The 10% Test: What RIA Sellers Should Ask an Acquirer About AI

An RIA owner comparing a dozen acquirers has a strange problem.

Every buyer has an AI story.

One points to an enterprise technology budget. Another has a new AI leader. A third has a partnership with the vendor everyone is talking about. The presentations are polished. The roadmaps are ambitious. The demos all work.

None of that tells the seller what will happen to the firm they spent twenty years building.

Will the platform make the advisors better?

Will it preserve the workflows clients actually value?

Will it help the firm grow, or simply move everyone onto a larger company's systems?

That is the question underneath the AI conversation in RIA M&A. DeVoe & Company characterized the first half of 2026 as a record for RIA deal activity, so more owners are being asked to evaluate that question now.

Buyers are not paying an AI premium today. Rush Benton, the former head of M&A at Captrust, recently told InvestmentNews that he has not seen buyers apply a premium or discount based on a seller's AI usage. Large acquirers are still buying client relationships and advisor talent, then moving those firms onto the acquirer's platform.

But Benton also said sellers are beginning to evaluate prospective buyers on how well they implement AI and whether they are devoting real resources to it.

Those two ideas are not in conflict.

AI may not be a seller valuation premium yet. It is becoming a buyer-selection test.

I have watched this from both sides: first in RIA M&A, then running a scaled RIA platform. The M&A work taught me what buyers pay for. The platform work taught me what centralization can improve, and what it can accidentally flatten.

The seller's job is not to find the buyer with the biggest AI budget.

It is to find the buyer that understands the 10%.

The 90% Should Be Standardized

Most of an advisory firm's technology should not be custom.

CRM, portfolio reporting, financial planning, cybersecurity, communications archiving, identity management, billing, and the rest of the core stack benefit from scale. A well-run acquirer should be able to buy, secure, integrate, and support those systems better than a $500 million RIA can on its own.

That is part of the bargain. The seller gives up some independence and receives infrastructure, capital, expertise, recruiting support, succession options, and relief from work that has become too complex to manage alone.

AI does not reverse that logic.

Kitces recently examined how AI-powered development changes the traditional build-versus-buy decision for advisory firms. The conclusion was not that RIAs will replace their core technology with homegrown applications. Even the most technically curious advisors are mostly building around their existing software: combining outputs, improving reporting, and solving narrow problems that broad platforms do not address.

That is the right model.

Buy the core 90%.

Build and configure the differentiating 10%.

Govern all of it.

The Last 10% Is Where the Firm Lives

The last 10% is not a collection of decorative preferences.

It is how a firm prepares for a business-owner review. It is the sequence a service team follows after a client loses a spouse. It is the way an advisor brings planning, portfolio, estate, and family context into one conversation. It is the reporting package built for a specific niche. It is the institutional knowledge that lets a second-generation advisor deliver the founder's best thinking without pretending to be the founder.

No national platform will anticipate every one of those workflows for every acquired firm.

AI makes it cheaper to fill those gaps. A firm can now prototype a specialized planning workflow, internal knowledge tool, client deliverable, or data bridge without commissioning a six-figure software project.

But cheap development is not the same as production readiness.

The moment a custom tool touches client data, influences client-facing work, writes to a system of record, or becomes part of a supervised workflow, the questions change. Who validates it? Who approves changes? What data can it access? Where are the records retained? What happens when the underlying model or integration changes?

FINRA's 2026 oversight report is written for member firms, not RIAs. But its technology practices provide a useful operating benchmark: clear accountability, documented controls for AI development and use, strong data management, ongoing monitoring, and continuity planning.

That is why the 10% test is not just about innovation.

It is about whether the acquirer can support firm-specific improvement inside an operating model the enterprise can actually govern.

Seven AI Questions to Ask Before Selling an RIA

An acquirer's technology presentation will tell you what it owns and what it plans to launch.

These questions reveal how the platform actually operates.

1. What will be standardized in the first twelve months?

Ask for specifics. Which systems will be replaced? Which workflows will change? Which reports, client portals, service processes, and integrations will disappear?

Standardization is not inherently bad. Ambiguous standardization is.

The seller needs to know which parts of the current operating model are being treated as redundant and which parts the buyer is willing to understand before replacing.

2. Where can an acquired firm still configure or build?

The useful answer is not "our platform is flexible."

Ask whether the buyer has an approved way to create firm-specific workflows, reports, automations, prompts, knowledge tools, or client experiences. Who can propose them? Who builds them? What testing is required? Who decides whether a prototype is safe enough for production?

A platform that standardizes everything may be efficient. It may also erase the operating details that made the firm worth acquiring.

3. Who owns what gets created?

Custom workflows produce assets: process maps, prompts, business rules, code, templates, reports, training materials, and client-facing deliverables.

The seller should understand who owns those assets, whether the acquired team can continue improving them, and what happens to them if an advisor or team later leaves the platform.

Data portability is only part of the question. Workflow portability matters too.

4. How does AI reach the firm's data?

Ask for the real architecture, not the vendor logos.

Which systems can AI access? Is access read-only or can tools write back to the CRM and other systems of record? How are permissions assigned? What is retained? Can firm or client data be used to train a model? What activity is logged and reviewable?

An AI tool is only as useful as the context it can reach, and only as trustworthy as the controls around that access.

5. Where does human accountability remain?

The buyer should be able to explain which AI-supported outputs require advisor, operations, legal, or compliance review before use.

This is especially important when a workflow creates client communications, changes records, supports planning analysis, or triggers an external action.

"There is a human in the loop" is not an operating model. The answer should identify the human, the decision, the evidence reviewed, and what happens when the output is wrong.

6. How do you know advisors are actually using it?

AI spend is not AI capability. Product announcements are not adoption.

Ask what the buyer measures after launch. Active usage can matter, but business measures matter more: meeting-to-follow-up time, service-cycle completion, CRM quality, advisor capacity, client response time, exception rates, and growth supported by the workflow.

If the platform cannot show how a tool changed the work, the seller is being asked to underwrite a roadmap.

7. What happens when the technology changes?

Models change. Vendors change terms. Integrations break. Products get acquired or retired. A tool that looked strategic during diligence can become unsupported eighteen months later.

Ask who owns continuity, how replacements are evaluated, whether workflows are documented independently of the vendor, and how the platform protects the business from becoming dependent on one tool or one technical employee.

The best technology strategy leaves the firm more resilient, not more trapped.

Three Answers That Should Make a Seller Nervous

There are three common answers that sound better in a presentation than they do in an operating review.

• "We are investing heavily in AI." Investment is an input. Ask what changed for advisors and clients.
• "Our vendors handle the governance." Vendors manage their products. The firm still owns its use cases, data decisions, supervision, and records.
• "Every acquired firm moves to the same model." Consistency can reduce risk and cost. It can also destroy useful differentiation when nobody separates what is merely different from what is genuinely valuable.

The point is not to catch a buyer with a bad answer.

It is to understand whether AI exists as a collection of announcements or as a managed capability inside the platform.

Will AI Eventually Affect RIA Valuations?

Probably. But not because a seller bought the right licenses.

If AI affects RIA valuations, it will be through evidence that the business itself is better: less dependent on the founder, more consistent across advisors, easier to integrate, capable of serving more clients without matching headcount growth, and supported by workflows that survive employee turnover.

Those are not AI metrics. They are business-quality metrics.

That distinction matters because an acquirer cannot simply inherit a seller's experimentation and call it enterprise value. The workflows need owners. The data needs controls. The results need evidence. The operating knowledge needs to be transferable.

The buyer also needs to prove that its own AI capability will strengthen the seller's business after closing. For owners rolling equity or remaining with the platform for years, that may matter more than whether the current transaction receives an explicit AI premium.

The wrong platform decision does not end at closing.

It compounds.

Build the 10% Memo Before the Buyer Meetings

Before evaluating an acquirer's AI roadmap, document the firm you are trying to preserve.

Create a short 10% memo with:

1. The five workflows that most clearly differentiate the client or advisor experience
2. The systems, data, and people each workflow depends on
3. The parts that should be standardized and the parts that should remain configurable
4. The controls and human review required
5. The measures that show each workflow is valuable

Then use the memo to test every buyer.

The question is not whether the acquirer has more technology than your firm. It will.

The question is whether the platform knows how to apply that technology without sanding away the reasons clients and advisors chose your firm in the first place.

That is the 10% test.

See What Your Firm Needs to Preserve and Improve

ThrivAI's AI Readiness Snapshot helps advisory firms identify the workflows creating real differentiation, the operating gaps limiting execution, and the priorities that should shape the next technology or platform decision.

Take the AI Readiness Snapshot

Sources

Sources: InvestmentNews, "AI hasn't moved RIA valuations yet, but it's rewriting the seller's checklist"; Kitces, "Why 'Vibe Coded' AI Tools Won't Threaten Your Tech Stack"; DeVoe & Company, RIA M&A Deal Books; FINRA, 2026 Annual Regulatory Oversight Report: Third-Party Risk Landscape.