Table of Contents
FINRA's Annual Conference draws the compliance community's full attention every year. This year, artificial intelligence was everywhere - on stage, in the corridors, and in nearly every conversation we had. For MirrorWeb, it was also the week we launched Mira, our new AI communications supervision agent built for regulated firms.
What the Sessions Told Us
The AI Revolution plenary set the tone early. The panel - technologists, compliance leaders, and regulators together - landed on a question the industry is still working through: how do you harness what AI can genuinely do while keeping every decision defensible to an examiner? It's a tension compliance teams know well, and it ran through the rest of the conference.
Compliant Communication in the Digital Age covered familiar ground for anyone in this space - books and records obligations across personal email, text, and social media, and the ongoing challenge of monitoring off-channel communications. The public enforcement posture may have softened under the current administration, but the underlying rules haven't changed. Rule 3110 still requires firms to supervise. The channels advisors use have multiplied. The obligation hasn't.
Modernizing Supervision for Today's Workforce addressed how supervision frameworks have had to evolve alongside distributed teams and hybrid working. The practical question - how do you maintain a supervision program that reflects how the firm actually operates today - is one we hear constantly. The gap between policy and practice is where exam risk lives.
The Rule Modernization session was worth attending for the gift limit reform discussion alone. Proposed changes to gifting thresholds and G&E policies are exactly the kind of nuanced, context-dependent compliance area where keyword-based supervision falls apart. Flagging the word "gift" is not a supervision program.
What We Heard on the Floor
The sessions reflected the agenda. The booth conversations were more candid.
AI was the dominant topic, but the mood around it was more complicated than enthusiasm. Compliance professionals are feeling pressure from boards and senior leadership to adopt AI tools - and in many cases, to adopt them quickly. The problem is that the tools being pushed don't always talk to each other, creating fragmented stacks that add operational complexity rather than reducing it. For people whose job is to manage risk, being handed a collection of disconnected solutions and told to make them work is not a comfortable position.
Data privacy was a recurring concern. Most AI tools require data to leave the firm's environment in some form, and compliance teams are acutely aware of what that means. The appetite to incorporate AI is genuine - but so is the determination to do it in a way that doesn't create new exposure in the process.
The broader consensus has shifted. The question is no longer whether AI has a role in compliance supervision - that debate is largely settled. The conversation now is about what good looks like: which approaches are defensible, which can be adopted without introducing new risk, and which were actually built for this environment rather than retrofitted into it.
Why We Were There
We launched Mira at FINRA because this is the room it was built for. Mira is MirrorWeb's communications supervision agent - built on a decade of native data from more than 1,000 regulated firms, trained to read your compliance handbook as written, and designed to surface genuine risk rather than generate noise. It's the kind of context-aware judgment that makes the difference in areas like G&E — Mira knows a Michelin-starred dinner is $400 a head and that your G&E policy caps at $250. Across initial customers we have seen an 80% reduction in time spent on communications review, while remediating more issues, not fewer.
The conversations we had this week told us the industry is ready. Learn more about Mira here
