November 27, 2024 • 43 mins
The importance of green financing has grown in recent years and analyzing its impact has been a hurdle for many investors, one that may potentially be resolved with advances in artificial intelligence. In today’s episode Romina Reversi, Credit Agricole CIB’s head of Sustainable Investment Banking, Americas, Marcel Bock and Andreas Hoelzl, co-founders of Mensura AI, join Chris Ratti, Bloomberg Intelligence’s senior ESG credit strategist, to discuss how quantifying the impact of sustainable debt, using proceeds for green or social projects or linked to social and environmental KPIs, could be done using AI. Sustainable-finance professionals are faced with a myriad of challenges and AI can assist with automated workflows to decipher and improve comparability of the impact and allocation data points.
For more insights, register here for BI ESG’s Dec. 11 conference.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:09):
ESG has become established as a key business theme as
companies and investors seek to navigate the climate crisis, energy transition,
social mega trends, mouting, regulatory attention and pressures from other stakeholders.
The rapidly evolving landscape has become inundated with acronyms, buzzwords
and lingo, and we aim to break these down with
industry experts. Welcome to ESG Currents, your guide to navigating
(00:33):
the evolving ESD space, one topic at a time. Brought
to you by Bloomberg Intelligence, which is part of Bloomberg's
research department, with five hundred analysts and strategists working across
all major world markets. Our coverage includes over two thousand
equities and credits, as well as well as outlooks on
more than ninety industries and one hundred market indices, currencies
(00:55):
and commodities. I'm Chris Rady, senior ESG credit strategists and
your hosts for this episode. Today, we're talking with Romina Reversi,
ben engine director and head of Sustainable Investment Banking America's
at Credit Agricol CIB, along with Marcel Bach and Andreas Holzel,
who are co founders of mensura AI. Welcome everyone, and
(01:15):
thank you for joining us today.
Speaker 2 (01:17):
Thank you, Chris, thank you.
Speaker 1 (01:19):
Today we're gonna kind of jump in and just discuss
how artificial intelligence can be utilized by stakeholders in sustainable financing.
And for me, when I think about the sustainable debt markets,
I'm looking at you labeled use of proceeds bonds that
have seen continued growth. There's over five trillion, depending on
which numbers you look at now in bonds outstanding, with
(01:41):
the most common being green bonds, followed by social and
sustainability bonds. There's also other various forms of investments out
there like blue bond, sustainability link bonds out with bonds debt,
for nature swaps. The list goes on. So when I
then turn and look at our official intelligence, I think
more about generative AI and content creation notes on complex topics,
(02:07):
you know, helping you summarize very complex things, or machine
learning which might help you look at large data sets
and analyze it in a different way. So I think,
before we get started, maybe it'd be good if maybe
Romina can define sustainable finance in terms of the context
of today's discussion, and then maybe Marcel Or Andres can
(02:28):
enlighten us with types of AI that you guys are
most focused on.
Speaker 3 (02:33):
Thanks Chris, it's a pleasure to be here today. I'm
Romina University heading up sustainable banking for the America's at
Credit I recall CIB so our team is involved in
all the structuring of ESG related debt. And to answer
your question, Chris, how would I define sustainable finance? At
(02:53):
its core, it is the structuring of labeling of ESG
related products, both on the bond and the loan side.
So that includes green bonds, sustainability linked bonds, sustainability linked
r cfs, all of these products that are issued and
borrowed across a variety of various companies and sovereigns that
(03:18):
either use proceeds for green and or social projects as
well as have a link to some form of environmental
and social KPI.
Speaker 1 (03:29):
Right, and then over to Marcela or Andres whatever wants
to kind of let us know how you're looking at
AI in terms of Mansirah.
Speaker 4 (03:38):
Yeah, thank you Chris, and thanks again for having us on.
Great to be here. So this is myself from Mansura.
Andreas and I have been building Mansua on the belief
that there's all of these great technologies out there and
all of the superpowers that AI can help you get today,
but we want to really tailor this to the context
(03:59):
of sustainable finance, and we want to in the end
equip sustainable finance professionals with the superpowers. The belief is
that ultimately, if we can, if we can enable sustainable
finance professionals, that will then increase the flow of capital
into green investments, green bonds, and accelerate the green transition.
(04:19):
So this is something that we're very excited about. You know,
on our end, we've observed a couple of challenges that
sustainable finance professionals very often face. There's a lot of
kind of manual and time consuming analysis which are very repetitive,
and you know those are those are happening by analysts
(04:42):
that are working in Romina's team, for example, but those
are happening really across the board. And there's a lot
of disparate data sources that are not really being aggregated,
and so folks have kind of trouble bringing us all
together and making sense of it. And then there's the
kind of current AYE models that are very powerful, but
(05:03):
they're not really tailored to the context of sustainability. So
this is where we come in. So we're building a
platform that allows those professionals to perform automated analytics workflows,
to have a one stop shop for all of the
different information on the issue on bond level that is
out there, and really kind of a co pilot by
their side that helps them navigate through some of these
(05:26):
data sources and analytics workflows.
Speaker 1 (05:29):
Yeah, I think there's a lot more in terms of
what a I could do, and I think we're gonna,
obviously in today's discussion, just talk about a lot of
those a little more with both of you, all of you. Sorry,
but let me turn back to Romina for a second
and just maybe we could talk about some of the
(05:50):
opportunities or challenges that green financing is currently facing.
Speaker 3 (05:54):
So at a score, green financing is a market where
there is no prescriptive standard and there is no one
definition of green globally, And before I get into what
some of the opportunities and the challenges are, I think
it's important to remember that there is no one definition
of green, and in many cases, the perfect is the
(06:14):
enemy of the good, and that has worked quite well
for the last decade of the market. Now, while that
has been a proponent of propelling the market forward, it's
also created some challenges that sustainable desks face, that are
issuers face, and even that are investors face. And I
would bucket those challenges into two formats that we typically see.
(06:39):
The first is the comparison and the benchmarking of all
of these frameworks. It's actually quite difficult when you have
an issuer who has either never created a framework before
or had a sustainable finance framework and they're trying to
update that framework.
Speaker 2 (06:58):
I'll give you a real life example.
Speaker 3 (07:01):
Let's say we have an issuer in the corporate space
who has a framework, they published it a couple of
years ago, and now they would like to come back
to the market with a new framework. And so what
does that mean. It is a process that requires sifting
through and reading every single other peer framework globally that has.
Speaker 2 (07:24):
Been published since the.
Speaker 3 (07:26):
Last couple of years when the issuer last published their framework,
going through all of the market standards, all of the
investor viewpoints, and the second party opinions, and to date,
there's no one way to do that efficiently. It is
someone on a sustainable banking team going through each of
(07:47):
these documents, which is exceedingly time consuming. That's the first
problem that I see in the market. And then the
second is the comparison of all those viewpoints and the
thresholds within these documents. And how do we ultimately have
a way to go back to our issuer who would
(08:07):
like to update their framework and give them advice on
where those updates should come from and root them in
the market developments and have a basis for those opinions.
Is the question that we are trying to answer. And
today it's done manually, it's done well, but it certainly
(08:29):
can be done better with the use of some of
these AI tools, both efficiency and also to increase the
credibility of the market.
Speaker 1 (08:38):
Yeah, I'm having I'm having some sweats and flashbacks of
you know, earlier in my career when these were first
coming out and there was just really no way to
even figure out what the what the investment was going
to be used for, and there wasn't all these second
party opinions really and people that are digging in and
per providing more data now so I think it has
(09:00):
gotten better, as you said, but there's plenty more work
that could be done. So maybe we could turn towards
Mensura now and ask how AI can help in this context.
Speaker 5 (09:10):
Yeah, definitely high doses. Andreas from Mansoura also big thank
you for having us on the podcast and the great
collaboration we're having. Since I have a background in data
engineering machine learning, I want to double click on the
technical aspects of of how how AI basically brings a
solution to the table. And I mean Romina greatly outlined
(09:33):
the challenges the work flow, all the pains and Alysta
going through and and these are things we've also seen
with with other partners and companies we've been working with.
