Financial institutions are redirecting their artificial intelligence investments from detection systems to investigation workflows to combat rising compliance costs. Baran Ozkan, co-founder and CEO of Flagright, said the operational bottleneck has now shifted downstream to resolving the resulting alerts, even as detection models improve.
The cost of false positives
Banks, fintechs and crypto firms face overwhelming alert volumes despite continuous improvements in detection technology. Every flagged transaction requires manual review, documentation and audit-ready reasoning, even if it turns out to be a false alarm.
"Every single false positive has a cost associated to it because someone actually has to work on it," Ozkan said. "They have to investigate, make sure it's not a true positive, there is no real risk and dispose it. But that all takes time. And time means money."
Ozkan estimated that resolving a single **false positive** costs between $10 and $30. For large institutions processing millions of alerts annually, these expenses scale quickly and strain traditional compliance infrastructure.
Automating investigative workflows
Historically, compliance technology focused on refining detection through better data and behavioral models. However, the downstream investigation layer has remained highly manual. Analysts spend significant time pulling fragmented information from onboarding systems, transaction records and spreadsheets before they can evaluate risk.
Generative AI models are now being deployed to accelerate this synthesis process. "With LLMs, we got a lot more firepower," Ozkan said. "We can process many more signals. We can do basic reasoning. We can write narratives on behalf of humans, but still keep humans in the loop."
Instead of replacing investigators, these tools function as a force multiplier. By applying Generative AI and LLM technology to investigative workflows, institutions can compress the time spent gathering data. A typical Level 2 investigation that once required 15 to 20 minutes of exporting and consolidating CSV files can now have those preparatory tasks automated.
Solving the explainability problem
The integration of artificial intelligence introduces a major hurdle for compliance organizations: **explainability**. Regulators demand clear audit trails detailing how investigations are conducted and why specific decisions are made.
"If you think about LLMs, they are inherently unexplainable," Ozkan said. "In a regulated environment, everything needs to be explainable."
Simply attaching a public large language model to a compliance workflow is insufficient. Organizations must build infrastructure that orchestrates mathematical functions and maps outputs back to standard operating procedures. Finding risk is only the beginning; resolving it efficiently and defensibly is the new competitive advantage.
Why this matters for general management
Management teams overseeing financial operations must account for the hidden costs of false positives, which can drain budgets at a rate of $10 to $30 per alert. Investing in AI requires a strategic shift: rather than solely funding better detection models, leaders should prioritize tools that automate the downstream investigation process. Crucially, any adopted solution must guarantee audit-ready explainability to withstand regulatory scrutiny, making infrastructure design just as critical as the AI model itself.
Your membership also unlocks:
