On Monday 27th July whilst visiting the Isle of Man I presented to a group of compliance professionals about Human Led, AI enabled Due Diligence. A summary of what I said appears below and there is also a link to the slides in the footnote to this text.
The due diligence landscape has changed dramatically. The web is now bigger, noisier and harder to navigate than ever, and that creates a serious challenge for compliance professionals, investigators and enhanced due diligence teams who are expected to turn vast amounts of information into clear, defensible intelligence.
That was the central point behind my presentation. Because whilst AI has value, it can only be when it is used with discipline, context and professional judgement.
Due diligence, at its core, is about conducting careful pre-relationship or pre-transaction enquiries to identify risk, avoid poor decisions and protect organisations from regulatory action, financial crime exposure and reputational damage. Enhanced due diligence goes further, demanding credible source selection, proper records, analysis, evaluation of information, and the investigative skill to connect facts in a meaningful and defensible way.
The difficulty is that investigators are now operating in what the presentation describes as a “data tsunami” far too much information, poorly maintained sources, digital noise and material that is often difficult to verify. Front-line feedback reflects the same pressures: not enough time, challenges with foreign-language material, record-keeping burdens, and the constant struggle of turning raw information into usable business intelligence.
This is where AI can help. Used properly, it can support faster first-pass research, help identify potentially credible online sources, assist with translation, generate step-by-step guidance, improve drafting quality and save time in report preparation. It can also help investigators think through more complex cases, especially where information is fragmented, difficult to access or spread across multiple jurisdictions.
We have to be equally clear about the limits. AI is only part of the solution. Speed and scale are valuable, but trusted due diligence still depends on human judgement: scrutiny of source quality, corroboration across independent sources, contextual analysis, and the experience to recognise what does not look right.
That distinction matters because AI introduces its own risks. Those concerns include hallucinations, single-source reliance, bias, lack of transparency, data security issues and continuing regulatory caution. A polished answer is not the same as a reliable finding, and a convincing summary is not evidence.
Professionals must think like investigators when using AI. That means planning before prompting, deciding exactly what you need, using focused instructions, asking follow-up questions, verifying outputs and challenging the response rather than accepting it at face value. In other words, AI should be handled as an assistant, not treated as an authority.
Compliance-grade research should be framed by specifying credible sources, directing the search toward company registers, sanctions and watchlist data, reputable media, and require corroboration, source citation and clear identification of anything that cannot be verified. That approach does not remove the need for human review, but it does improve the quality of the starting point.
Another important theme is governance. Before using AI tools in any due diligence environment, teams need to understand where information comes from, how accurate and complete it is, where data and outputs are stored, whether third parties can access them, and how much control the user retains. For regulated firms and reputation-sensitive clients, those are not technical footnotes; they are core risk questions.
The FCA’s broad direction of travel in 2026 suggests that existing frameworks may still apply, AI offers significant opportunity, but the risks of fraud, bias, opacity and manipulation are growing, and regulatory capability will need to evolve alongside adoption. For compliance professionals, that reinforces a familiar principle, innovation may be welcome, accountability does not disappear.
AI cannot be ignored, and it can add real value as an overview tool, a drafting aid and a way to improve efficiency. But it should be used cautiously, alongside other trusted sources, and every important output should still be checked, edited and validated by an experienced human investigator.
Would you stake your reputation on an AI result? In due diligence, compliance and investigations, that is the standard that matters. The future is not AI alone. It is human-led, AI-enabled work that produces decisions you can defend.
Colin Tansley – CEO Intelect
Link to presentation can be found here.