At the core, I mean, language model technology has come
a long way since the event of the very first
(09:55):
big models, and we're now really at the at the
pivot time where these technologies really become applicable to business applications.
That means they're getting more and more accurate, They're more
and more capable of supporting complex workflows. So this is
(10:17):
really great. This is a great time to apply these
technologies specifically, I mean, what we are providing with Minsouri
I really targets all the individual pain points Romana highlighted
and her team is having one is access to primary data,
right so this is involved searching the internet for company reports.
(10:41):
Obviously you can use search engines for that, but you
never get a comprehensive answer from there. So this is
something we're providing access, comprehensive access to primary data for
company reports and these can be quite diverse Sustainable Financing Framework,
second party opinions, but also the SINY reports and any
(11:02):
other other company documents. Having these documents at your fingertips,
this is already a big, big enhancement saves you a
lot of time. But obviously we don't stop there, and
the the next thing we can do with AI is
so called semantic indexing, basically storing the data in a
(11:22):
format that you can that allows you to chat with
the document so you don't have to do keyword search
or read the whole thing. I basically just ask what
what information you want to get out of the document
and you get a good, very good result from that,
including references to the to the sources you can verify that.
(11:46):
That's the next step, but obviously it doesn't doesn't scale
for an analyst, right they need to go through multiple
documents and typing in the same question again and again
doesn't it doesn't work. So we're building workflows on top
of that where we can do complex data extraction turning
unstructured data into structured data, and then on top of
(12:08):
that providing a synopsis comparison for example of the eligible
projects mentioned in the use of proceeds, or just list
out the climate goals, extract the greenhouse guards emissions and
basically merge them together to see if the company is
actually on track with these goals and it doesn't stop.
(12:31):
There as many other things we can do, and yeah,
last but not least, with the completed analysis, there's also
the reporting aspect. We can support pretty much giving giving
the analyst, as ourselves said, superpowers, like rather than riding
a bike, they're riding an e bike, which is super
(12:53):
powerful and they can get the work done in a
more fun way fast, They can go deeper into the
subject and that which will ultimately have also begain picked
on the market. Right, we're not trying to replace people here,
but really crow the market size and having an oval impact.
Speaker 1 (13:16):
Yeah, and we've seen it more now, I think even
on the corporation side too, where yeah, they're they're providing
better data and more information and and more data points
than we've ever seen so I think, you know, a
I can definitely help in terms of digesting all this
data that they are now putting out. I mean it
used to be like there was nothing, and now there
(13:37):
there's just report after report and all this data that's
out there and needs to be vetted. So, uh yeah,
great points on how AI could kind of help an
analyst dig through all that. As we look at at
long term success of companies, I mean, sustainable plants, sustainability
plans and targets have become you know, vital, a vital
(13:59):
part of the equation in terms of a company's success
and funding such plans continues to evolve. So what are
some areas maybe of growth that you've seen recently or
areas that are your surprise, haven't seen as much growth
lately in terms of some of these companies plans. I mean,
maybe i'll throw that at Remina.
Speaker 3 (14:21):
We've seen a lot of growth lately from various breakthrough technologies.
So if you think about the hydrogen space, the growth
in green hydrogen, or the growth in even nuclear, the
various types of CCUS technologies, all of these varieties that
enable a company to ultimately reach their decarbonization pathway and
(14:45):
are as I see at levers for how they can
reach those ultimate goals, whether it be an SBTi one
and a half degree validated goal that they've set for
two thousand and forty, which is excellent, But how are
they actually going to reach those goals? What green technology
are they investing in? And when I bring this back
to how can AI help us, it's a matter of
(15:11):
through these tools, are we able to unlock some of
this capital and add it into use of proceeds frameworks
or are even showcase in the context of a sustainability
linked bond, how various green investments are creating this roadmap
for an issuer to meet their goals in a credible way.
Speaker 2 (15:34):
Is the benefit to the.
Speaker 3 (15:37):
Tools that are in the process of being created and
developed they will ultimately unlock this capital. From the side
of the bank, who's identifying what are these projects, are
they viable, what's the amount of money spent on them?
And also on the side of the issuers as they
sift through their various lines of business. That takes a
(15:59):
really long time for an issuer of a green bond
to go through all of their different departments. Where are
you spending money? Is it green is it not green?
How do we organize that and condense it ultimately into
a framework that is acceptable to the ESG investor base.
Speaker 1 (16:16):
I guess I'll ask a very similar question in regards
to AI as a follow up to that, and just
ask how AI can help you know, is it contributing
in this context? Like, you know, how can it help
companies when with all the data and all the projects
that they're trying to put together?
Speaker 5 (16:33):
Yeah, for sure.
Speaker 4 (16:34):
And so I think I can play a role in
two ways. So one is transparency. And so for example,
with the topics that Romina described, with the new advances
and AI, you can essentially monitor in real time what
companies goals are and where there might be gaps, and
you can very easy and like very early sport certain trends. So,
(16:54):
for example, if you're looking into topics like biodiversity, which
right now you have the TNFD, companies are starting to
prepare this initial adoption. With the app of AI, you
can really track the advancements of that space across a
large set of companies and identify patterns. Right, so where
companies making these initial investments, you can then figure out okay,
(17:17):
for given industry or four given peer set, who are
the leaders, who are the laggards? You what do the
leaders do right? And thereby you kind of reduce the
friction in the market by just you're leveling the playing
field if you so well, right, and that helps, you know,
the sustainable Finance desk like really also come up with
ideas of how to guide companies towards the adoption of
(17:39):
some of these new some of these new trends and measures.
I think that's like this one aspect is really having
toransparency across the board and being kind of really early
in identifying who is doing what and leveling the playing field.
The other aspect that I would throw in is harmonization
because as companies, you're peci in these kind of new
(18:00):
evolving industries. As companies are defining their goals and KPI
is very often they're reporting on these in a very
different way, right. And like a simple example could be
something like Okay, one company is breaking down their clubs
three emissions in one way, the other company is doing
it in another way. And with the advantage of AI,
you can then very easily create basis of comparison and
(18:22):
make different units kind of your standardized and harmonized and
really track you know where individual player is going and
how does this compare to some of the peers. So
I think those are two main things that I would
throw in her. So one transparency and the other one
is really harmonization of indicators.
Speaker 1 (18:43):
Yeah, I think, I think you know, that's a good
way that I A, I can definitely help because even
even as something like I think of scores sometimes too,
like people are creating all these scores now for different
companies in a way they do it, but then somebody
else creates go or and it's a different score, right,
It's a different the way that score is even scored.
(19:04):
So it's just there's just every time somebody tries to
attack this, there's just no regulation on how it should
be done, so it's always done differently. So yeah, if
AI could help break that down into a more easily
identifiable way for you yourself to look at the data, I'm
sure that that would be a positive for the market.
(19:24):
I like to maybe ask now, in terms of going
back to focus more on on like this the sustainable
debt space, how how AI can maybe help companies involved
in sustainable debt across the life cycle from maybe like
trying to figure out the type of issuess they're going
to do, or or potential pricing or reporting, et cetera. Like,
how can it AI help in that context?
Speaker 3 (19:45):
Definitely, I think there are a couple different ways. You
mentioned pricing, Chris, This is a topic that comes up
time and time again for our issuers. What is the
greenium and what does the oversubscription look like if we
are going to issue a sustainable bond versus a traditional bond.
And while we have data points in a variety of
(20:06):
different forms and anecdotes, we don't have one tool yet
that can aggregate all of those data points and actually
start to create trends or sift through documents where these
data points may have been included and have them formatted
in a way that we can actually use as part
(20:29):
of the pitches to help inform our clients. And then
the second way that I see it used throughout the
life cycle is reporting. So we mentioned that earlier, but
when an issuer comes to market with a green bond,
they're required to do a green bond report annually until
full allocation of the proceeds in order to be aligned
(20:50):
with the green bond principles.
