A-Team Group’s RegTech Summit New York returns on 19 November for its tenth year, bringing the RegTech community together with regulators and practitioners to examine the latest in AI deployment and market integrity.
Recent news highlights evolving AI policies across Tier-1 firms with Goldman Sachs restricting certain AI coding tools for bankers and traders, while Morgan Stanley has allowed wealth advisers to build their own agents. These approaches raise practical questions about who can develop applications, when reviews become necessary and what firms require before approving wider use.The afternoon programme will address those decisions through discussions of AI supervision and security. Following the US midterms, enforcement and prediction-market specialists will also examine how venues detect abuse and demonstrate that their controls work.
Deployment Decisions
The two banks’ approaches illustrate the choices firms face as staff gain the means to build their own tools. Goldman’s restrictions followed a review of productivity and risk, with access retained for engineers and quantitative specialists.
At Morgan Stanley, advisers have made nearly 50,000 attempts to build agents using natural-language prompts. Most remain experiments. Agents adopted across adviser teams receive accuracy and completeness reviews, with further testing for potential firmwide use.
For compliance teams, the challenge is deciding which uses require review and when that review must happen. An assistant that retrieves information presents different risks from an agent authorised to act across systems. As staff create more applications, firms need review processes that assess those differences and establish what can proceed.
Reviewing each application takes time, particularly when assessors must gather evidence across technical documentation and code repositories. Vincent David, Vice President AI Engineering at Smarsh, will examine how AI could support that work in his keynote, AI That Works for the Compliance Team, Not Round It.
“Agentic AI is moving into regulated workflows faster than most review processes can keep up. Making Well-Architected reviews repeatable and evidence-based is how we get agents ready for regulated customers before they go live,” he says.
David will explore how an AI reviewer can assess repository evidence against best practices and reduce review times from days to hours. The session will address how to test the reviewer, ensure it flags missing evidence and determine whether these assessments could become a required check before deployment.
Prediction Markets Focus
Once systems enter live markets, firms must show how their controls perform under trading conditions. Following the US midterm elections, prediction-market operators face questions about their ability to detect insider activity and manipulation.On 29 October, Senators Elizabeth Warren and John Curtis sought information from Kalshi and Polymarket about safeguards for election-related trading. Their enquiry raises questions about who can trade on sensitive information and how venues identify prohibited activity.
At 2:00pm, David I. Miller, Director of Enforcement at the Commodity Futures Trading Commission, will discuss market integrity in emerging markets with Dessa Glasser, Principal, Financial Risk Group & Independent Board Member, Oppenheimer Holdings Inc. The interview will examine enforcement priorities, and the surveillance capabilities firms need as new products and technologies develop.
The panel, Managing compliance, surveillance and risk for prediction markets and event-based contracts follows at 2:20pm, moderated by Tölt Strategies CEO & Founder Dorothy D. DeWitt, with Kalshi’s Sudhir Jain and Novig’s Elie Mishory.
Together, these two sessions offer enforcement and operator perspectives on identifying suspicious trading, investigating potential abuse and documenting decisions. For attendees, the post-election discussion provides an opportunity to examine how surveillance connects trading activity with information that could influence a contract’s outcome.
Supervisory Expectations
As firms use AI to support surveillance and other compliance decisions, supervisors will need to understand how they oversee those systems. At 3:00pm, Senior Risk Advisor and Educator Penny Cagan will interview Nick Krafft, Executive Deputy Superintendent of Research and Innovation at the New York State Department of Financial Services. Their discussion will examine accountability, oversight and third-party AI risk as applications enter production.
The 3:20pm panel: AI Innovation is outpacing security: Managing cyber, third-party and model risk governance moderated by David Hirsch, Partner, McGuireWoods LLP will extend that discussion into cyber security and model risk. A supplier’s model update or a change in an agent’s permissions may require firms to revisit their original assessment. Panellists will examine how organisations manage those dependencies alongside threats such as prompt injection and data leakage.
These sessions return to the challenge Smarsh’s Vincent David raised in his earlier keynote: how firms maintain evidence that supports deployment decisions as applications change.
Join the Discussion
Attendees can explore these implementation questions with peers during the 4:30pm champagne roundtables. Discussions will cover regulatory reporting architecture, perpetual know-your-customer monitoring, AI governance and AI-enabled surveillance.
Bring a challenge from your own organisation, whether approving a staff-built agent, testing an automated review or assessing surveillance effectiveness. The sessions offer time to compare approaches and discuss what has worked, followed by networking drinks.
Join RegTech Summit New York on 19 November 2026 at @Ease, 605 Third Avenue. View the full agenda HERE, and secure your spot HERE.
Subscribe to our newsletter
