Six months on an AI-first stack: three lessons for RIAs

Lead-Lag Media, a sales and marketing firm serving financial advisers, ran an AI-first technology stack beginning in late 2025 and evaluated results after six months in production. The firm reports three primary lessons for registered investment advisers weighing automation choices.

The company replaced manual research, list-building, outreach coordination and compliance pre-review with automated workflows. Specific tasks moved to software included pulling Form 5500 retirement-plan filings, deduplicating leads gathered from cold email, LinkedIn and personal outreach, and preflagging potential compliance issues before human review. Lead-Lag reports those back-office automations cut hours that advisors and producers previously spent on manual work and reduced error-prone processes, while leaving client-facing interactions handled by staff.

Lead-Lag attempted to automate prospect and client communications but reversed those automations after observing poorer outcomes. Instant autoreplies and highly polished AI emails produced an artificial tone that prospects noticed, and replies that arrived seconds after an inbound message felt inappropriate even when content was correct. AI-generated personalized first lines did not improve response rates because the follow-up sentence that shows real knowledge of a recipient’s role or firm remained important.

Michael A. Gayed, founder of Lead-Lag Media, recommended, “Automate the boring work, not the client conversation.” He framed that rule as a guide for where firms should focus early automation efforts.

The firm also requires written governance for each AI workflow. Before automation goes live, Lead-Lag documents the point at which machine decision-making ends and human judgment begins. The company says those written boundaries serve internal compliance and client transparency. Under the SEC marketing rule, advisors retain responsibility for communications whether they are drafted by staff, a large language model or a third-party vendor, and that responsibility cannot be delegated to a vendor.

Gayed emphasized the same point on governance: “Each AI workflow should come with an explicit answer to that question, in writing, before it goes live.” He said compliance officers and firm principals must own that boundary.

Lead-Lag warns that decisions about which tasks to hand to software and which to keep with staff will affect a wealth-management firm’s operating model, capital spending, compliance posture and team composition over the next 12 to 18 months. The firm reports that its approach combined targeted internal automation, retained human control over client conversations and documented governance while it tested the AI stack.