ATAH is an open framework for moving users from AI systems to verified professionals across insurance, legal, healthcare and financial services
Epoq, a leading insurtech that makes legal protection accessible through insurance products, today announced its founder and CEO of Epoq North America, Grahame Cohen, has published ATAH, the Agent to Authenticated Human Protocol, an open framework released under Apache 2.0 and published independently of Epoq.
As consumers increasingly turn to AI platforms for professional advice, there is no accepted standard for how those platforms should connect them with trusted human professionals. Across systems such as ChatGPT, Claude, Gemini and Perplexity, those interactions usually end with a disclaimer or a handoff back to traditional search, creating a discoverability gap.
The AI discoverability gap has two sides:
• Users need a verified human versus seeing traditional search results or an AI guess
• Professionals need fair discoverability as AI replaces search
ATAH is designed to close both ends of that gap, with verified professionals, transparent sourcing and a clean handoff. Without a neutral layer, only the largest brands and marketing budgets may surface inside AI conversations.
Read More on Fintech : Global Fintech Interview with Rob Young, Managing Director – UK at InDebted
“AI is becoming the first stop for people seeking professional advice; legal, financial, medical, insurance, tax and beyond,” said Cohen. “What is missing is the trust layer that gets someone safely from the AI to a verified human expert. That gap is going to become one of the most important trust and discoverability challenges of the AI era.”
Cohen developed ATAH drawing on more than 25 years building digital legal services at Epoq, working with major insurance and financial services partners across the United Kingdom, United States and Canada. The protocol was built with contributors from the company’s team and published by Cohen personally under Apache 2.0 rather than as an Epoq product because trust infrastructure for AI cannot belong to any one company.
How ATAH’s two components work
Discovery. An AI agent asks ATAH to find verified professionals matching the user’s situation. ATAH returns a candidate set with the source and date of every verification attached; out-of-date, unlicensed or otherwise ineligible records are filtered out before the AI sees them. ATAH does not rank, score or recommend; the AI platform decides which candidates fit.
Handoff. ATAH moves the user from the AI conversation to a working introduction with a chosen professional. For regulated work, the professional confirms they can take the matter—conflict check, scope, fees—before any contact details are shared, so the user is never passed to someone who then has to decline. Personal data flows through a transient encrypted vault rather than central storage and is erased once the professional has retrieved it.
Working with professional organizations
Identifying verified professionals from trusted sources, including medical boards, bar associations, licensing authorities, engineering institutions, professional accounting bodies and other credentialing organizations, remains a gap in the AI search experience. ATAH lets those organizations make authenticated member and licensing data available to AI platforms while keeping control over verification, privacy and governance, so platforms can surface real, verified experts rather than best guesses.
“If we don’t set open standards around identity, consent and provenance today, those decisions will be controlled by opaque commercial systems,” said Cohen. “ATAH is published in the open so this gets built once, before paid placement and bad incentives become the default infrastructure of trust on the internet.”
ATAH v0.8 is published as a release candidate for technical review. Developers, AI platforms, standards organizations and professional bodies are invited to review the specification, contribute to its development, and build reference implementations.
Catch more Fintech Insights : The AI Shift in Fraud: Why Banks Need a New Playbook
[To share your insights with us, please write to psen@itechseries.com ]
