Figma’s stock jumped 37.4% last month, and the reason wasn’t a flashy new feature. It was the market rewarding a company that lets AI work directly inside real, sensitive files without breaking the trust of the people using them. That’s the exact tension every financial advisory firm owner feels right now. You want AI to touch client data, portfolio numbers, meeting notes, and compliance files. You just don’t want it touching all of it, all the time, for anyone who asks.

That’s the real story behind “AI with data guardrails.” It’s not about whether AI can help your firm. It already can. It’s about whether you can control precisely what it sees, who it acts on behalf of, and what gets logged when it does. Get that wrong and you’ve got a compliance problem. Get it right and you’ve unlocked 5 to 10 hours a week per adviser that’s currently going into unbillable admin work.

We built our Omni ops agents around this exact principle. Granular permissioning first, automation second. Here’s what that looks like for a firm your size, and where the money actually is.

The three places your firm is losing money right now

Most advisory firm owners in the $1M to $25M range already sense something is off in their cost structure. Revenue per adviser looks fine. Margins don’t. The leakage usually hides in three places.

Meeting prep and write-ups. Before every client review, someone has to pull portfolio performance, check recent communications, and reconstruct where the client’s goals stand. After the meeting, someone writes it all up again. We typically see advisers spending 5 to 10 hours a week on this, none of it billable, most of it repetitive across clients who have similar plan structures.

Compliance documentation. SOAs, ROAs, and file notes are where paraplanner time disappears. A single advice document can eat $3,000 to $8,000 of paraplanner cost once you count drafting, review cycles, and compliance sign-off. Cycle times routinely stretch into weeks, and every week is a week the adviser can’t act on the advice.

Onboarding and KYC. New clients are the moment a firm should feel most efficient, and it’s often the opposite. Document collection, fact-finding, and risk profiling drag a 30 to 60 day onboarding window into the norm rather than the exception. New clients lose momentum before their first trade even settles.

Run the numbers across a firm with 8 to 15 advisers and this typically adds up to $70,000 to $200,000 a year in labor cost that produces no additional revenue. That’s not a rounding error. That’s often the gap between a firm that’s comfortable and one that’s genuinely profitable.

What “AI with guardrails” actually means for your firm

Here’s where the Figma comparison earns its place. Figma’s AI tools work inside design files that are full of intellectual property nobody wants leaking. The reason it works is scoped access. The AI sees exactly the layer, the file, the permission level it’s been granted. Nothing more.

Financial advisory data is a harder problem. You’ve got SOA templates, portfolio holdings, KYC documents, and personal financial details, all subject to regulatory retention and disclosure rules that vary by jurisdiction and by client type. An AI agent that has blanket access to “everything in the CRM” is a liability, not an asset. An AI agent that’s scoped to exactly the record type, the client relationship, and the document class it needs is a genuine productivity tool.

That’s the architecture behind every Omni ops agent we build. Each agent has a defined data boundary. It reads what it’s permitted to read, writes what it’s permitted to write, and every action is logged in a way your compliance team can actually audit, not just trust. This is the difference between “we use AI” and “we can show a regulator exactly what our AI touched and why.”

If you want to see how this fits your firm specifically, the AI audit for financial advisory firms walks through where your data boundaries currently sit and where they need to be before any agent goes live.

The Meeting Prep Agent, end to end

Take the meeting prep problem first, because it’s the fastest win for most firms.

Right now, an adviser or their assistant spends 30 to 60 minutes before every client review pulling together a picture of where things stand. Portfolio performance since last review. Any recent emails or calls logged in the CRM. Progress against the client’s stated goals, whether that’s a retirement date, a house purchase, or a drawdown target.

The Meeting Prep Agent does this automatically, on a schedule tied to the calendar. It pulls portfolio data from your platform of record, scans recent communications for anything relevant, checks goal progress against the plan on file, and assembles a one-page brief the adviser reads five minutes before walking into the room. No adviser has to remember to run a report. No assistant has to manually compile three systems into one document.

The guardrail piece matters here too. The agent only has read access to the specific client’s records, not the full book. It doesn’t retain data beyond what’s needed to generate the brief, and it doesn’t write anything back into the CRM without a human confirming it first. That’s the difference between a tool that saves time and a tool that creates a new audit headache.

Firms running something like this usually get 3 to 5 hours a week back per adviser, which at a blended cost of $80 to $150 an hour is real money, before you even count the quality improvement of walking into every meeting properly prepared.

The Advice Document Agent, and why paraplanners actually like it

The second agent worth naming is the one that touches your compliance documentation directly, because it’s the one owners are most nervous about and, once they see it, most relieved by.

The Advice Document Agent takes the meeting transcript, or the adviser’s dictated notes, and drafts the SOA, ROA, or file note using your firm’s own compliance template. It’s not writing generic advice language. It’s pulling from the structure your paraplanners already use, populating client-specific detail, and flagging any sections where information is missing or where the adviser’s notes don’t clearly support a recommendation being made.

This doesn’t replace your paraplanner. It replaces the first draft, the part that consumes the most hours and adds the least judgment. Your paraplanner reviews, tightens, and signs off, which is exactly the role they should be playing. Firms typically see the paraplanner time per document drop by 40 to 60%, which on a $3,000 to $8,000 fully loaded cost per document is a meaningful shift in how many clients a single paraplanner can support.

The data boundary on this agent is tighter than the Meeting Prep Agent because it’s generating client-facing regulated documents. It only accesses the specific meeting transcript and the client’s existing file, and every draft it produces is logged with a timestamp and a version history before a human ever edits it. That log is what makes this defensible to a regulator, not just efficient for your team.

The Client Onboarding Agent and the 30 to 60 day problem

The third agent worth mentioning runs the fact-find and KYC process for new clients. Instead of an adviser or admin chasing documents over email for six weeks, the Client Onboarding Agent runs a guided fact-find directly with the new client, collects the required KYC documentation through a structured intake, and prepares a clean onboarding pack the adviser reviews before the first substantive meeting.

This one has the widest data boundary of the three because it’s collecting new personal and financial information rather than working from records you already hold. That means the permission structure has to be built around consent capture, secure document handling, and clear retention rules from the first interaction. Done properly, it’s actually a stronger compliance position than a lot of firms have today with email-based document collection, where sensitive files sit in inboxes indefinitely.

The business case is straightforward. Every week shaved off onboarding is a week sooner a client’s assets are working and a week sooner your firm is earning on the relationship. Firms running structured intake agents typically compress onboarding from the 30 to 60 day norm down to 10 to 20 days, and the client experience improves because they’re not chasing forms, the process is chasing them.

Why this isn’t a generic AI rollout

Every one of these agents sits inside what we call Omni ops, the operational layer of how we build AI into a business. It’s worth being clear about what this isn’t. It’s not a chatbot bolted onto your website, and it’s not a general assistant with access to your whole tech stack. If you’re curious about the difference between operational agents like these and other categories such as voice-based client contact or lightweight internal apps, our Omni ops page breaks down where each type fits, and how they differ from a voice agent built for phone-based client interaction.

The reason the data guardrail conversation matters more for advisory firms than almost any other vertical is straightforward. You’re bound by advice documentation rules, KYC and AML obligations, and often a licensee or dealer group compliance framework on top of your own. An AI agent that ignores those boundaries isn’t a productivity gain, it’s a new source of risk that lands on the adviser’s license, not the vendor’s balance sheet.

This is why we start every engagement with an audit rather than a build. Before any agent touches a live client record, we need to know exactly what data sits where, who currently has access to it, and what the compliance team needs to see logged. That’s not a slide deck exercise. It’s 60 minutes with your operations lead going through your actual systems.

If you want to see what that looks like before committing to anything, Book a 60-min Omni Audit and we’ll map your specific leakage points rather than talk in generalities.

What the audit actually produces

We keep this deliberately simple because most firms have sat through enough vendor pitches to be tired of them. The Omni Audit runs 60 minutes, and it produces three things. First, a map of where your advisers and paraplanners are losing hours right now, benchmarked against what we typically see in firms your size. Second, a data-access assessment that shows exactly what an agent would and wouldn’t be permitted to touch given your current compliance obligations. Third, a written estimate of the dollar impact, tied to your actual headcount and fee structure, not an industry average pulled from a case study that doesn’t resemble your firm.

No deck. No multi-week discovery phase. If you want the fuller picture of how this fits alongside other AI use cases we’re building across advisory firms, our insights section has more detail on specific agent builds, and the guides library covers the operational side of getting a firm ready for this kind of change, including data governance basics that tend to matter most to compliance officers.

The math that should drive the decision

Strip away the AI framing for a second and just look at the numbers. A firm with 10 advisers losing 6 hours a week each to meeting prep, at a fully loaded cost of $100 an hour, is spending roughly $312,000 a year on work that produces zero incremental revenue. Add paraplanner time on advice documents and slow onboarding cycles, and the $70,000 to $200,000 leakage band we see across this vertical starts to look conservative for larger firms.

None of that requires replacing staff. It requires giving your best people back the hours they’re currently spending on work a properly scoped agent can do faster and with a cleaner audit trail than a rushed human process usually manages anyway.

The firms that move on this in the next year won’t be the ones with the flashiest AI story. They’ll be the ones who scoped the data access correctly from day one, kept their compliance team in the loop, and let the agents handle the repetitive 80% so advisers could spend their time on the judgment calls that actually justify the fee.

If that’s the direction you want your firm heading, See Omni for financial advisory firms and take a look at what the first 90 days typically looks like. Or if you’d rather talk through your specific setup first, Book my Omni Audit and we’ll go through your numbers directly, no generic pitch required.