Professional services firms love to talk about AI. Client proposals mention it constantly. Partner interviews emphasise AI investments. But behind the marketing, what’s actually happening?

I’ve been tracking AI adoption across law firms, accounting practices, and consultancies throughout 2025. The reality is more nuanced than the messaging suggests.

The Law Firm Landscape

What’s Actually Deployed

Document review and discovery: AI-assisted document review for litigation is now standard at large firms. eDiscovery platforms use AI to prioritise documents for human review. This works and delivers real efficiency gains.

Contract analysis: AI tools that extract key terms, identify risks, and compare documents against standards. Adoption is growing, particularly for M&A due diligence and high-volume contract review.

Legal research: AI-powered research tools that find relevant cases and synthesise precedents. Most firms have access; usage varies significantly by practice area and partner.

Document generation: Drafting assistance for routine documents – contracts, letters, filings. Useful for first drafts that lawyers refine.

What’s Still Hype

Autonomous legal work: AI that handles legal matters end-to-end without lawyer involvement. Despite some vendor claims, this isn’t happening at serious firms.

Strategy and judgment: AI advising on litigation strategy, deal structure, or regulatory approach. The judgment calls remain human.

Client-facing AI: Chatbots handling client enquiries directly. Rare, and usually limited to simple FAQ-type questions.

The Partnership Challenge

Law firm AI adoption faces a structural challenge: partner economics. Partners bill by the hour. Efficiency means fewer billable hours. Where’s the incentive?

Firms that have addressed this are moving toward value-based pricing, where efficiency benefits clients and firm margins simultaneously. But this transition is slow.

The Accounting Landscape

What’s Actually Deployed

Audit sampling and testing: AI identifying unusual transactions and prioritising audit attention. The Big 4 (including Deloitte and PwC) have invested heavily here, and it’s in production use.

Tax research: AI-assisted research into tax implications and precedents. Similar to legal research – tools exist and are used.

Anomaly detection: Identifying potential fraud, errors, or compliance issues in financial data. This is where AI adds genuine value – pattern recognition at scale.

Document extraction: Pulling data from financial documents (invoices, statements, tax forms) into structured formats. Works well for standardised documents.

What’s Still Hype

Automated audit completion: The audit opinion still requires professional judgment. AI supports the work; it doesn’t replace the opinion.

Complex tax planning: Creative tax structuring remains human expertise. AI helps with research, not strategy.

Full automation of compliance: Regulatory compliance still requires human review and sign-off.

Regulatory Constraints

Accounting AI adoption is constrained by regulatory requirements. Professional standards require human judgment for key decisions. Auditors must demonstrate they’ve exercised professional scepticism. AI can inform these decisions but not make them.

This regulatory clarity actually helps – it defines clear boundaries for AI use.

The Consulting Landscape

What’s Actually Deployed

Research synthesis: AI summarising market research, competitive intelligence, and client documents. Every large consultancy uses this.

Presentation creation: First-draft slides generated from outlines or prior decks. Useful for routine deliverables, less so for strategic work.

Transcription and analysis: Meeting transcripts, interview summaries, and qualitative data analysis.

Knowledge management: AI-powered search across firm knowledge bases. Variable quality, but improving.

What’s Still Hype

Strategy generation: AI developing business strategies or transformation roadmaps. The strategic thinking remains consultant-driven.

Client-ready deliverables: AI producing final client deliverables without significant human refinement. Quality isn’t there.

Senior partner augmentation: AI enabling senior partners to be more productive. Adoption at senior levels remains low.

The Leverage Model Problem

Consulting’s traditional leverage model – seniors sell, juniors deliver – creates AI friction. If AI replaces junior work, who staffs projects? How do juniors develop?

Firms are grappling with this. Some are reducing junior hiring. Others are shifting junior work toward tasks AI can’t do. The model is evolving, uncomfortably.

Cross-Cutting Themes

Several patterns appear across professional services:

Support Functions, Not Core Work

AI excels at supporting work: research, document processing, first drafts. It struggles with the core professional judgment that clients actually pay for.

This is appropriate. The value of professional services is professional judgment, not document production. AI that helps professionals spend more time on judgment and less on mechanics is valuable.

Adoption Variation by Role

Usage varies dramatically:

• Administrative staff: High adoption for relevant tasks
• Junior professionals: Growing adoption, variable quality
• Senior professionals: Low adoption, limited interest
• Partners: Very low adoption

This creates an awkward dynamic where AI tools are purchased by leadership and barely used by leadership.

Client Expectations Outpacing Reality

Clients increasingly expect AI-enabled services. But expectations are often based on vendor marketing rather than realistic capabilities.

This creates pressure to claim AI capability that exceeds actual deployment. The gap between marketing and reality is uncomfortable.

Training Investment Required

Professional staff need training to use AI effectively. Firms that invest in prompting skills, workflow integration, and best practices see better adoption.

Firms that just deploy tools without training see expensive shelfware.

The Business Model Question

The deeper question: what is professional services AI for?

Efficiency story: AI makes existing work faster and cheaper. This helps margins but threatens revenue.

Quality story: AI enables better work – more thorough research, fewer errors, more options considered. This justifies fees.

Capacity story: AI enables handling more work with the same resources. This grows revenue without proportional cost growth.

Firms that have answers to these questions are deploying AI strategically. Firms without answers are deploying AI reactively.

What Clients Should Know

If you’re a client of professional services firms:

Ask what AI they actually use, not what they can use. Firms often describe capability rather than practice.

Understand how AI affects pricing. If AI makes work faster, are you sharing the efficiency gains?

Expect AI-assisted, not AI-replaced. Professional judgment still matters. AI supports humans; it doesn’t replace them.

Assess AI claims critically. Firms have incentives to exaggerate AI sophistication. Ask specific questions.

Final Thought

Professional services AI in 2025 is useful for support work and limited for core professional judgment. This is appropriate – the judgment is what you’re paying for.

The firms succeeding with AI are those with clear business model answers, investment in training, and realistic expectations. The firms struggling have deployed tools without strategy.

The hype exceeds the reality, but the reality isn’t bad. AI in professional services is doing exactly what it should – supporting humans to do better work.