Bal Shukla on Why Agentic AI Is Rewriting the Rules of Financial Services

Jeff Kavanaugh:

I'm Jeff Kavanaugh, head of the Infosys Knowledge Institute, here at the Infosys Connect conference in Los Angeles. Glad to be here with Bal Shukla, who heads transformation and AI for financial services. Good to see you.

Bal Shukla:

Good to see you, Jeff. Thanks for having me here.

Jeff Kavanaugh:

You bet. Well, the word of the day, topic of the day is AI, especially unlocking value from AI. And financial services, what's top of mind now in the next level of AI and value?

Bal Shukla:

See, financial services has always been into AI space for quite some time. So AI is not new in financial services space in financial services space as we call.

Jeff Kavanaugh:

But still evolving.

Bal Shukla:

Exactly. But what has happened now is we have got this GenAI, agentic AI, which is going to take next leap of faith what's going to happen and all. And that is changing the business model completely from how things can happen from today versus tomorrow. So the biggest thing that's happening in FS space now is reimagine every process that they are working on today. That's a big change that you can think of.

Jeff Kavanaugh:

That's a big deal for banks who don't like to change things much.

Bal Shukla:

Yeah, exactly. So it start with simple thing like onboarding. A customer onboarding, which will take couple of weeks, can be done in days and hours. That's what people are looking at. If a private banker customer and I want to make sure how can I do my onboarding, my entitlement done much, much faster. So not only you do for me, you do for family officer as well. If I'm a financial advisor, how can I do the job that I can do to you, which is much, much advanced and making more intelligent to you, not just going and going back to spreadsheets and figuring out the information from policy documents and all. So all the roles that we're seeing now is getting completely transformed. From a front office perspective, you look at the call center business, what's going on. That is getting disrupted to it's a conversational AI to start off and then go to humans only as need be, an example like that. Or financial advisors who get enough information available on their fingertip before they come and present to customers now. That makes your life much, much faster to go back from a financial planning perspective and all.

Jeff Kavanaugh:

Well, and the customers' expectations are so much higher. We're used to seeing pretty graphs, instant feedback. And so before it was, okay, great. I got this kind of return. The bank's doing this. Now it has to be pretty and fast and simulations.

Bal Shukla:

And not only this, they want interactive. And then a lot of new things also coming up. In the crypto world coming up, tokenization coming up and all. So looking at how do I go to new asset class completely. What should happen to me if I do this class now? And the most important thing, in order to do this part, it will take years to do that piece because there is so much of legacy platform behind it. Code banking platform, you know, mortgage platform, your payments platform, wealth platform, all were on legacy mainframe and arcane code base. What they're saying now is, oh, you know what? My rules, my policies were all embedded somewhere. Now can I expedite that piece from three years, five years online to next 18 months? How much change can I drive now?

Jeff Kavanaugh:

You have to make what was hidden and implicit explicit and relevant.

Bal Shukla:

Exactly. It's almost like an iceberg. FS space had an iceberg where what you see at the top is a very small, huge legacy. The iceberg is coming up now.

Jeff Kavanaugh:

Oh, and by the way, that financial institution may have had a bunch of acquisitions.

Bal Shukla:

Absolutely. So. That makes even more complex. Yeah. You have got five core banking softwares running underneath you. And then what you do now? And now you're looking at how do you strip off, hollow the core. You hollow the core, what can happen? How can I change my branch experience completely? What can happen now? And the fraud is a big, big thing for us. And that's a big thing on trust. How do you make sure, you're making sure that you're giving safeguard to the customer by ensuring that you allow me to avoid any fraudulent activity. But if I come and complain you from a complaints management, you're able to resolve my complaint much, much faster now. So dispute and claims management becomes much faster for me. What it takes me five hours, eight days to come back to me, can you do within a day to me now? If I come and complain a transaction, can you do complaint much faster now? So all those things to this point of customer experience and expectation is coming down, and you want to bank with someone who can give you this peace of mind all the time. One important thing that's coming now as well in the peace of mind is we charge you NSF fees, non-sufficient delivery fees because you didn't have enough balance in the account. Customer is saying, why do you have to do that? You exactly know how much money I have in my account, how much I do You know payment for credit card, for mortgage, for all the things in a month, how much I get You exactly know which day, what comes in. Give me a heads up. If I would do it, give me a heads up a week back, your account is low. Auto transfer, do whatever. And auto transfer, do it all. So there's a lot of agentic work is coming around. That's why agent commerce is coming as well. How can I do agentic commerce now, where you have my permission to do something at this level The control is, of course, given with human in the loop as it goes beyond certain entitlement and limits. How can you do that?

Jeff Kavanaugh:

Like auto-pay on steroids, yeah.

Bal Shukla:

Yeah, auto-pay on steroids. Yeah. Huge opportunity in financial service space. There's all the personas, An RM persona, financial advisor persona, in retail banking, coming to payments and all as we do customer experience, customer service center, all persona getting transformed in financial services space. That's why you see every AI lab partners, If you talk to them, they say, 80% of my client base is finance services. That's a general trend I see. It's coming from financial services in some shape and form. Because they're the first-mover advantage. They want to try it out because they've been doing AI for a while. They have these challenges, and it's also market driven. If two players have done, two have not done, it's an impact to them. So rather everybody wants to be on the same page.

