GenAI in the Enterprise: From Hype to Real Business Value

Artificial intelligence is here to stay, but enterprises can't afford to get it wrong, even though AI has been adopted faster than any other technology.

Most enterprise projects stalled before delivering real value.

In fact, a new report from Moody's shows 70% never move past proof of concept.

Well, the era of AI experiments is over, and CFOs now demand measurable business outcomes.

Can companies ensure success by aligning AI with clear business goals and also build strong data foundations and embed compliance from day one?

Joining me live is Pare Savik, senior director of Gen AI Solutions at Moody's.

Pale, great to have you here.

Thank you so much for joining us.

Great to be here.

Well, we know that there are a lot of headlines regarding AI, but why do you think most AI projects stop?

So I think there's a difference between the mass adoption that's happening with generative AI in the public sector versus enterprises.

Of course this technology is pretty new in terms of mass adoption.

The last 2.5, 3 years with Chat GBT for enterprises they need a lot more rigor when it comes to adopting these Gen AI tools.

So if you look at Fortune 1000 organizations and you look at the average spend in terms of IT, it's probably around about 7 to 8 million across these organizations, but it mainly sits in the innovation budget, in the innovation IT budget, which means it's kind of a nice to have.

It's a learning, it's an experiment, and it's a little bit siloed from the rest of the organization.

What is happening now is this is actually shifting to the central IT budget, so it was about 25% in the innovation budget.

Now it's in the central IT budget and business as usual as well.

So that's a huge shift that's happening.

And the other thing which is really important is the data.

How clean is the data?

Is it standardized?

Is it AI ready? and enterprises need this data to be secure.

They need to be cleansed in order for them to actually use them in their operations.

And probably it's befitting that you're here at the New York Stock Exchange having this conversation because we know that all the publicly listed companies here grapple with this.

So how do you align artificial intelligence with business goals?

So what isn't really critical is that the generative AI that is going to be used or even the agentic solutions is aligned with business outcome from the get-go, particularly when it comes to compliance.

You don't want to do this as an afterthought, right?

And so what that means is we have to look at what are the use cases that generative I can increase productivity.

They can speed up, speed up assessments and that they can really free up employees to have more kind of high value added input into the operations.

So we're talking credit risk management.

We're talking treasury management.

We're talking about KYC compliance and even enhanced due diligence reporting, all of these kind of central operations within businesses that really require a more efficiency, more accuracy, and more speed of productivity.

What you're seeing in the markets, a lot of excitement, right?

And this excitement is really coming from the kind of infrastructure spend that's happening within generative AI.

So we're talking data center spend.

We're talking about energy supply, you know, we're talking about semiconductor factories, and if you look at the hyper scales, they are.

Purported to invest more than $400 billion or going to be investing more than $400 billion in this infrastructure and that's really making the whole market very buoyant.

But when it comes to implementation, it's a lot more difficult to implement that and so that's where the disconnect comes and that's what enterprises are grappling with and what Moody's is providing with their genetic solutions and our entire data estate.

Yeah, and at the crux of this conversation is data, and this is something that you touched upon earlier as well.

So why do you think data is a common failure point here?

Can you tell us a little bit more?

Yeah, so when it comes to data, it's not all data fits all products and solutions.

It really needs to be cleaned and standardized for AI.

The other thing that is incredibly important is there is an interoperability across these data sets.

So if you look at an average agentic solution use case that we are deploying with our with our clients, you're going to be looking at 20 to 25 plus different data sources in those workflows.

So they're very complicated and across any organization, the large organizations, you're going to have more than 500 different data sets.

So these data sets, not only do they have to be compliant, but there also has to be transparency there.

So you have to be able to know where it's coming from, where it sorts.

You need to identify it and then also audit it as well and make sure that there's a there's a connect between all of these data sets so that the artificial intelligence tools can be accurate and they can interplay between these different data sets.

That's really what Moody's doing.

We have a large data estate of more than 5. million financial organizations, both public and private, we have everything from sanctions information, KYC information, news information all the way through to economic indicators, and of course credit credit risk tools which we're very, very known for.

And finally, Pablo, before I let you go, we have about 60 seconds here, so I do want to talk about the role that AI will play in the labor market.

So tell me, how can AI augmented and not replace people?

Yes, it's all about change management, and I think what is probably going to happen in terms of the, let's say the pyramid and the structure of employees within. it may go from a triangle to a diamond, meaning there's going to be a lot more folks that are required in the mid part of an organization to implement these tools, to interpret these tools, to translate these tools, and really it needs to start with employees and training to make sure that it's part of the core processes and that the tools are used like Moody's Agentic Solutions to Increase efficiencies and productivity and so that the human in the loop, the people have the ability and more time to have that more higher higher added value input in the processes, everything from risk to finding new opportunities in the market.

Well, it was wonderful having this conversation with you on such an important topic, so thank you so much for joining me today at the New York Stock Exchange.