2026 Predictions: What's Next for AI in Professional Services

December is prediction time. Let me lay out what I think we'll see in professional services AI through 2026. These are educated guesses, not certainties. But they're based on what I'm seeing in the market and where the trends are pointing.

Prediction 1: Agentic Workflows Become Viable, But Not Yet Dominant

In 2026, we'll see more firms successfully deploying agentic AI—systems that handle multi-step processes autonomously with human checkpoints. A research task, analysis, draft generation, internal review loop, refinement. All orchestrated by AI with humans intervening at critical decision points.

But we won't see them replace core consulting work. Agentic AI will be used for support work: research, analysis, draft preparation. The strategic judgment and client interaction still needs humans.

Why? Because agentic systems still make mistakes that are costly in professional services contexts. They're good enough for 80% of routine work. The 20% where they fail still needs human judgment.

Prediction 2: Vertical-Specific AI Tools Will Consolidate

Right now, there are 50 "AI for consulting" tools and 100 "AI for legal" tools. Most are barely differentiated. In 2026, consolidation will accelerate. The winners will be the ones integrated with existing systems (Salesforce, NetSuite, SAP, etc.) or the ones that solve a very specific, painful problem better than anything else.

Expect M&A in this space. Larger software vendors will acquire the most promising vertical AI tools to integrate them into their platforms.

Prediction 3: Regulatory Guidance Will Increase and Get More Prescriptive

We got some guidance in 2025. In 2026, expect more. Different regulators (legal, financial, healthcare) will release specific guidance on AI disclosure, AI governance, and AI liability. Some of it will make adoption harder. But it will also create clarity, which is better than uncertainty.

The firms that built governance into their processes early will have easier compliance. The firms that skipped it will face painful retrofits.

Prediction 4: Model Capability Improvements Will Slow, But Reliability Will Improve

Expect smaller incremental improvements in AI capability, not breakthroughs. What will improve is reliability, long-context handling, and predictability. Models will get better at doing what they do without surprising failures.

This is good for professional services. We don't need smarter models. We need more reliable models.

Prediction 5: Cost Advantage Shifts to Smaller, Specialized Models

Large models are expensive at scale. As firms scale AI usage, costs become material. Expect investment in smaller models (Llama, Mixtral, etc.) fine-tuned for specific tasks. They'll cost 10–20% of large model costs and be good enough for 60–70% of use cases.

This is a year where cost becomes a real competitive factor. Firms that figured out the right model-to-task matching will have better margins.

Prediction 6: AI Will Start Affecting Professional Services Pricing

For three years, we've been saying "AI won't change pricing." I think that changes in 2026. We'll start seeing:

• Firms that confidently say "AI makes us faster" and charge premium rates for faster delivery.
• Firms that use AI to commoditize low-value work and compete on price.
• Clients demanding discounts because they know AI is doing part of the work.

Pricing will be a bigger strategic question in 2026 than it was in 2025.

Prediction 7: The "AI Maturity Gap" Will Become More Obvious

The firms that invested in AI in 2023 and 2024 are noticeably more productive and profitable than firms that started in 2025. In 2026, that gap will become increasingly visible and increasingly hard to close.

This will accelerate adoption. Firms that see competitors winning will move faster, not slower.

Prediction 8: Data Governance Will Become a Major Liability

We're already seeing initial questions about AI trained on client data, AI that sees confidential information, AI that shares data with vendors. In 2026, this becomes a major conversation. Some high-profile cases of data misuse or exposure will happen. Firms that don't have clear data governance will face real liability.

Expect this to be a major insurance and compliance issue by mid-2026.

Prediction 9: Smaller Firms Will Start Catching Up

2025 advantages went to large, well-resourced firms that could hire AI talent and invest in infrastructure. In 2026, as tools get better and cheaper, smaller firms will start using AI effectively. The advantage won't disappear, but it will narrow.

Prediction 10: The Real Value Will Be in Transformation, Not Tools

Firms that treat AI as a tool will see tool-like impact. Firms that treat AI as a lever for fundamental business model change will see bigger impact. In 2026, the distinction will become clearer.

The winners won't be the ones with the best AI. They'll be the ones that rethought their business model around AI's capabilities.

What I'm Least Confident About

I'm genuinely uncertain about regulatory direction. It could be light-handed, or it could be heavy. I'm uncertain about how fast model improvement will slow. I'm uncertain about whether the productivity gains from AI will stick or evaporate as people adjust expectations.

But I'm pretty confident about the direction: AI will matter more, governance will matter more, and competitive pressure will increase.

The Bottom Line

2026 is the year where AI goes from "the new thing" to "table stakes." Firms that haven't started will have to accelerate. Firms that have started will have to deepen. Firms that treat AI as infrastructure will thrive. Firms that treat it as a novelty will struggle.