Speaker 2 (20:52):
And once that.
Speaker 3 (20:53):
Report is created and it's on their website. It goes
back to one of the challenges that Marcel mentioned is
transparency and that harmonization. How do you take that report
from Issuer A and compare it to the same report
from Issuer bcd N E and use that comparison in
(21:17):
order to help inform the next deal that you're going
to structure. That is a void that is in the
market right now. And I don't think that we leverage
these reports as much as we possibly could in order
to inform the next structure in the next deal. And
I see that as a yet another way that this
(21:39):
AI tool, the functionality can aid in not only the
life cycle from pitching the deal, winning the deal, structuring
the deal, executing the deal, but how do we then
keep it going to here's the greenbawn report, here's what
we learned, here's how it compares to others, and that
entire higher, full circle evolution of the green bond market
(22:04):
to ultimately what are we trying to do make these
green bond transactions more transparent and stronger for the ESGA
bester base.
Speaker 5 (22:13):
Yeah, and to add to what Romina just said, Yeah,
it's really I think the most important thing is to
think about the life cycle, right. It involves all the
different aspects in the market players specifically, I mean we've
seen it with other from other partners as feedback issues
(22:33):
are struggling with data collection and management and then also
harmonization standardization pretty much getting all the necessary information together
to issue sustainable debt offering. This is something where we
can help in the future and then also kind of
(22:55):
help them validate how this offering compares to the market.
Like you're like give reverse benchmarking that AMINA team is
currently doing, we can also apply to on the issuing side,
so it's kind of more. It helps them to already
ahead of the issuans tweak weak points, help understand what
(23:15):
the rest of the market is doing, where the gifts
they need to work on. This would certainly greatly help
streamline the issues process. And yeah, as I mean said,
the data management reporting is key that after the issuance,
staying on top and reporting them on the progress the
allocation of the money and the KPIs how they're being fulfilled.
Speaker 1 (23:39):
Yeah, and I think of it too. I really see
it benefiting you know some of the smaller firms out
there too, right, that don't have and less capital to
just throw at the problem and and hire more people
and hire more people. So I think a lot of
you know, in the early days of like some people
coming with green bonds or the like, you know, it
was a large issuers because they could afford it, and
(24:01):
they had the infrastructure to be able to create these
reports and find the data. And so for some of
the smaller companies out there, I think they could really
benefit from the aspect of AI being able to help
put all these reports and data together for.
Speaker 2 (24:17):
Them without a doubt.
Speaker 3 (24:19):
Chris and I think that's a really important point that
to date, a majority of the issuers, let's just say
ninety percent of the issuers in the market are large
investment grade companies. Then we do have a subset of
the market that's high yield companies. But resourcing to put
towards these types of projects within treasury teams and sustainability
teams is thin, and so it becomes a question of
(24:43):
what is the benefit versus going to market with a
traditional bond, And in my opinion, if there are ways
that even on the issuer side, these AI tools can
be used to help streamline the process. On the issuer side,
forget about the bank side and the structures, but how
can the issuers who have small teams actually use these
(25:07):
tools in order to help prepare them. Foreign issuance is
yet another way that we can propel the market forward
for some of those issuers who would like to do
a green bond but they simply don't have the capacity
and they're stuck.
Speaker 4 (25:21):
And maybe to underline, and I think it's it's it's
really a great point. Is it's it's it's about issuing
and kind of structuring these instruments to begin with, but
then also how do you do the ongoing reporting? And
what we're also seeing in the market that the requirements
for for ongoing reporting, the appetite to receive more and
more granular information is really increasing. And we could imagine
(25:43):
a world where this reporting happens real time right where
you know, you can collect project devil data, not just
spend education, but really kind of project devil data on
the on the impact that some of these projects are
having and really in real time, bit about this post
to the to the to the underwriting banks, but then
(26:04):
even more importantly to the investor base that is that
is wanting that kind of granular tracking of information and
then also can dedicate dedicated capital to flow into certain
instruments based on those criteria that investors are ultimately interested in.
Speaker 1 (26:18):
Hmm, yeah, I think. I think a lot of those
are very good points in terms of especially for banks
that are trying to structure these things or issuers. But
what about if if we turn a little bit pivot
a little bit towards the buyers of some of these
instruments and buyers of these that in regards to asset
managers or other institutions that are involved in the space.
Speaker 3 (26:41):
The buyer side for ESG investors, I view it quite
similarly to the struggles that we face on the investment
bank side. How do these ESG investors compare and contrast
all the different frameworks out there and use that in
order for them to create their own methodology in their
opinions on the frameworks and the green or social criteria.
(27:06):
One part, and then the second is how do other
investors look through the methodology documents.
Speaker 2 (27:14):
Of their peers, of the other asset.
Speaker 3 (27:16):
Managers and asset owners and come up with a way
to compare the green methodology of asset manager A versus
the green methodology of asset manager B. It's the same
challenge that we face and lack of prescriptive standard of
how these green bond funds are constructed. So that's yet
(27:37):
another way that from the asset manager side the investors side,
they can compare.
Speaker 2 (27:44):
Documents and frameworks, but also.
Speaker 3 (27:47):
Compare green bond funds green bond methodologies amongst the asset
management community to help better inform how they're going to
create their next green bond fund or tweak the criteria
that they'd like to see. I'll give a real world
example that consistently comes up in terms of for a
green bond issuance, you have a look back and a
(28:10):
look forward period that you can allocate the proceeds. And
if you take a sample size of fifty different green bonds,
there's probably going to be certain percent that has with
two year look back, some have a three year look back,
some have no look back. And if I asked today,
I would like to see the last fifty green bonds
(28:31):
out of the United States across all the sectors, and
I want to know the look back and the look
forward the number of years for that question, and please
can I have that done in the next hour? Could
anybody get that to me? I think it would be
really really hard to do that, but I do think
that AI probably could do it. We will hope, or
(28:55):
we're very close, or we are there in many respects,
but to go look that back to your question about
the ESG investor, this is something that ESG investors are
asking for and they're giving their opinion on. So it's
not language and a document that has overlooked. It's informing,
in a way, a small part of an investment decision process,
(29:19):
and it is data and information that we use on
the investor side, the bank side, and the issuer side
in order to figure out what are those projects.
Speaker 2 (29:28):
Within that look back period.
Speaker 3 (29:29):
That's one example amongst many of how this practically could
be used for the investor base.
Speaker 5 (29:35):
It's a great example of a yeah, and it's exactly
what we are enabling, yeah, with the tooling and platform
we're providing and building out. I want to just highlight
on the bias side, what we also have seen is
bias come with their own specific focus, right, So pretty
(29:59):
much the same analysis we can perform on multiple issues documents,
we can we can perform on data requirements of bias investors.
That's one thing we can do to make basically get
an understanding what where where others co investors or competitors
(30:21):
stand in the market. Uh, this will be greatly beneficial.
But also these specific focus areas these investors bring with them,
we can nicely fold them into the the issues analysis
right to see if there's an alignment with what the
investor is looking for. The interesting technical part here is
(30:46):
that at this point where we're no longer working on
public data, so we need to integrate private data into
the AI tooling and that's obviously comes with a couple
of things you need to acknowledge and provide around security, privacy,
et cetera. Good news is that models and technology advanced
(31:11):
in a way that for most of the use cases
you can actually nicely accomplish that in a private public
cloud setup, so where the banks other financial institutions are
in control of their data, the processing happens within the
boundaries and you don't need to reach out to any
(31:32):
of these big model providers. I just wanted wanted to
highlight the there's also technical implications bringing in private information
that we're very much aware of that and also want
to provide that as a core capability, and I think
that's the strength where you actually go beyond tapping into
(31:52):
the public information or the purchased information from a data
provider and merging it with their own in house data
and and that way become a you really get very
efficient analysis and results from that really helps unlock quite
(32:13):
a lot of other new interesting workflows. And one other thing,
if if if it may add to that, So we
in terms of investors buyers, we're also looking certainly looking
beyond sustainable debt more broadly also into the equity space
and private equity, and yeah, it's kind of it's really
(32:38):
super surprising to see what what investors are looking for,
certainly specifically if you work in the impact investing space.