Jeff Kavanaugh:

Yeah, you can be making incremental progress on an absolute basis, but if your competitor is relatively doing it faster, you're losing out.

Bal Shukla:

Yeah. Exactly. The example I did, a private banker example, if I get the same experience, same product capabilities from another entity, who do I go to?

Jeff Kavanaugh:

Right. You spend your workdays thinking about transformation. What are the essential aspects that our financial companies should be thinking about for transformation?

Bal Shukla:

There are three very important things that every company has to think of in transformation. First thing is, lay the rails right, because your rails are very important. It's almost like how you lay the rails right for any transformation that you do. The rails are very important and to set up, which means the foundations have to be set right. One of the key component of foundation is change management as well. Which is like you let the organization structure set right with the right mindsets. The change is very, very important in the transformation space, and change is happening to every responsibility, every role, every process that's coming up. So people having this approach of saying, look, it's going to change my own way of working for the right benefit, it's very important. Third important thing I would say to you, change will not come on day one, everything will be perfect. It will not be. It's a gradual progress, and you have to believe in this, be with this to do that. Like hallucination, as we call, right When you come with process, you redesign process, suddenly, oh, something is wrong. And I've seen examples, certain banks who took six months to do contact center knowledge management transformation, just knowledge management. Because the document that I'm using is probably not right. Some document is in my head. When I come back and use document, say, oh, you didn't give the right information. I have so much in my head. So how do you bring back the knowledge from a head back to knowledge document? That comes through agentic way of working. It takes a while. So important thing in financial services for transformation is three important things. First is set the foundation rails right. With the tools, technology, everything has to come right. B, set the right change management operating model right. Third is the ability to go towards it has to come in a manner which you call lab-based work, experimental-based work. Lab, iteration, learnability. Iteration. Be open to some failure points. Don't think it'll be perfect day one, but be with this. It will just flywheel effect will come. These to me are the three key ingredients to make it happen. Believe in this and pace it up. It will have flywheel effect very, very soon.

Jeff Kavanaugh:

We talked about AI a lot. What about the people aspect? What is the impact on the workers in this whole relationship with the machines?

Bal Shukla:

So I mentioned about change management. That's the 80% of work is change management in this case. Because the work that every individual is doing from the whole hierarchy, from the CEO office all the way to the lowest manager level, every job is changing completely the way you have to augment with AI today.

Jeff Kavanaugh:

That's hard for people to internalize that much change.

Bal Shukla:

It might be a little bit, right. A financial team may understand my job is to only make spreadsheets, and AI will going to help me in this way. That's not enough. You talk about scenario planning. AI can help you do scenario plan much, much faster with all the attributes you can look at. And you can augment your thinking with this. As you augment thinking, it becomes reinforcement learning now. Next time it has got the knowledge as well back into it. So next time it does much better on this. So every role. So what is happening in the organization is that first is embrace technology AI. That's one thing to do, which is lab-based, experiment-based work to do. Train everyone to understand AI, not just AI at the higher level. Go down and do it yourself. My mantra is very simple. Until you do it yourself, you'll not understand the power of AI today. For personal use, you do it. Then you start using for professional use as well and apply it, and then you say how much power is there. The other day someone asked me a very simple example. I want to do a presentation to client, and in the banking world, presentation to your commercial banking customer, to your financial advisor, private banker in our group. How do you do about it? Traditional approach, you get all the data, put into PPT, and then prepare it and think through the story and go back. Today, you can get the data, give it to one of the LLM providers, get the information back from them. It's much better story now because it has got the ingredients of all the intelligence of people have put in. And now you apply your thinking to make it further productive. As you do it, apply it next time. So it's a very much template-driven approach going to come across. But most importantly, people have to try it themselves. So in every organization, what we see is technology is shifting very fast. This is a gap I see. Business is coming little later. So technology, of course, everybody should learn some of the technology, which AI, augmented technology, get GitHub, Copilot or use Devin, use Claude. Absolutely you have to know Gemini, you have to use all the tools. Business side has to do it as well. That's where why Viber is so important nowadays. Just to believe in it. I'm not telling you vibe and put into production. I'm not saying that. But everybody has to know that what I should think in my head, I can put on paper. It could be interactive. Let's see how it looks like as I get in my thoughts. That cuts down the big time in the organization from six months to a few weeks now. Because that used to be longest pole in my view. Of course, engineering has been solved, but this part of product management used to be a long pole, which you get to shorten now. The benefit of this is what could happen in nine months or twelve months, then go to market, then take another twelve months to test out in the market. Now you can do three to six months. Test out three months, iterate, make new features. The learning is going to be much faster. So that show and tell will help everybody learn faster. And once one organization done it, one group, next group also picks on, third, fourth. You will not believe how soon it can become much faster that entire organization now looking at it and say, how can we be a part of it? That's the way we see it.

Jeff Kavanaugh:

Product management, that's the way to go, right?

Bal Shukla:

Yes.

Jeff Kavanaugh:

Awesome.

Bal Shukla:

You're welcome.

Jeff Kavanaugh:

Well, listen, I really appreciate your time. Thank you so much.

Bal Shukla:

Thank you so much.

Jeff Kavanaugh:

You bet. I'm Jeff Kavanaugh. Unill next time, keep learning and keep sharing.