Until now, you've been making your lending, servicing, and collections decisions based on balances, payment histories, and third-party data like scores and credit reports.
That's been fine, mostly because you didn't have anything else.
But generative AI is changing that, with immediate insights into what's happening in your customers' financial lives, right from the interactions you have with them every day.
We chatted with Shantanu Gangal, CEO of Prodigal, and Scott Hamilton, Prodigal's Banking and Lending Strategy Leader, to see how you can use new AI tools to leap ahead by answering questions you've been wondering about for years.
Shantanu:
"For the last 50 years or so, all credit decisions have been governed by a standard set of third-party resources. And some of them were definitely cutting edge and ahead of their times when they were introduced.
Those sources rely pretty narrowly on your ability to make a payment, but don’t take into account all of the other kinds of interactions you have with your lenders and banks throughout the course of a month.
We are really excited to expand how lenders, banks, and financial services at large understand their customers and borrowers by looking at not just a series of dots and dashes on the payment file, but also more holistically - a 360-degree view on all interactions they have with these customers."
Scott:
"It's historically been quite difficult to truly understand where the customer is sitting, what their needs are, whether they're happy, what they plan to do next.
I was at a top 25 bank. I think the issue at the time was, “What was the purpose of the call?”
It was a very simple question that was incredibly difficult to answer. The tools that they had, and virtually the whole industry has, were old-school tools that were helpful, but not super accurate.
The models that we've built at Prodigal together with the data that the models have been trained on - repeating that hundreds of millions of times - the accuracy rate has now exceeded a human, while other alternatives are sub-50%.
That unlocks all these insights that are truly actionable, that you don't have to go back and validate."
Scott:
"The conversation often starts with call QA automation. And very quickly, when you get comfortable with the accuracy rate of models, you learn what is now possible, another set starts to come out in the servicing case, understanding the reason for the call. As simple as that sounds, that's really hard.
And then that opens up automated complaint capture, true calculations of first-call resolution rate, life events for cross-selling and upselling and churn. These are all problems that I wrestled with for a couple of decades. But the data was impossible to get.
In the collections space: being able to extract new data attributes that are predictive of payment is one bucket for segmentation strategies, dialer strategies, treatment, assignment, legal, etc. And then being able to share that data with your agencies, your network, so that every time you transfer the account, the new firm is not starting over.
You can extract real substantive agent and borrower behaviors, the reason for delinquency, their sentiment, their duration of hardship. These are things that are now available that are incredibly powerful."
Shantanu:
"A lot of lenders specifically rely very heavily on credit scores, which is fine for a particular segment of borrowers, but a lot of borrowers are thin file or no file, or they have limited histories. They just can't get the advantages of it.
But that is where the revenue lies: being able to identify insights and signals before they are common knowledge is how we help our customers make money.
It’s about being contrarian and right.
In a matter of six to nine months from a particular life event, someone getting a new job, someone moving to a new city, someone getting married or looking for a new car, that becomes public knowledge. All of that intelligence is baked into your credit score eventually.
But our ability to tap into it immediately and provide it to our customers, six months ahead of it being public knowledge, is what gives us an edge. And because we hear it from the horse's mouth, because we've captured and verified these things at the moment they happen, and we are able to stitch it together across all engagements you've heard, gives us verified information.
And so being contrarian and right helps us make our customers make more revenue, reduce risk, and the added fact that they don't have to rely on a lot of people doing repetitive, mundane tasks, reduces cost as well.
We help them get ahead of things that become complete and kind of public knowledge six to nine months later. And that helps them get into a relationship with the great customer, that helps them get out of a relationship with the not great customer, and really drives profitability all in all."
Shantanu:
"We are agnostic to the type of interaction you're having, because we do a lot of work at our end to standardize and catalog it in a very simple-to-understand manner.
You might have a text message exchange, you might have an email exchange or a phone call. And we can quantify all of these interactions into standardized forms. What we do is understand the intent of the conversation, which is why we informally call all of our work, all of the data science and machine learning work that we do, our AI Intent Engine.
All of our statistical expertise is deployed to understand the underlying intent of the interaction. And that may be over an email, phone call, text message, web chat, or any communication. And in understanding intent, we are able to find insights as to who's unsure versus who's sure, who is expressing a lot of hope for the future, who expects a change in their life station in the next 30 to 90 days.
These are things that banks and lenders really, really care about, and our ability to capture things that are mentioned in passing, our ability to understand things that may even be unsaid. Just a catch in your voice or intonation is something we're able to capture, expound upon, quantify, and make available to financial services in a way that they can act on it.
Our ability to do it stems from, firstly, a very strong team, but also our understanding this space extremely well. And then learning and repeating and training and testing across close to 300 million calls at this point, that gives us an edge that continues getting deeper and wider."
Shantanu:
"We are built on a very scalable architecture. That means as you grow as a business, you capture data from more consumers, but you also capture more data per consumer.
It is also completely modular. So for example, if you decide tomorrow that you want to expand into a different line of business, we are able to give you an ability to get insights into that line of business or that book of business with a click of a button or a flip of a parameter in an API call.
Those are the kinds of things that make our modern software stack easy to use for a fast-growing business, any fintech or any consumer finance application that's looking to scale pretty quickly."
Scott:
"It’s all the tools and all the data and the hundreds of millions of calls and insights that we can create is what I find transformative.
The top five banks don't have this. But it's available to the smallest firms at the same price with very little lift, and very, if not any, upfront setup. It's super powerful and accessible to the small, medium, and large and the super large all at the same time.
Having built it is the hard part. Now providing access to people to plug into it is exciting."
.svg){kind=icon}
-min.png)
.jpg)
-modified.png)
.png)
.png)
.png)
.jpg)
.jpg)
.jpg)
.jpg)
