Founded in 2019, Velocity FSS offers a comprehensive AML platform acts as an all-in-one solution for AML compliance and reporting requirements. Offering a FRAML – fraud and AML – platform, the company is well placed to meet the evolving needs of a widening market.
Vineet Mishra, managing partner & chief product officer at Velocity FSS, explained that as a FRAML platform, the company offers real-time sanctions screening and payment fraud monitoring. In addition, the company recently launched its agentic AI pack, which is now in production with its clients.
He explained, “We offer agentic AI investigators to our clients who are generating sanctioned alerts, transaction monitoring alerts, or fraud alerts. We also use AI-driven fraud models to screen payments in real time, predicting whether a transaction is likely to be fraudulent. We also apply AI-based decisioning models to adverse media screening, helping institutions identify and assess potential risks more efficiently.
How AI changes the AML and fraud approach
How is AI changing the way that companies approach AML and fraud prevention? For Mishra, his industry perspective is that its making a good impact amongst its clients, with some of Velocity’s clients using its AI offerings. “They’re seeing a huge impact when it comes especially to efficiency,” said Mishra. “I think there’s a great opportunity right now for AI to come into the efficiency and productivity play, and these are the much bigger plays at the moment.
An area where Mishra is already starting to see growing adoption is around AI decisioning. Whilst appetite for decisioning is still lower that efficiency and productivity, one of the company’s clients has adopted the decisioning offering, and in his view, is starting to go parallel.
“From our client perspective – we work with a lot of small to medium sized clients like beginning money transfer companies and community banks – whilst the appetite is still less, we’re starting to see more uptake,” said Mishra.
The FinCrime challenges
What are for Mishra the biggest challenges that financial institutions face in detecting financial crime in 2026?
Mishra believes the challenge for financial institutions remains fundamentally unchanged: criminals are “always one step ahead”, forcing firms to continuously adapt their financial crime controls.
However, he argues that smaller institutions face a distinct set of challenges. Community banks and smaller financial institutions are often still trying to keep pace with wider industry developments and may not have access to the same capabilities as their larger peers. While many financial crime solutions are available across the market, they are frequently built with enterprise-scale organisations in mind.
“The products and solutions are available for everybody,” he says, but in practice, they are not always accessible to smaller institutions. In some cases, enterprise-grade tools are simply out of reach financially, while in others they are not designed for organisations operating at a smaller scale.
As a result, many smaller institutions continue to face a gap in both technology access and adoption as they look to strengthen their financial crime defences.
Balancing innovation and regulation
A pivotal balancing act to consider how organisations can balance AI innovation with the expectations of regulators.
In the view of Mishra, when the regulators come in, the only important thing they look for is explainability and auditability. In this case, if a firm is building AI or offering AI which offers enough explainability and the auditability is fine for the regulators, AI doesn’t throw up many challenges.
However, the challenge that lies for regulators, Mishra states, is the explainability and auditability of AI. “Regulators don’t want a black box. For example, you put AI in an alert investigation – they don’t want to know just that the investigation looks false positive. They’re looking to see what you did to arrive on that recommendation as to why it looks like a false positive. That’s when your AI is accepted.”
Agentic AI opportunities
Companies in the financial services space are currently grasping with an important question: what is the greatest opportunity for agentic AI in financial crime operations?
Mishra is succinct in his belief that investigations is the greatest opportunity, as this has traditionally been the most people-hungry process.
“The moment you start to grow, and this is a bigger problem in the FinTech space, they come in, start to grow fast, their alert volume grows and they need to pump in the people to look at those alerts. This is where AI is, and can continue to play a major role,” he said.
Next-gen AML and fraud
As AI continues to evolve at a break-neck pace, areas such as the AML and fraud market will need to keep up. What capabilities will distinguish the next gen of such solutions?
Mishra raises the example of stablecoin, which is a backed currency and means it is protected. Right now, he views the future solution in financial crime prevention with AI as one where you can manage fiat and digital currency – especially stablecoin – together, as a future will arise where users can have both such currencies in one bank account.
With this, monitoring of the customer will become even more critical. Mishra raised the recent acquisition of BVNK by Mastercard, with the former offering a one payment system where both currency types can be managed under one payment rail.
He added, “That’s a unique offering, and I think that is what the future is looking like, combining everything into one aspect, as customers are now spending in both fiat and stablecoin currency, so you need to understand all the spending patterns as this will be the new digital currency.”
Measuring AI success
How should companies measure the success of AI within financial crime prevention? The number one here for Mishra is decision making – and how good the decisions the AI is making.
Critical here for Mishra is the human-in-the loop. “If you read the EU AI Act very carefully, there is one component standing out loud, which is human-in-the-loop.” In nearterm, this for the Velocity CPO is a critical touchpoint for how firms can measure the success of their AI within financial crime.
Future developments
A pressing debate in the industry right now centers around what AI developments are likely to have the biggest impact on AML and fraud over the next few years.
For Mishra, a big part will be building around their data. Mishra made clear that for anyone publicly launching a model, this won’t help.
The way to take this to its conclusion for Mishra is that the most impactful AI development in financial crime won’t be foundation models, but the ability of institutions to build and train AI around their own proprietary data, customer behaviour and risk appetite.
The AIFinTech100, which this interview was a part of, can be downloaded here.
Copyright © 2018 RegTech Analyst