Just the other day, I did a did an analysis
of how many criteria we analyze so far for partners
and pilot partners, and it's it's one hundred plus different
sustainability criteria ranging from a very wide range of climate
(33:04):
related information. They interested into governance, workplace safety, prison, labor
and supply chain, community involvement, whether the chief sustainability officer
they have one, and the compensation is tied to sustainability goals.
Just put a little bit more color on what we're
(33:26):
looking here. What investors might be looking for is information
that is crucial to.
Speaker 1 (33:32):
Them, and that's very interesting. I think investors have a
lot of ways they'd love to look at it, they
just haven't been able to really yet. So I'm looking
forward to that.
Speaker 6 (33:41):
Taking a quick break here to invite you to Bloomberg
Intelligence's End of your Conference ESG is here today five
investment themes for twenty twenty five. The event will be
held Wednesday, December eleventh at twelve pm at Bloomberg Headquarters
in New York. Come here from leading experts in public
and private sector discussing the rise and energy demand, supply
(34:04):
chain regulations, and more. The link to register will be
in our show description.
Speaker 1 (34:11):
As we're as we're coming close to time here, why
don't we just kind of finish up with just what
future innovations you're seeing or expect to see in the
sustainable debt markets and from other AI applications.
Speaker 3 (34:27):
I think, looping back quickly to what Andres just mentioned
on the private data front is the next frontier to
the usage of this AI tool being for example, say
we have a issuer who comes to market with a
green bond, and as part of that execution, we obtain
(34:48):
twenty pieces of feedback on the green criteria from the
investor base, and we take notes on all that and
we store it in a folder within our systems and
time passes, and then two months from now, we have
another green bond in the same sector, and this one
is coming to market in France, or it's coming to
(35:11):
market in Japan, and that team does the same thing.
They take notes, they learn what they learn, and it
goes into a folder. And perhaps those teams send around
these notes in these questions and these feedback across via email,
and then time goes on and all this is magically
lost or it's forgotten about in some folder. And are
(35:33):
we actually capturing a lot of this esg feedback to
veteran form the structuring for the next deals. The reality is,
in many cases we're not. So how do these AI
tools actually help us with those processes beyond the public
information in those public documents, I think is one of
(35:53):
the next frontiers. But in addition to that, what do
I see as the future of AI and sustainable finance?
Personally working with Marcel and Andreas as design partners and
creating and helping see how we can use these tools
on the bank side, it has really changed how I
(36:15):
think about structuring deals and what I'm going to put
into a framework. Whereas yesterday, knowing nothing about AI, I
would have done it this way. But today, knowing about
the capabilities of AI and where these tools are going
and how they could be used, We're thinking about the
(36:36):
structuring and the language that we put into frameworks quite differently.
So not only viewing it as here's a problem, how
do we solve this problem with AI, but knowing how
AI can solve some of our problems, how do I
create a positive impact on that feedback loop to structure
deals and to write frameworks in a way that makes
(36:58):
it simpler for AI to help us in the end.
And that I think is a mindset shift that unless
you are working with a company or a platform that
is integrating AI and learn the capabilities of some of
this technology, you don't fully appreciate. And that's something that
(37:19):
I've learned and how to think through the next frontier
of sustainable finance from the bank perspective.
Speaker 4 (37:26):
Yeah, and maybe I mean Christopher if I major up
in because I think this is like one really important
aspect here. I think up to now, most of the
kind of unlocked from the II component has come through efficiency, right,
so it allows you to do existing tasks, you know,
three times faster, four times faster, five times faster. And
we've seen that and this is great. This is obviously
(37:48):
already a great unlock. I think what we are very
excited about is that, you know, technology will only be
be better from here. And if we just look at
kind of where we've come from from last year to
the advancements in a large language models, the pace is
really really accelerating, and we're excited about you know, what's
what's after the efficiency play, right, So just maybe to
(38:11):
throw out a few things that we're quite excited about
a few themes that we're seeing. Your one is like
the arrival of agents, and what that means is that
you know, now you can define different roles within a
large language models that does that do different types of tasks.
So you could define Okay, now you're an analyst, and
now there's a reviewer, so analysts please perform this task,
(38:34):
but also reviewer please then check the work that the
analyst has done and come back with suggestions on how
to make that better. Yeah, you can implement that and
you will better get better results from that implementation. I
think that's like one one area that we're trying to
implement and super excited about the other pieces. Predictive analytics.
So if you think about you're not just tracking progress
(38:57):
towards a certain KPI, a certain climate goals, so to say,
really thinking about, okay, will the companies actually hit their
goals right, like for sustainability being bonds? Will they hit
the targets that they have defined? How likely is that?
What's surprising impact? So I think those topics are super interesting.
And then kind of thirdly, and we alluded to that
earlier when we talked about the issue dimension is you know,
(39:21):
real time tracking of KPI level project data by by issuers. Right,
so rather than just understanding how much money have they
allocated to certain areas of green finance or certain spend buckets,
so to say that are being tied back to green bonds,
what is really the granular impact of these initiatives? And
you know, how can we kind of report that you're
(39:43):
throughout the the the value chain and across all stakeholders
that are involved in both issuing but then also kind
of investing.
Speaker 5 (39:51):
Into these instruments.
Speaker 4 (39:52):
So those are a few things, but there's so many more.
We could probably spend you know, the next next hour
to to talk about the exciting things to come.
Speaker 5 (40:00):
But it's it's a great.
Speaker 4 (40:01):
Time to work on this problem at the moment. I
think also a really good time to kind of get
on board and get yourself acquainted with these new technologies
and if it may.
Speaker 5 (40:11):
May just add a little bit more on top of
that and be a little bit more visionary. I mean,
first of all, we just reiterate it's really exciting to
see these these technologies mature and really the future becomes tangible.
I mean to me, there's really there's a technical pass
to accomplishing all which is discussed. This is a question
(40:34):
of building out the systems, fine tuning them and shipping
them and obviously iterating with partners on them to make
the most most useful. But this is this is super tangible.
The I think the going beyond it, the exciting part
is really also scale, like really getting down the investable universe,
(40:55):
everything that's out there that really allows for way deeper
and overviews, but also things with the market and that
space is evolving. There's also the aspect of the not
building a knowledge base right there's information coming from regulation,
(41:15):
from standards, from science, science research. How do you stay
on top of all these areas? If you want to
invest in a in a specific niche market, how do
you know that? What are you unknowns? How do you
know what is currently being discussed as the most relevant
aspects in nuclear or whatever there is? And there's a
(41:40):
there's a there's a path to to build such a
knowledge knowledge base and leverage it so to inform your decisions,
to derive new frameworks and standards, and practically, for example,
to apply it to things like regulatory risk assessment. Use
a company you in that space, the there is regulation upcoming,
(42:02):
How does this impact your your operations? How much do
you have to adapt? Do you have liabilities legacies on
your emission side and other areas of your company operations?
So this is just very very tangible. I mean also
in terms of data sources going beyond that, controversies, discovery,
(42:25):
red flag analysis, something companies already do out there. It's
also very tangible and many more. We didn't dive into
biodiversity or physical climate risk assessment. All these areas will
benefit from from these technologies. So super super exciting.
Speaker 1 (42:45):
Yeah, without a doubt. I mean in terms of expanding
the knowledge base. I mean the scalability, being able to
compare largely regular data sets, and just gaining efficiencies. Overall,
I think there's plenty of opportunities for AI to continue
to assist in the roles of analyst, companies, bankers, a
(43:05):
whole host of interested parties. So thank you all of
you for this interesting discussion. I'd like to thank Ramita,
Marcel and Andreas for joining me today to discuss sustainable
financing and the role AI can play. And thank you
to all our listeners out there. Keep downloading belove it.
