Global Fintech Interview with David Pope, COO at DeepDive

David Pope, COO at DeepDive shares a few thoughts on how regulated firms can enhance their security to protect customers from online financial crimes in this fintech interview:

___________

Hi Dave, tell us about DeepDive’s latest AI platform?

We’re re-imagining Enhanced Due Diligence so that AML teams don’t have to trawl through Google to build intelligence on their high-risk customers. Instead of manual human effort with a search engine, DeepDive is a knowledge engine that answers the questions that MLROs have before they make decisions.

DeepDive combines five core capabilities: advanced multi-language search across global open-source data and compliance databases, Natural Language Processing to analyse vast amounts of data, proprietary entity resolution algorithms to eliminate false positives, AI prompt engineering to extract key intelligence with full citations, and generative AI to produce comprehensive reports with an interactive chatbot for interrogating findings.

It’s not only 10X quicker but it also improves the strength of EDD BY building a wider dataset of intelligence that can be mined to reveal risk.  

How can regulated firms be more vigilant when trying to protect their brands / customers from financial crime?

Over the last 20 years we’ve built, scaled and exited three companies in identity verification and fraud investigation and over that time I’ve witnessed how regulated firms have moved beyond the tick box approach to preventing financial crime.  It’s no surprise that criminals probe their targets and strike at the weakest point of defence and the firm with the lowest defences in place is the one that will get hit. 

Technology has been the great enabler for regulated firms to move beyond box ticking and build defence in depth. Firms like Jumio or Mitek provide a range of KYC technology that doesn’t harm the user experience.  Companies like Seon or Feedzai offer fraud and money laundering (FRAML) risk engines that spot bad actors when they try to attack. Platforms like Cable enable firms to test and optimise the deployment of their financial crime controls.   

My advice to regulated firms is to keep layering new tech into their stack to protect their brand and customers from financial crime. 

Read More on Fintech : Global Fintech Interview with Mike Lynch, Principal, AI Strategy and Finance Transformation for Auditoria             

How is AI both enabling sophisticated threats while also helping teams protect their users against financial threats today?

It’s a classic arms race. Bad actors are using AI to create convincing synthetic identities and beat traditional KYC systems. No doubt they’re leveraging LLMs to create deepfakes that some document validation and biometric comparison providers will struggle with.

As an aside, I would caution against AI being the answer to everything. Believe it or not, AI is not good at some things!  For example, when looking at a set of raw sources that may relate to the subject of the EDD, it’s more effective to build relevant entity resolution cluster analysis rules than it is to ask an LLM to sort it out. 

The other key trick when using AI is to get the right LLM for the job.  Some models are better at certain tasks than others. As you might expect, we use AI for multiple tasks inside DeepDive; AI to transliterate search terms so that we can spread the knowledge gathering net far and wide when doing EDD that spans multiple jurisdictions; NLP to read and digest hundreds of sources quicker than the human analyst; And of course, generative AI to provide narrative in our reports.   

We’re also careful about AI governance, using version control for LLMs, documented prompt engineering standards, and confidence scoring that provides transparency into AI decision-making.

The organisations that will thrive are those that embrace AI while maintaining human oversight—using technology to extend their investigative reach while applying human judgment to risk assessment and decision-making.

Can you talk about some recent financial crime events from around the world that impacted business outcomes in a big way and the learnings/takeaways you’d like to share from it?

What staggers me is that firms continue to be fined for poor AML controls.  It’s not as if KYC, EDD or transaction monitoring is a new thing, right?!  Globally, regulators issued a third more fines last year compared to the year before.  In 2025 we’ve already seen hundreds of millions of pounds of fines issued this year and last year a large American bank received the largest fine ever.  A whopping $3 billion for poor due diligence and failing to detect and report suspicious activities involving high-risk customers. 

A critical learning is that surface-level compliance screening is no longer sufficient. Many organisations are discovering risks they missed during onboarding when those relationships later generate adverse media coverage or regulatory scrutiny. The challenge isn’t that the information didn’t exist—it was publicly available but scattered across multiple languages, jurisdictions, and source types.

A few thoughts on the current state of financial crimes and urgent to-do’s that businesses must pursue?

We’re at an inflection point. Financial crime is becoming more sophisticated.  But it is detectable if you have the right tools, deployed in the right way. For me, three urgent priorities stand out:

First, eliminate information blind spots. If your EDD processes can’t ingest many sources, some of which could be across multiple languages and jurisdictions, you’re creating exploitable vulnerabilities. Financial criminals specifically target organisations with known geographic or linguistic limitations. This isn’t about working harder—it’s about deploying technology that extends your field of view.

Second, make AI governance a priority, not an afterthought. Many organisations are rushing to adopt AI without establishing proper controls around explainability, version management, and confidence scoring. Regulators are increasingly asking not just what you found, but how you found it. Black box AI won’t satisfy regulatory scrutiny. Ensure your AI implementations provide complete audit trails and transparent decision-making processes.

Third, rethink your capacity model. The traditional approach of scaling compliance teams linearly with business growth is economically unsustainable given both cost pressures and talent shortages. The solution? Thoughtfully divide compliance labour between humans and machines—letting AI handle systematic data gathering and analysis while reserving human judgment for risk assessment and decision-making. This enables you to extend EDD across a wider customer segment without proportional cost increases.”

Catch more Fintech Insights : The Disappearing Payment: How Embedded Finance Is Quietly Reshaping B2B Transactions?

[To share your insights with us, please write to psen@itechseries.com ]