rating industry workswith.
And from the rating agency'sperspective, I think it's
(26:37):
inevitable.
I feel because in our signalswe see that we are frequently
six to eight months ahead of anyrating changes, because just
the markets know, and just beingable to pick that up and
propagate it itself alreadymakes it very useful.
Uh, from a to your questionabout what's the role of rating
agencies going forward, I thinkit's rating agencies owe it to
(26:59):
make the latest and mostimportant ratings available to
the users or anyone who needs it.
In fact, we are beginning tothink a little more broadly and
we feel maybe we should beavailable to everybody, anyone
who is operating in the space.
Like you know, earlier data wasfree.
Now, maybe, going forward,maybe insight will be free or
(27:23):
opinions will be free, and so inthat world world, how does
things change?
Like uh, and I think, uh, if Iwere.
Uh.
So, in fact, we don't think ofourselves.
We don't think of ourselves asa rating agency.
We think of ourselves as acredit interpolation company,
and what we're saying is, giveneverything that we know right
now, what's the best estimate ofthe risk?
(27:44):
And that's the question we'retrying to answer.
We are not trying to say, hey,this is our process and this is
how we rate this company.
Instead, we're just saying,given everything that's
available, what's the best wecan do?
And for a rating agency, Ithink this is still valuable.
I think it would be especiallyfor some of the bigger investors
and bigger investments.
(28:07):
I think some of the processesthat the rating agencies have
for due diligence andunderwriting will still be very
valuable, but I think solutionslike Martini will be far more
powerful for two big parts ofthe credit workflow.
One, pre-trade due diligence,when you have hundreds of deals
are coming in and you're tryingto figure out which ones should
(28:29):
you go after.
That's one place where we thinkwe'll have big impact.
The other part that I'm franklyvery excited about is portfolio
monitoring and creditsurveillance.
Like once you have portfolios,you could just be a S&p 500
company doing business withmaybe 10 000 companies and
you're worried about youraccounts, receivable risk and
(28:52):
you need to monitor risk on adaily basis, and I think that's
where a solution like martin youit's super powerful because it
lets you do things like uh, howdo you monitor your portfolio on
a daily basis, how do you seewhat happens to your business in
the face of inevitable bigevents happening?
Or things like, how do youensure your portfolio on a daily
(29:14):
basis?
So those are kind of big usecases that I think Martini
unlocks, which is going beyondthe world of just labeling
companies for risk but sayinghow do I operate in this world?
And in fact, a couple of yearsago when I started, there used
to be a slide I used to show allthe time, which was?
You remember that incident ofthe ship getting stuck in the
(29:35):
swiss canal I think it's theever given, or something like
that and like the ship swervedjust a little bit, got stuck in
the sand and then all theseships got backed up behind it.
And then price of oil nextthing, you know, the price of
oil was shooting up just because, like, the ship had drifted
like a few feet and this is justlike a fantastic example of the
(29:55):
butterfly effect and I don'tknow how many companies got
impacted then.
And uh, you know, like you know, it would be difficult to
understand what's happening.
And what we are excited aboutnow is like you can just type
out those situations in ourscenario builder, and it'll tell
you what happens, whichcompanies get impacted, and so
(30:15):
it's just very exciting to beable to see that come to life
and being able to have tools tooperate in such a world.
Jim (30:22):
And what advice do you have for firms that are just
beginning to explore AI-poweredscenario modeling?
Rajiv (30:27):
The first thing is just getting familiar with the tools.
The second thing is building avocabulary.
I feel like a vocabulary doesnot necessarily exist for
understanding what kind ofscenario should you worry about?
Because I know that theregulators require banks to do
some of this.
They have stress tests andthose stress tests are mostly
(30:48):
around interest rate hikes ormodeling like a past scenario
saying hey, if the GFC happensagain, what happens to your
portfolio?
Do you have enough capital tocover that?
But I think scenarios can be somuch broader, risks can be so
much deeper, and just being ableto build a set of scenarios and
(31:08):
responses and action plansaround each of those will just
make firms much more resilient,and I think so.
That would be my first piece ofadvice, which is just saying
what does your playbook looklike?
What risks are you set up for,what can you handle, what can't
you handle and which of theseare consequential.
And the rate at which thesescenarios are unfolding now is
(31:32):
mind-boggling.
And like I was in New York afew weeks ago, the week of the
tariffs, every single day thetariffs changed and like for the
traders on the floor, I wastalking to someone at a japanese
bank and she just threw herhands up in the air saying, hey,
I just don't know how torespond to any of this because,
(31:53):
uh, things are changing wayfaster than I can, uh, uh,
anticipate and or I can plan for.
So, given that these, now thesescenarios have just become, you
know, on a daily or hourlybasis, they're unfolding at that
rate it's just so much moreimportant to have a good handle
(32:14):
on what kind of scenarios couldcome up.
Two is what's your responseplans?
And three is what kind oflevers do you actually have?
Do you have enough levers?
And I think that's the thirdpart.
The third part is where I thinka lot of users will end up
introspecting a lot, becausewhat good is being able to
(32:37):
understand scenarios if youcan't act on it and just being
able to build the levers tooperate in this?
Jim (32:44):
that's the part they need to focus on so unfortunately
we've made it to the lastquestion of the podcast.
We call it the trend drop.
Uh, it's like a desert islandquestion.
And if you can only watch ortrack one trend in ai and
investing, what would that be?
Rajiv (32:58):
so well, it's the obvious one and but it's uh obvious and
very powerful it is the trackof, like you know, agentic AI in
helping making your decisioningand understanding what's
happening and shaping yourportfolio and your decisions.
And the way I think of it ishow can agentic AI be a co-pilot
for you?
(33:18):
So you're already seeing thisin code, in coding.
You have companies like Cursorwho act like engineers for you.
You can just sit and talk to itall day long and it generates
phenomenal pieces of code thatyou can use to get.
That does work for you.
So what's the analogy in finance?
So it's going to be anassistant which sits with you
(33:40):
and helps you think through allthese huge data sets, helps you
think through all of thescenarios, helps you think
through decisioning and where itwill go from.
There is, once you have thedecisioning, the next thing is
the actions, and so AgentDK aremaking actions for you, like
(34:01):
deciding limits, decidinginvestment decisions and then
eventually shaping policies.
So then you could be aninvestment manager who's
spending time thinking aboutwhat should my portfolio kind of
roughly look like to meet theinvestment goals of my LPs and
then working with your agent tocraft that policy and execute on
(34:22):
that policy.
So I think that is becomingreal much, much faster than
people anticipate, and I wouldkeep an eye open for that.
Jim (34:31):
Well, rajiv, I want to thank you so much for your time,
your insights and theinformation you shared with all
of us.
Thank you so much.
Rajiv (34:38):
Thanks, Jim.
All the questions are fantastic.
I absolutely love being on here.
Absolutely love being on here.
Jim (34:49):
Thanks so much for listening to today's episode and
if you're enjoying TradingTomorrow, navigating trends in
capital markets, be sure to like, subscribe and share and we'll
see you next time you.
Popular Podcasts
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.