(43:26):
You can find more information on sustainable debt markets, climate investments,
carbon markets, and a whole array of other topics by
going to b I E. S G on the Bloomberg terminal,
which opens up to the Bloomberg Intelligence ESG Research dashboard.
If you have any other questions, please reach out to
one of our BI experts send us an email at
(43:47):
ESG Currents at Bloomberg dot net. Thank you everyone,
ESG has become established as a key business theme as
companies and investors seek to navigate the climate crisis, energy transition,
social mega trends, mouting, regulatory attention and pressures from other stakeholders.
The rapidly evolving landscape has become inundated with acronyms, buzzwords
and lingo, and we aim to break these down with
industry experts. Welcome to ESG Currents, your guide to navigating
(00:33):
the evolving ESD space, one topic at a time. Brought
to you by Bloomberg Intelligence, which is part of Bloomberg's
research department, with five hundred analysts and strategists working across
all major world markets. Our coverage includes over two thousand
equities and credits, as well as well as outlooks on
more than ninety industries and one hundred market indices, currencies
(00:55):
and commodities. I'm Chris Rady, senior ESG credit strategists and
your hosts for this episode. Today, we're talking with Romina Reversi,
ben engine director and head of Sustainable Investment Banking America's
at Credit Agricol CIB, along with Marcel Bach and Andreas Holzel,
who are co founders of mensura AI. Welcome everyone, and
(01:15):
thank you for joining us today.
Speaker 2 (01:17):
Thank you, Chris, thank you.
Speaker 1 (01:19):
Today we're gonna kind of jump in and just discuss
how artificial intelligence can be utilized by stakeholders in sustainable financing.
And for me, when I think about the sustainable debt markets,
I'm looking at you labeled use of proceeds bonds that
have seen continued growth. There's over five trillion, depending on
which numbers you look at now in bonds outstanding, with
(01:41):
the most common being green bonds, followed by social and
sustainability bonds. There's also other various forms of investments out
there like blue bond, sustainability link bonds out with bonds debt,
for nature swaps. The list goes on. So when I
then turn and look at our official intelligence, I think
more about generative AI and content creation notes on complex topics,
(02:07):
you know, helping you summarize very complex things, or machine
learning which might help you look at large data sets
and analyze it in a different way. So I think,
before we get started, maybe it'd be good if maybe
Romina can define sustainable finance in terms of the context
of today's discussion, and then maybe Marcel Or Andres can
(02:28):
enlighten us with types of AI that you guys are
most focused on.
Speaker 3 (02:33):
Thanks Chris, it's a pleasure to be here today. I'm
Romina University heading up sustainable banking for the America's at
Credit I recall CIB so our team is involved in
all the structuring of ESG related debt. And to answer
your question, Chris, how would I define sustainable finance? At
(02:53):
its core, it is the structuring of labeling of ESG
related products, both on the bond and the loan side.
So that includes green bonds, sustainability linked bonds, sustainability linked
r cfs, all of these products that are issued and
borrowed across a variety of various companies and sovereigns that
(03:18):
either use proceeds for green and or social projects as
well as have a link to some form of environmental
and social KPI.
Speaker 1 (03:29):
Right, and then over to Marcela or Andres whatever wants
to kind of let us know how you're looking at
AI in terms of Mansirah.
Speaker 4 (03:38):
Yeah, thank you Chris, and thanks again for having us on.
Great to be here. So this is myself from Mansura.
Andreas and I have been building Mansua on the belief
that there's all of these great technologies out there and
all of the superpowers that AI can help you get today,
but we want to really tailor this to the context
(03:59):
of sustainable finance, and we want to in the end
equip sustainable finance professionals with the superpowers. The belief is
that ultimately, if we can, if we can enable sustainable
finance professionals, that will then increase the flow of capital
into green investments, green bonds, and accelerate the green transition.
(04:19):
So this is something that we're very excited about. You know,
on our end, we've observed a couple of challenges that
sustainable finance professionals very often face. There's a lot of
kind of manual and time consuming analysis which are very repetitive,
and you know those are those are happening by analysts
(04:42):
that are working in Romina's team, for example, but those
are happening really across the board. And there's a lot
of disparate data sources that are not really being aggregated,
and so folks have kind of trouble bringing us all
together and making sense of it. And then there's the
kind of current AYE models that are very powerful, but
(05:03):
they're not really tailored to the context of sustainability. So
this is where we come in. So we're building a
platform that allows those professionals to perform automated analytics workflows,
to have a one stop shop for all of the
different information on the issue on bond level that is
out there, and really kind of a co pilot by
their side that helps them navigate through some of these
(05:26):
data sources and analytics workflows.
Speaker 1 (05:29):
Yeah, I think there's a lot more in terms of
what a I could do, and I think we're gonna,
obviously in today's discussion, just talk about a lot of
those a little more with both of you, all of you. Sorry,
but let me turn back to Romina for a second
and just maybe we could talk about some of the
(05:50):
opportunities or challenges that green financing is currently facing.
Speaker 3 (05:54):
So at a score, green financing is a market where
there is no prescriptive standard and there is no one
definition of green globally, And before I get into what
some of the opportunities and the challenges are, I think
it's important to remember that there is no one definition
of green, and in many cases, the perfect is the
(06:14):
enemy of the good, and that has worked quite well
for the last decade of the market. Now, while that
has been a proponent of propelling the market forward, it's
also created some challenges that sustainable desks face, that are
issuers face, and even that are investors face. And I
would bucket those challenges into two formats that we typically see.
(06:39):
The first is the comparison and the benchmarking of all
of these frameworks. It's actually quite difficult when you have
an issuer who has either never created a framework before
or had a sustainable finance framework and they're trying to
update that framework.
Speaker 2 (06:58):
I'll give you a real life example.
Speaker 3 (07:01):
Let's say we have an issuer in the corporate space
who has a framework, they published it a couple of
years ago, and now they would like to come back
to the market with a new framework. And so what
does that mean. It is a process that requires sifting
through and reading every single other peer framework globally that has.
Speaker 2 (07:24):
Been published since the.
Speaker 3 (07:26):
Last couple of years when the issuer last published their framework,
going through all of the market standards, all of the
investor viewpoints, and the second party opinions, and to date,
there's no one way to do that efficiently. It is
someone on a sustainable banking team going through each of
(07:47):
these documents, which is exceedingly time consuming. That's the first
problem that I see in the market. And then the
second is the comparison of all those viewpoints and the
thresholds within these documents. And how do we ultimately have
a way to go back to our issuer who would
(08:07):
like to update their framework and give them advice on
where those updates should come from and root them in
the market developments and have a basis for those opinions.
Is the question that we are trying to answer. And
today it's done manually, it's done well, but it certainly
(08:29):
can be done better with the use of some of
these AI tools, both efficiency and also to increase the
credibility of the market.
Speaker 1 (08:38):
Yeah, I'm having I'm having some sweats and flashbacks of
you know, earlier in my career when these were first
coming out and there was just really no way to
even figure out what the what the investment was going
to be used for, and there wasn't all these second
party opinions really and people that are digging in and
per providing more data now so I think it has
(09:00):
gotten better, as you said, but there's plenty more work
that could be done. So maybe we could turn towards
Mensura now and ask how AI can help in this context.
Speaker 5 (09:10):
Yeah, definitely high doses. Andreas from Mansoura also big thank
you for having us on the podcast and the great
collaboration we're having. Since I have a background in data
engineering machine learning, I want to double click on the
technical aspects of of how how AI basically brings a
solution to the table. And I mean Romina greatly outlined
(09:33):
the challenges the work flow, all the pains and Alysta
going through and and these are things we've also seen
with with other partners and companies we've been working with.
At the core, I mean, language model technology has come
a long way since the event of the very first
(09:55):
big models, and we're now really at the at the
pivot time where these technologies really become applicable to business applications.
