85% of Your Clients Want to Know You Use AI. 39% of Firms Say Nothing.

85% of Your Clients Want to Know You Use AI. 39% of Firms Say Nothing.

Here is the math that should bother every professional services firm owner right now: 85% of clients say their firm should disclose when AI is used on their matters. Only 16% of firms actively communicate AI use. And 39% say nothing at all.

That is a 69-point gap between what clients expect and what most firms deliver — and it comes from the most credible data we have. The Thomson Reuters Institute 2026 AI in Professional Services Report surveyed more than 1,500 professionals across legal, accounting, and tax sectors. This is not a vendor poll. It is the clearest picture yet of where your clients stand and what they know about how you work.

The gap is not a compliance story. There is no single federal law requiring most firms to disclose AI use to clients right now. The gap is a client trust and retention story. The client who does not know you use AI is not neutral on the question — they are forming a view of your firm based on incomplete information. When they find out (and they will), the conversation about trust has already happened without you.

---

What the Thomson Reuters 2026 Data Actually Shows

The Thomson Reuters Institute 2026 AI in Professional Services Report (1,500+ professionals surveyed, US and international) gives us five numbers that, taken together, describe an industry-wide failure of communication — not technology.

1. 85% of clients expect disclosure. When asked directly, 85% of clients said professional services firms should disclose when AI is used on their matters. This is not a fringe expectation. It is a near-consensus view.

2. Fewer than one-third of clients know if their firm uses AI. Despite 85% expecting disclosure, fewer than one in three clients were aware whether their current firm — the firm they are paying right now — uses AI at all. That gap does not exist because clients haven't asked. It exists because firms haven't told them.

3. 39% of firms communicate nothing. 16% actively communicate. Firms have sorted themselves into three groups: those who communicate actively (16%), those who have said something vague at some point (the remaining 45%), and those who have said nothing at all (39%). The 39% silence is concentrated in smaller practices — the 5–50 person firm running without a marketing team and without a formal AI communication policy.

4. 40% of firms have received conflicting orders from clients. This is the most underreported finding in the TR 2026 data, and it is covered in detail in the next section.

5. 75% of clients and firm respondents agree: firms should lead these conversations. Not wait for clients to ask. Not respond when challenged. Lead. Three-quarters of everyone surveyed — on both sides of the relationship — believe the firm should initiate the disclosure conversation.

The #1 skills gap named in the Thomson Reuters Institute 2026 AI in Professional Services Report is not technical. It is not AI proficiency, model selection, or data security. It is stakeholder communication and change management — cited by 58% of respondents. Firms have deployed AI tools. They have not built the internal consensus or external messaging to talk about it.

This is a people-and-messaging problem.

---

The 40% Problem: Conflicting Orders

The most original insight in the Thomson Reuters Institute 2026 AI in Professional Services Report is not the headline disclosure gap. It is the 40% finding buried two sections in.

Forty percent of firm respondents have received conflicting instructions from clients: one client contact telling them to use AI on a matter, another telling them not to. Often from the same client organization. Sometimes on the same matter.

Picture what this looks like at a 12-attorney firm. The general counsel at a mid-size company tells you to use AI tools to accelerate contract review. Three months later, the CFO at the same company raises a data security concern during a quarterly review and says they expect you to avoid AI tools on their matters until they review your vendor list. You have no written policy. You have no documented exception process. You have no defensible default when these instructions conflict.

That is not a communication preference problem. That is an operational risk that will eventually produce a difficult conversation — or worse.

Firms without a documented AI policy are exposed to exactly this scenario. A written firm-level AI policy defines your default position, requires client-specific exceptions to be documented, and gives you a paper trail when conflicting instructions arrive. Without it, you are improvising on the fly with client data and client trust at stake.

The conflicting orders finding also reveals where this is headed. Clients are increasingly informed about AI. They have internal debates about it. They are passing those debates upstream to their outside firms. The firms that have a clear, written policy are able to respond. The firms that do not have a policy cannot respond in a way that satisfies either stakeholder.

---

What the 16% Who Communicate AI Use Know That Others Don't

The 16% of firms that actively communicate AI use to clients are not primarily large firms with dedicated marketing departments, though some are. The firms that communicate are the ones that made a decision: that disclosure is a position, not a risk.

Here is what proactive communicators have figured out:

Clients who know are more likely to stay. When you tell a client you use AI and explain how it works in their matters, you are demonstrating competence and honesty simultaneously. Clients who have been briefed on your AI use have made an informed choice to stay. That is a stickier relationship than one built on assumption.

Clients who use AI themselves prefer advisors who use it. A growing share of your clients have AI tools in-house. They are not afraid of AI — they are using it. To those clients, silence about your AI use does not signal caution. It signals that you might be behind.

Proactive disclosure converts a liability into a differentiator. Every firm using AI is carrying the disclosure question as a background risk. The firms communicating actively have converted that risk into a position: "We use AI. Here is how. Here is what it means for your work. Here is what we verify." That is a differentiated pitch that 84% of competing firms cannot make.

The Karbon 2026 State of AI in Accounting report adds a data point that should settle any internal debate: 82% of accounting firms that disclosed AI use said it positively impacted client relationships. Transparency did not hurt them. It helped.

The fear that clients will push back on AI disclosure is understandable. But the data does not support it. What clients push back on is being surprised.

---

The Crossing Report Verdict

The gap is 69 points wide. 85% of clients expect disclosure. Only 16% of firms deliver it. That gap does not stay comfortable — it closes reactively or proactively. Firms that close it now own the client conversation. Firms that wait will close it after a client asks why you didn't mention it.

The disclosure gap is not a problem that resolves itself with time. As AI use in professional services deepens — as more work is AI-assisted or AI-generated — the conversation becomes harder to avoid and more loaded when it happens late.

The firms that will win the next three years of client relationships are not necessarily the ones using the most advanced AI. They are the ones whose clients understand and trust how they use it.

For the specific language to close the gap — how to draft a client AI disclosure, what to include in an engagement letter update, and how to sequence the conversation with existing clients — see How to Tell Your Clients You Use AI →

---

The Crossing Report covers AI adoption for professional services firm owners — no vendor spin. Subscribe free