Professional services firms are entering the ROI phase of AI

Professional services firms are moving past the novelty phase of AI.

That is the good news.

The harder news is that usage alone is no longer very interesting.

In 2026, the real question is whether firms can turn AI activity into measurable operating leverage.

What the market data says

Thomson Reuters' 2026 AI in Professional Services Report shows how quickly adoption is moving:

• 40% of professionals say their organizations now use generative AI, up from 22% the year before
• more than 80% of current users engage with it weekly
• only 18% say their organizations track ROI
• only 15% say their organizations currently use agentic AI, though another 53% are planning or considering it

That is a very specific signal.

The industry has moved from curiosity to routine use faster than it has moved to economic discipline.

Why that gap matters

Professional services firms do not win by adopting technology in the abstract.

They win by improving some mix of:

• realization
• turnaround time
• client experience
• margin
• staff leverage
• quality and consistency

If AI use does not connect to one of those, it will eventually look like overhead with better branding.

This is already visible in buyer expectations.

The Thomson Reuters report says two-thirds of corporate respondents want their outside firms to use AI, but fewer than 20% formally require it.

That creates an awkward middle state:

clients expect more speed and efficiency, but many firms still cannot explain clearly where AI changes delivery economics or service quality.

Where the operational opportunity actually is

The easiest trap is to think the biggest win is generic drafting.

Drafting matters. But many firms lose more time in the surrounding workflow:

• intake cleanup
• matter opening
• document collection
• checklist routing
• evidence or support packet assembly
• internal review handoffs
• status follow-up across email and systems

Those are the tasks that quietly consume billable time, delay response cycles, and create hidden margin pressure.

That is why the best AI use cases in professional services are often operational before they are creative.

What stronger implementations look like

The firms that get value first will usually define a narrower target:

• client onboarding package completion
• tax document collection and normalization
• audit support request follow-up
• claims, compliance, or risk evidence gathering
• engagement status routing across partners, managers, and ops teams

These workflows are good candidates because they tend to have:

• repetitive coordination
• structured definitions of done
• obvious exception paths
• clear economic value when cycle time drops

That is how firms move from "people use AI" to "the firm runs better."

Why ROI discipline is the next separator

The Thomson Reuters numbers are the warning sign.

If adoption keeps climbing while ROI measurement stays weak, firms will eventually face the same question from leadership and clients:

What changed?

Not:

What tools do we have?

But:

• Did turnaround time improve?
• Did utilization or realization improve?
• Did non-billable admin fall?
• Did service delivery become more consistent?
• Did client communication get faster without adding headcount?

That is the phase professional services is entering now.

It is the ROI phase.

What firms should avoid

Two mistakes are common.

First, firms buy AI access without redesigning the workflow around it.

Second, firms try to measure value too late, after usage has already spread.

The better approach is:

1. Pick one expensive workflow.
2. Define the current manual cost and turnaround time.
3. Automate the coordination around the work, not just the draft.
4. Track the resulting throughput and margin impact.

That creates a cleaner operating case than a broad "AI transformation" program.

The practical takeaway

Professional services firms are not being judged on whether they have heard of AI anymore.

They are starting to be judged on whether they can use it to improve delivery without making quality, trust, or economics fuzzier.

That means the important shift is not from no AI to some AI.

It is from unmeasured AI activity to measured workflow improvement.

That is where the next real advantage will come from.

Sources

• Thomson Reuters, "2026 AI in Professional Services Report"

If your firm is already using AI but still losing time to intake, follow-up, and handoff work, our professional services page shows the workflow angle. If you want a simple cost baseline first, run the calculator.