That means they're getting more and more accurate, They're more
and more capable of supporting complex workflows. So this is
(10:17):
really great. This is a great time to apply these
technologies specifically, I mean, what we are providing with Minsouri
I really targets all the individual pain points Romana highlighted
and her team is having one is access to primary data,
right so this is involved searching the internet for company reports.
(10:41):
Obviously you can use search engines for that, but you
never get a comprehensive answer from there. So this is
something we're providing access, comprehensive access to primary data for
company reports and these can be quite diverse Sustainable Financing Framework,
second party opinions, but also the SINY reports and any
(11:02):
other other company documents. Having these documents at your fingertips,
this is already a big, big enhancement saves you a
lot of time. But obviously we don't stop there, and
the the next thing we can do with AI is
so called semantic indexing, basically storing the data in a
(11:22):
format that you can that allows you to chat with
the document so you don't have to do keyword search
or read the whole thing. I basically just ask what
what information you want to get out of the document
and you get a good, very good result from that,
including references to the to the sources you can verify that.
(11:46):
That's the next step, but obviously it doesn't doesn't scale
for an analyst, right they need to go through multiple
documents and typing in the same question again and again
doesn't it doesn't work. So we're building workflows on top
of that where we can do complex data extraction turning
unstructured data into structured data, and then on top of
(12:08):
that providing a synopsis comparison for example of the eligible
projects mentioned in the use of proceeds, or just list
out the climate goals, extract the greenhouse guards emissions and
basically merge them together to see if the company is
actually on track with these goals and it doesn't stop.
(12:31):
There as many other things we can do, and yeah,
last but not least, with the completed analysis, there's also
the reporting aspect. We can support pretty much giving giving
the analyst, as ourselves said, superpowers, like rather than riding
a bike, they're riding an e bike, which is super
(12:53):
powerful and they can get the work done in a
more fun way fast, They can go deeper into the
subject and that which will ultimately have also begain picked
on the market. Right, we're not trying to replace people here,
but really crow the market size and having an oval impact.
Speaker 1 (13:16):
Yeah, and we've seen it more now, I think even
on the corporation side too, where yeah, they're they're providing
better data and more information and and more data points
than we've ever seen so I think, you know, a
I can definitely help in terms of digesting all this
data that they are now putting out. I mean it
used to be like there was nothing, and now there
(13:37):
there's just report after report and all this data that's
out there and needs to be vetted. So, uh yeah,
great points on how AI could kind of help an
analyst dig through all that. As we look at at
long term success of companies, I mean, sustainable plants, sustainability
plans and targets have become you know, vital, a vital
(13:59):
part of the equation in terms of a company's success
and funding such plans continues to evolve. So what are
some areas maybe of growth that you've seen recently or
areas that are your surprise, haven't seen as much growth
lately in terms of some of these companies plans. I mean,
maybe i'll throw that at Remina.
Speaker 3 (14:21):
We've seen a lot of growth lately from various breakthrough technologies.
So if you think about the hydrogen space, the growth
in green hydrogen, or the growth in even nuclear, the
various types of CCUS technologies, all of these varieties that
enable a company to ultimately reach their decarbonization pathway and
(14:45):
are as I see at levers for how they can
reach those ultimate goals, whether it be an SBTi one
and a half degree validated goal that they've set for
two thousand and forty, which is excellent, But how are
they actually going to reach those goals? What green technology
are they investing in? And when I bring this back
to how can AI help us, it's a matter of
(15:11):
through these tools, are we able to unlock some of
this capital and add it into use of proceeds frameworks
or are even showcase in the context of a sustainability
linked bond, how various green investments are creating this roadmap
for an issuer to meet their goals in a credible way.
Speaker 2 (15:34):
Is the benefit to the.
Speaker 3 (15:37):
Tools that are in the process of being created and
developed they will ultimately unlock this capital. From the side
of the bank, who's identifying what are these projects, are
they viable, what's the amount of money spent on them?
And also on the side of the issuers as they
sift through their various lines of business. That takes a
(15:59):
really long time for an issuer of a green bond
to go through all of their different departments. Where are
you spending money? Is it green is it not green?
How do we organize that and condense it ultimately into
a framework that is acceptable to the ESG investor base.
Speaker 1 (16:16):
I guess I'll ask a very similar question in regards
to AI as a follow up to that, and just
ask how AI can help you know, is it contributing
in this context? Like, you know, how can it help
companies when with all the data and all the projects
that they're trying to put together?
Speaker 5 (16:33):
Yeah, for sure.
Speaker 4 (16:34):
And so I think I can play a role in
two ways. So one is transparency. And so for example,
with the topics that Romina described, with the new advances
and AI, you can essentially monitor in real time what
companies goals are and where there might be gaps, and
you can very easy and like very early sport certain trends. So,
(16:54):
for example, if you're looking into topics like biodiversity, which
right now you have the TNFD, companies are starting to
prepare this initial adoption. With the app of AI, you
can really track the advancements of that space across a
large set of companies and identify patterns. Right, so where
companies making these initial investments, you can then figure out okay,
(17:17):
for given industry or four given peer set, who are
the leaders, who are the laggards? You what do the
leaders do right? And thereby you kind of reduce the
friction in the market by just you're leveling the playing
field if you so well, right, and that helps, you know,
the sustainable Finance desk like really also come up with
ideas of how to guide companies towards the adoption of
(17:39):
some of these new some of these new trends and measures.
I think that's like this one aspect is really having
toransparency across the board and being kind of really early
in identifying who is doing what and leveling the playing field.
The other aspect that I would throw in is harmonization
because as companies, you're peci in these kind of new
(18:00):
evolving industries. As companies are defining their goals and KPI
is very often they're reporting on these in a very
different way, right. And like a simple example could be
something like Okay, one company is breaking down their clubs
three emissions in one way, the other company is doing
it in another way. And with the advantage of AI,
you can then very easily create basis of comparison and
(18:22):
make different units kind of your standardized and harmonized and
really track you know where individual player is going and
how does this compare to some of the peers. So
I think those are two main things that I would
throw in her. So one transparency and the other one
is really harmonization of indicators.
Speaker 1 (18:43):
Yeah, I think, I think you know, that's a good
way that I A, I can definitely help because even
even as something like I think of scores sometimes too,
like people are creating all these scores now for different
companies in a way they do it, but then somebody
else creates go or and it's a different score, right,
It's a different the way that score is even scored.
(19:04):
So it's just there's just every time somebody tries to
attack this, there's just no regulation on how it should
be done, so it's always done differently. So yeah, if
AI could help break that down into a more easily
identifiable way for you yourself to look at the data, I'm
sure that that would be a positive for the market.
(19:24):
I like to maybe ask now, in terms of going
back to focus more on on like this the sustainable
debt space, how how AI can maybe help companies involved
in sustainable debt across the life cycle from maybe like
trying to figure out the type of issuess they're going
to do, or or potential pricing or reporting, et cetera. Like,
how can it AI help in that context?
Speaker 3 (19:45):
Definitely, I think there are a couple different ways. You
mentioned pricing, Chris, This is a topic that comes up
time and time again for our issuers. What is the
greenium and what does the oversubscription look like if we
are going to issue a sustainable bond versus a traditional bond.
And while we have data points in a variety of
(20:06):
different forms and anecdotes, we don't have one tool yet
that can aggregate all of those data points and actually
start to create trends or sift through documents where these
data points may have been included and have them formatted
in a way that we can actually use as part
(20:29):
of the pitches to help inform our clients. And then
the second way that I see it used throughout the
life cycle is reporting. So we mentioned that earlier, but
when an issuer comes to market with a green bond,
they're required to do a green bond report annually until
full allocation of the proceeds in order to be aligned
(20:50):
with the green bond principles.
Speaker 2 (20:52):
And once that.
