Author

Conor Burke

Biography

Writing

Speaking

Co-founder and CTO

Inscribe AI

Expertise

Artificial Intelligence

Machine Learning

Financial Services

Banking

Lending

Fintech

Document Fraud Detection

AI Fraud Reasoning

Agentic AI Systems

Explainable AI

LLM-Powered Fraud Detection

Biography

Last updated: May 28, 2026

Conor Burke is the co-founder and CTO of Inscribe, the agentic document fraud detection platform trusted by leading U.S. banks, credit unions, and fintechs. He leads the AI and engineering systems that power Inscribe's fraud detection capabilities, and has spent nearly a decade working on a single hard problem: building AI that can reason through document fraud the way a skilled analyst would.

His technical focus has sharpened around a core insight — that rules-based detection and pattern-matching were built for a world where fraud was predictable and infrequent. Generative AI has broken that assumption. Fraud is now mass-produced, fast-iterating, and designed to pass basic checks. Catching it requires systems that can reason across inconsistencies, adapt to patterns they've never seen before, and explain their conclusions clearly enough for risk teams to act on and regulators to audit.

Conor's work at Inscribe reflects that. He has led the development of Inscribe's fraud reasoning architecture — moving from one-to-one detector models toward agentic systems where LLMs orchestrate multiple tools, validate their own outputs, and surface findings with transparent reasoning. He has also written extensively on what explainable AI actually means in practice for financial institutions: not just a score, but a documented chain of logic that shows exactly what the system checked, what it found, and why it matters.

He writes on the Machine Mind Blog on topics including LLM reasoning, agentic systems, and practical AI deployment in fraud detection. He is a regular presence on Inscribe's Good Question podcast and has spoken at Money20/20 and Fintech Nexus, and been featured on the State of Identity Podcast and Irish Tech News. He is a 2020 Forbes "30 Under 30 Europe" honoree and studied Electronic Engineering at University College Dublin before completing Y Combinator.

What is Conor Burke's area of expertise?

Conor Burke specializes in AI systems architecture, LLM-powered fraud detection, and the practical deployment of agentic AI in financial services. As CTO of Inscribe, he leads the engineering behind document fraud detection systems that reason across inconsistencies, adapt to novel fraud patterns, and produce audit-ready explanations for risk teams.

What is explainable AI in fraud detection?

Explainable AI in fraud detection means a system doesn't just output a fraud score — it documents exactly what it checked, what it found, and why it reached its conclusion. This matters in financial services because risk teams need to justify decisions to compliance teams and regulators, and because analysts need to understand what the AI is telling them before they act on it. Inscribe's AI Agents show their reasoning in real time: what documents were read, what tools were called, what inconsistencies were found, and how those signals add up.

How do LLMs improve fraud detection?

Standard fraud detection models are trained to recognize known patterns. LLMs add a reasoning layer — they can interpret context, cross-reference signals, and identify inconsistencies that don't match any existing rule. At Inscribe, LLMs are used to orchestrate fraud investigations, generate explanations for flagged documents, and improve parsing of complex document types like bank statements. Switching to LLM-powered parsing improved Inscribe's field coverage on non-English bank statements by up to 5x and reduced false positives.

What is the difference between traditional fraud detection and fraud reasoning?

Traditional fraud detection follows a fixed playbook: run a document through a set of detectors, each built for a specific fraud type, and return a score. Fraud reasoning works differently. An AI agent plans its own investigation, selects the tools it needs, adapts based on what the evidence shows, and explains its conclusions. This allows the system to catch fraud types it has never seen before — including the first instance of a new scheme — rather than being limited to patterns it was explicitly trained on.

What has Conor Burke written and spoken about?

Conor Burke writes on the Machine Mind Blog at inscribe.ai on topics including explainable AI, LLM reasoning, agentic systems, and AI deployment in financial services. He has spoken at Money20/20 and Fintech Nexus, appeared on the State of Identity Podcast and Irish Tech News, and is a regular contributor to Inscribe's Good Question podcast.