EXCLUSIVE: "Power Points" - Ravi Nemalikanti, Abrigo in 'The Paytech Magazine'

Community banks and credit unions are key to local economies in the States. Abrigo is helping to supercharge them with a new agentic experience

Small Banks, Big Impact . That was the title of a report by the Federal Reserve Bank of St Louis on the impact that community banks and credit unions were having on the regional economy. It illustrated it with some key figures: in 2023, small business loans made up 12.6 per cent of assets held by local banks with $250million in assets or less. That compared to just 3.6 per cent at banks with more than $10billion on the books.

The author concluded that was largely a reflection of the proximity of these smaller institutions to their customers – who are often, literally, just down the road – and how that allowed them to make a better judgement call than those more remote lenders. The previous year, according to the report, 82 per cent of small-business applicants were at least partially approved for loans from small banks. Just 68 per cent of small business applicants received at least partial loan approval from big banks.

But there is a problem. Operating conditions are becoming increasingly difficult for community banks and credit unions in the U.S. due to mounting regulatory, technological, and competitive pressures. Their numbers are shrinking year on year. If they are to continue to be the flywheel for economic growth, they need to find more cost-effective models, without sacrificing the very thing that distinguishes them from their bigger rivals: local knowledge and strong relationships with the communities they serve. 

For that, they still need to rely on super intelligence of the human kind. That apparent paradox is at the heart of Abrigo's vision for the future of lending at community banks and credit unions. The U.S. fintech, whose technology is already deployed across more than 2,400 financial institutions, made its new Abrigo APX (Abrigo Agentic Platform Experience) generally available this autumn to automate and scale lending workflows at these institutions. Abrigo APX is capable not simply of helping a banker complete a task, but of understanding an objective, drawing together data, policies and context, and moving work across an entire lending process.

Abrigo believes that by reducing the time that bankers spend on paperwork, admin and fragmented systems, AI can give them more time to understand customers, spot opportunities and become even better advisors, driving more ‘yeses’, more efficiently. In effect, it’s putting sophisticated technology that was once the preserve of the world’s biggest banks at the fingertips of local bankers, democratising lending capabilities and helping to get more capital to small businesses that are struggling to access it elsewhere.

“J.P. Morgan might have 1,100 or so data scientists working day in, day out. The institutions we work with, thei entire workforce might be 200 or 300 people,” says Ravi Nemalikanti, Abrigo’s Chief Product and Technology Officer.

He’s aware that when transformational technologies such as AI emerge, smaller organisations like this struggle to develop the expertise and keep pace with competitors whose IT budgets dwarf their own.

“So, that’s why we bring it home, apply it and build on-ramps to new technologies for them,” says Nemalikanti.

A 2026 survey by the American Bankers Association (ABA) found that technology limitations and integration gaps were among community institutions’ biggest barriers to growing commercial lending.

Abrigo APX is ‘not about an AI feature that’s just bolted on to an existing platform’, says Nemalikanti. It paves the way for a step change in how these smaller banks operate.

“Where conventional AI might extract information from documents, answer a policy question or generate a credit memo, the banker still determines what comes next and moves work between systems. With an agentic platform, you’re helping the entire institution complete an outcome. Think of that as an operating model shift.”

Abrigo’s technology understands the desired outcome and surrounding context, breaks work into steps and coordinates specialist agents across the workflow. Nemalikanti describes these assistants as being closer to ‘digital humans’, and, just like humans, when a task goes beyond their pay grade, the system escalates it to a superior rather than improvise a solution.

“Humans are getting elevated to a higher threshold of intelligence where they have to apply a level of judgement,” says Nemalikanti.

Agentic versus automation

A surprisingly large proportion of modern lending remains administrative. Abrigo estimates that agentic AI could reduce manual work in lending operations by more than 40 per cent. Nemalikanti himself estimates that 60 to 70 per cent of the time surrounding some lending decisions sits outside underwriting itself. 

“It is all around pushing digital paper from one screen to the next,” he says.

He cites the example of a local manufacturer seeking financing for new equipment. An Abrigo agentic workflow could recognise the loan type, determine which financial documents are required, identify missing information and communicate with the borrower where authorised.

Other agents could extract and reconcile data, calculate cash flows and ratios, examine existing relationships and lines of credit, incorporate relevant fraud, identity or sanctions information and test the case against the institution’s policies before assembling a credit package.

In purely automated processes, small errors can cost big delays. Receiving the wrong year’s tax document is one example. Traditional automation can flag the problem and wait for somebody to intervene. An agent can identify it and immediately request the correct document, even at 7pm when the small business customer is still working but the branch has closed.

“We don’t want to lose time,” says Nemalikanti. “We want the time between a loan application being submitted and cash being dispensed to be as little as possible.”

That can be a lifeline for the kind of small businesses these banks and credit unions deal with. 

“Maybe the customer’s refrigerator is broken,” says Nemalikanti. “Which means their café is going to lose sales until it is replaced.”

But the U.S. Federal Reserve’s 2026 Small Business Credit Survey found only 42 per cent of applicants received all the financing they sought, while 22 per cent received none. The proportion seeking finance from online fintech lenders, meanwhile, rose from 17 per cent in 2020 to 29 per cent in 2025, although applicants approaching small banks were more likely to receive all the financing they sought than those using any other type of lender.

Bank and credit union borrowers also reported higher satisfaction than customers of online lenders and finance companies. Rather than diluting relationship banking, Nemalikanti believes that technology can remove the time-absorbing administration that prevents more of it from happening.

“The relationship between our customer and their customer – it’s not the paperwork,” he says. “Our goal is that automation should strengthen and deepen that relationship.”

Abrigo has already launched Ask Abrigo, which Nemalikanti describes as something akin to putting ChatGPT across an institution’s customer relationships. It can rapidly pull together a customer’s history, institutional data and industry context for a relationship manager.

“Maybe there’s a different way to structure the loan,” he says. “Or a different piece of collateral they need to bring to increase their chances of approval.”

AI may allow institutions to act before borrowers even ask. If data reveals a recurring seasonal cash-flow squeeze, for example, a bank could proactively suggest a line of credit while the business is strong enough to obtain it, or find a better way to structure finance.

“That is what a trusted advisor does,” says Nemalikanti. “Because you know the business so well that you can actually offer that line of credit even before the customer sees the issue.”

This is where productivity potentially becomes innovation and opportunity. Rather than simply reducing headcount or processing the same loans more cheaply, institutions can redeploy capacity towards deepening relationships, identifying overlooked borrowers, developing products and pursuing growth.

That gives community institutions competing against fintechs and national banks an opportunity not merely to automate, but also gain leverage they could never afford to build internally while at the same time freeing up their people to concentrate on what local institutions do best.

There is, nevertheless, an uncomfortable other side to making lending faster. The National Credit Union Administration says overall delinquency and rolling 12-month loss rates are at their highest levels in more than a decade, which has made underwriting, credit administration and portfolio monitoring supervisory priorities for the NCUA in 2026.

Importantly, Abrigo's agents get banks and credit unions to the point of decision-making, but no further. The most sensitive as well as the most impactful part of the lending process is so important, says Nemalikanti, that only human intelligence should have control over it.

“We are purposefully staying away from the credit decision itself,” he says.

While that remains governed by deterministic policies and accountable humans if, as Nemalikanti says, 60 to 70 per cent of the work around it can be transformed, that is nevertheless an enormous prize.

Every one of those steps in the process towards a decision is underpinned by strong governance, built into the APX platform. Policies must be translated into structured thresholds, evidence requirements, approval authorities, prohibited actions and exception paths.

“You do not create reliable guardrails by just uploading a policy PDF and hoping that the platform got it right,” says Nemalikanti.

And if information is missing or policies conflict, the agent must stop.

“It shouldn’t manufacture any kind of facts just to bridge to a decision,” he says. “I would much rather float up 90 per cent of the decisions to the humans than try to get to a 90 per cent rate that’s straight through the agentic platform.”

Beyond lending

Lending is only the beginning for Abrigo’s agentic platform. Nemalikanti ultimately sees the company becoming the ‘trusted intelligence and operating partner’ for banks and credit unions, also embedding AI across portfolio management, credit risk, financial crime and customer intelligence.

It can potentially bring deposit data, lending relationships, fraud activity, anti-money laundering information and CRM data together around the same customer. Nemalikanti’s vision is for that connected view to become an intelligence layer, running through an institution.

“We could be the main door into insights into their customer base,” he says.

If that means viable small businesses are identified earlier, understood better and connected with capital more quickly, local institutions will continue to be the backbone of their local communities.