Speaker 3 (20:53):
Report is created and it's on their website. It goes
back to one of the challenges that Marcel mentioned is
transparency and that harmonization. How do you take that report
from Issuer A and compare it to the same report
from Issuer bcd N E and use that comparison in
(21:17):
order to help inform the next deal that you're going
to structure. That is a void that is in the
market right now. And I don't think that we leverage
these reports as much as we possibly could in order
to inform the next structure in the next deal. And
I see that as a yet another way that this
(21:39):
AI tool, the functionality can aid in not only the
life cycle from pitching the deal, winning the deal, structuring
the deal, executing the deal, but how do we then
keep it going to here's the greenbawn report, here's what
we learned, here's how it compares to others, and that
entire higher, full circle evolution of the green bond market
(22:04):
to ultimately what are we trying to do make these
green bond transactions more transparent and stronger for the ESGA
bester base.
Speaker 5 (22:13):
Yeah, and to add to what Romina just said, Yeah,
it's really I think the most important thing is to
think about the life cycle, right. It involves all the
different aspects in the market players specifically, I mean we've
seen it with other from other partners as feedback issues
(22:33):
are struggling with data collection and management and then also
harmonization standardization pretty much getting all the necessary information together
to issue sustainable debt offering. This is something where we
can help in the future and then also kind of
(22:55):
help them validate how this offering compares to the market.
Like you're like give reverse benchmarking that AMINA team is
currently doing, we can also apply to on the issuing side,
so it's kind of more. It helps them to already
ahead of the issuans tweak weak points, help understand what
(23:15):
the rest of the market is doing, where the gifts
they need to work on. This would certainly greatly help
streamline the issues process. And yeah, as I mean said,
the data management reporting is key that after the issuance,
staying on top and reporting them on the progress the
allocation of the money and the KPIs how they're being fulfilled.
Speaker 1 (23:39):
Yeah, and I think of it too. I really see
it benefiting you know some of the smaller firms out
there too, right, that don't have and less capital to
just throw at the problem and and hire more people
and hire more people. So I think a lot of
you know, in the early days of like some people
coming with green bonds or the like, you know, it
was a large issuers because they could afford it, and
(24:01):
they had the infrastructure to be able to create these
reports and find the data. And so for some of
the smaller companies out there, I think they could really
benefit from the aspect of AI being able to help
put all these reports and data together for.
Speaker 2 (24:17):
Them without a doubt.
Speaker 3 (24:19):
Chris and I think that's a really important point that
to date, a majority of the issuers, let's just say
ninety percent of the issuers in the market are large
investment grade companies. Then we do have a subset of
the market that's high yield companies. But resourcing to put
towards these types of projects within treasury teams and sustainability
teams is thin, and so it becomes a question of
(24:43):
what is the benefit versus going to market with a
traditional bond, And in my opinion, if there are ways
that even on the issuer side, these AI tools can
be used to help streamline the process. On the issuer side,
forget about the bank side and the structures, but how
can the issuers who have small teams actually use these
(25:07):
tools in order to help prepare them. Foreign issuance is
yet another way that we can propel the market forward
for some of those issuers who would like to do
a green bond but they simply don't have the capacity
and they're stuck.
Speaker 4 (25:21):
And maybe to underline, and I think it's it's it's
really a great point. Is it's it's it's about issuing
and kind of structuring these instruments to begin with, but
then also how do you do the ongoing reporting? And
what we're also seeing in the market that the requirements
for for ongoing reporting, the appetite to receive more and
more granular information is really increasing. And we could imagine
(25:43):
a world where this reporting happens real time right where
you know, you can collect project devil data, not just
spend education, but really kind of project devil data on
the on the impact that some of these projects are
having and really in real time, bit about this post
to the to the to the underwriting banks, but then
(26:04):
even more importantly to the investor base that is that
is wanting that kind of granular tracking of information and
then also can dedicate dedicated capital to flow into certain
instruments based on those criteria that investors are ultimately interested in.
Speaker 1 (26:18):
Hmm, yeah, I think. I think a lot of those
are very good points in terms of especially for banks
that are trying to structure these things or issuers. But
what about if if we turn a little bit pivot
a little bit towards the buyers of some of these
instruments and buyers of these that in regards to asset
managers or other institutions that are involved in the space.
Speaker 3 (26:41):
The buyer side for ESG investors, I view it quite
similarly to the struggles that we face on the investment
bank side. How do these ESG investors compare and contrast
all the different frameworks out there and use that in
order for them to create their own methodology in their
opinions on the frameworks and the green or social criteria.
(27:06):
One part, and then the second is how do other
investors look through the methodology documents.
Speaker 2 (27:14):
Of their peers, of the other asset.
Speaker 3 (27:16):
Managers and asset owners and come up with a way
to compare the green methodology of asset manager A versus
the green methodology of asset manager B. It's the same
challenge that we face and lack of prescriptive standard of
how these green bond funds are constructed. So that's yet
(27:37):
another way that from the asset manager side the investors side,
they can compare.
Speaker 2 (27:44):
Documents and frameworks, but also.
Speaker 3 (27:47):
Compare green bond funds green bond methodologies amongst the asset
management community to help better inform how they're going to
create their next green bond fund or tweak the criteria
that they'd like to see. I'll give a real world
example that consistently comes up in terms of for a
green bond issuance, you have a look back and a
(28:10):
look forward period that you can allocate the proceeds. And
if you take a sample size of fifty different green bonds,
there's probably going to be certain percent that has with
two year look back, some have a three year look back,
some have no look back. And if I asked today,
I would like to see the last fifty green bonds
(28:31):
out of the United States across all the sectors, and
I want to know the look back and the look
forward the number of years for that question, and please
can I have that done in the next hour? Could
anybody get that to me? I think it would be
really really hard to do that, but I do think
that AI probably could do it. We will hope, or
(28:55):
we're very close, or we are there in many respects,
but to go look that back to your question about
the ESG investor, this is something that ESG investors are
asking for and they're giving their opinion on. So it's
not language and a document that has overlooked. It's informing,
in a way, a small part of an investment decision process,
(29:19):
and it is data and information that we use on
the investor side, the bank side, and the issuer side
in order to figure out what are those projects.
Speaker 2 (29:28):
Within that look back period.
Speaker 3 (29:29):
That's one example amongst many of how this practically could
be used for the investor base.
Speaker 5 (29:35):
It's a great example of a yeah, and it's exactly
what we are enabling, yeah, with the tooling and platform
we're providing and building out. I want to just highlight
on the bias side, what we also have seen is
bias come with their own specific focus, right, So pretty
(29:59):
much the same analysis we can perform on multiple issues documents,
we can we can perform on data requirements of bias investors.
That's one thing we can do to make basically get
an understanding what where where others co investors or competitors
(30:21):
stand in the market. Uh, this will be greatly beneficial.
But also these specific focus areas these investors bring with them,
we can nicely fold them into the the issues analysis
right to see if there's an alignment with what the
investor is looking for. The interesting technical part here is
(30:46):
that at this point where we're no longer working on
public data, so we need to integrate private data into
the AI tooling and that's obviously comes with a couple
of things you need to acknowledge and provide around security, privacy,
et cetera. Good news is that models and technology advanced
(31:11):
in a way that for most of the use cases
you can actually nicely accomplish that in a private public
cloud setup, so where the banks other financial institutions are
in control of their data, the processing happens within the
boundaries and you don't need to reach out to any
(31:32):
of these big model providers. I just wanted wanted to
highlight the there's also technical implications bringing in private information
that we're very much aware of that and also want
to provide that as a core capability, and I think
that's the strength where you actually go beyond tapping into
(31:52):
the public information or the purchased information from a data
provider and merging it with their own in house data
and and that way become a you really get very
efficient analysis and results from that really helps unlock quite
(32:13):
a lot of other new interesting workflows. And one other thing,
if if if it may add to that, So we
in terms of investors buyers, we're also looking certainly looking
beyond sustainable debt more broadly also into the equity space
and private equity, and yeah, it's kind of it's really
(32:38):
super surprising to see what what investors are looking for,
certainly specifically if you work in the impact investing space.
Just the other day, I did a did an analysis
of how many criteria we analyze so far for partners
and pilot partners, and it's it's one hundred plus different
sustainability criteria ranging from a very wide range of climate
(33:04):
related information. They interested into governance, workplace safety, prison, labor
and supply chain, community involvement, whether the chief sustainability officer
they have one, and the compensation is tied to sustainability goals.
Just put a little bit more color on what we're
(33:26):
looking here. What investors might be looking for is information
that is crucial to.
Speaker 1 (33:32):
Them, and that's very interesting. I think investors have a
lot of ways they'd love to look at it, they
just haven't been able to really yet. So I'm looking
forward to that.
Speaker 6 (33:41):
Taking a quick break here to invite you to Bloomberg
Intelligence's End of your Conference ESG is here today five
investment themes for twenty twenty five. The event will be
held Wednesday, December eleventh at twelve pm at Bloomberg Headquarters
in New York. Come here from leading experts in public
and private sector discussing the rise and energy demand, supply
(34:04):
chain regulations, and more. The link to register will be
in our show description.
Speaker 1 (34:11):
As we're as we're coming close to time here, why
don't we just kind of finish up with just what
future innovations you're seeing or expect to see in the
sustainable debt markets and from other AI applications.
Speaker 3 (34:27):
I think, looping back quickly to what Andres just mentioned
on the private data front is the next frontier to
the usage of this AI tool being for example, say
we have a issuer who comes to market with a
green bond, and as part of that execution, we obtain
(34:48):
twenty pieces of feedback on the green criteria from the
investor base, and we take notes on all that and
we store it in a folder within our systems and
time passes, and then two months from now, we have
another green bond in the same sector, and this one
is coming to market in France, or it's coming to
(35:11):
market in Japan, and that team does the same thing.
They take notes, they learn what they learn, and it
goes into a folder. And perhaps those teams send around
these notes in these questions and these feedback across via email,
and then time goes on and all this is magically
lost or it's forgotten about in some folder. And are
(35:33):
we actually capturing a lot of this esg feedback to
veteran form the structuring for the next deals. The reality is,
in many cases we're not. So how do these AI
tools actually help us with those processes beyond the public
information in those public documents, I think is one of
(35:53):
the next frontiers. But in addition to that, what do
I see as the future of AI and sustainable finance?
Personally working with Marcel and Andreas as design partners and
creating and helping see how we can use these tools
on the bank side, it has really changed how I
(36:15):
think about structuring deals and what I'm going to put
into a framework. Whereas yesterday, knowing nothing about AI, I
would have done it this way. But today, knowing about
the capabilities of AI and where these tools are going
and how they could be used, We're thinking about the
(36:36):
structuring and the language that we put into frameworks quite differently.
So not only viewing it as here's a problem, how
do we solve this problem with AI, but knowing how
AI can solve some of our problems, how do I
create a positive impact on that feedback loop to structure
deals and to write frameworks in a way that makes
(36:58):
it simpler for AI to help us in the end.
And that I think is a mindset shift that unless
you are working with a company or a platform that
is integrating AI and learn the capabilities of some of
this technology, you don't fully appreciate. And that's something that
(37:19):
I've learned and how to think through the next frontier
of sustainable finance from the bank perspective.
Speaker 4 (37:26):
Yeah, and maybe I mean Christopher if I major up
in because I think this is like one really important
aspect here. I think up to now, most of the
kind of unlocked from the II component has come through efficiency, right,
so it allows you to do existing tasks, you know,
three times faster, four times faster, five times faster. And
we've seen that and this is great. This is obviously
(37:48):
already a great unlock. I think what we are very
excited about is that, you know, technology will only be
be better from here. And if we just look at
kind of where we've come from from last year to
the advancements in a large language models, the pace is
really really accelerating, and we're excited about you know, what's
what's after the efficiency play, right, So just maybe to
(38:11):
throw out a few things that we're quite excited about
a few themes that we're seeing. Your one is like
the arrival of agents, and what that means is that
you know, now you can define different roles within a
large language models that does that do different types of tasks.
So you could define Okay, now you're an analyst, and
now there's a reviewer, so analysts please perform this task,
(38:34):
but also reviewer please then check the work that the
analyst has done and come back with suggestions on how
to make that better. Yeah, you can implement that and
you will better get better results from that implementation. I
think that's like one one area that we're trying to
implement and super excited about the other pieces. Predictive analytics.
So if you think about you're not just tracking progress
(38:57):
towards a certain KPI, a certain climate goals, so to say,
really thinking about, okay, will the companies actually hit their
goals right, like for sustainability being bonds? Will they hit
the targets that they have defined? How likely is that?
What's surprising impact? So I think those topics are super interesting.
And then kind of thirdly, and we alluded to that
earlier when we talked about the issue dimension is you know,
(39:21):
real time tracking of KPI level project data by by issuers. Right,
so rather than just understanding how much money have they
allocated to certain areas of green finance or certain spend buckets,
so to say that are being tied back to green bonds,
what is really the granular impact of these initiatives? And
you know, how can we kind of report that you're
(39:43):
throughout the the the value chain and across all stakeholders
that are involved in both issuing but then also kind
of investing.
Speaker 5 (39:51):
Into these instruments.
Speaker 4 (39:52):
So those are a few things, but there's so many more.
We could probably spend you know, the next next hour
to to talk about the exciting things to come.
Speaker 5 (40:00):
But it's it's a great.
Speaker 4 (40:01):
Time to work on this problem at the moment. I
think also a really good time to kind of get
on board and get yourself acquainted with these new technologies
and if it may.
Speaker 5 (40:11):
May just add a little bit more on top of
that and be a little bit more visionary. I mean,
first of all, we just reiterate it's really exciting to
see these these technologies mature and really the future becomes tangible.
I mean to me, there's really there's a technical pass
to accomplishing all which is discussed. This is a question
(40:34):
of building out the systems, fine tuning them and shipping
them and obviously iterating with partners on them to make
the most most useful. But this is this is super tangible.
The I think the going beyond it, the exciting part
is really also scale, like really getting down the investable universe,
(40:55):
everything that's out there that really allows for way deeper
and overviews, but also things with the market and that
space is evolving. There's also the aspect of the not
building a knowledge base right there's information coming from regulation,
(41:15):
from standards, from science, science research. How do you stay
on top of all these areas? If you want to
invest in a in a specific niche market, how do
you know that? What are you unknowns? How do you
know what is currently being discussed as the most relevant
aspects in nuclear or whatever there is? And there's a
(41:40):
there's a there's a path to to build such a
knowledge knowledge base and leverage it so to inform your decisions,
to derive new frameworks and standards, and practically, for example,
to apply it to things like regulatory risk assessment. Use
a company you in that space, the there is regulation upcoming,
(42:02):
How does this impact your your operations? How much do
you have to adapt? Do you have liabilities legacies on
your emission side and other areas of your company operations?
So this is just very very tangible. I mean also
in terms of data sources going beyond that, controversies, discovery,
(42:25):
red flag analysis, something companies already do out there. It's
also very tangible and many more. We didn't dive into
biodiversity or physical climate risk assessment. All these areas will
benefit from from these technologies. So super super exciting.
Speaker 1 (42:45):
Yeah, without a doubt. I mean in terms of expanding
the knowledge base. I mean the scalability, being able to
compare largely regular data sets, and just gaining efficiencies. Overall,
I think there's plenty of opportunities for AI to continue
to assist in the roles of analyst, companies, bankers, a
(43:05):
whole host of interested parties. So thank you all of
you for this interesting discussion. I'd like to thank Ramita,
Marcel and Andreas for joining me today to discuss sustainable
financing and the role AI can play. And thank you
to all our listeners out there. Keep downloading belove it.
(43:26):
You can find more information on sustainable debt markets, climate investments,
carbon markets, and a whole array of other topics by
going to b I E. S G on the Bloomberg terminal,
which opens up to the Bloomberg Intelligence ESG Research dashboard.
If you have any other questions, please reach out to
one of our BI experts send us an email at
(43:47):
ESG Currents at Bloomberg dot net. Thank you everyone,
Host
Eric Kane
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com