For years, I paid my accountant €300 a year. Not because he was cheap — because I had done half the work myself. Then he retired and handed over his files to another accountant. New tools, automation in place. New invoice: over €1,000 a year. The automation was specifically designed to reduce human work. And the price had more than tripled. This paradox is not a billing anomaly. It is structural — and it is running through every professional services sector.
As I dug deeper into my own experience, I realised there are two very different situations playing out simultaneously, with radically different implications for what clients actually pay.
Two clients. One revolution. Opposite outcomes.
Case 1 — The firm automates. The client never sees the benefit.
This is the most common situation today. The professional adopts automation tools — document capture software, AI-powered transaction categorisation, integrated management platforms. Productivity increases. But the client continues to bring documents as before, unchanged in habits or contribution.
The question is straightforward: where does the productivity gain go?
In accountancy, tools such as Clearfacts, Dext, and Pennylane automate document capture, transaction categorisation, and bank reconciliation — tasks that historically represented a substantial share of billable time for small clients. A October 2025 report by LexisNexis found that 64% of tax professionals now use AI in their daily work, up from 40% just months earlier, with the main reported benefit being the ability to deliver work faster. Yet pricing structures have been slow to follow: billable hours remain dominant, with firms only beginning to experiment with fixed or flat-fee models.
In law, a 2023 Goldman Sachs report estimated that 44% of tasks within the legal profession could be automated by AI — document drafting, legal research, due diligence. Tasks historically billed at premium hourly rates. Early-adopting firms gained in capacity. Their rates did not move.
In medicine, cardiologist and researcher Eric Topol, in his landmark book Deep Medicine (Basic Books, 2019), describes how AI has the potential to transform everything physicians do — from medical imaging to diagnosis and treatment — greatly reducing the cost of care and freeing doctors to focus on genuine human connection with patients. Yet medical consultations continue to be priced on a pre-AI model. The value created by automation flows to software vendors and institutions, not to patients.
Case 2 — The client augments themselves. The firm didn't see it coming.
This is the less documented case — and the more disruptive one. The client adopts AI. They arrive at the professional appointment with a structured file, precise questions, and a refined understanding of their own situation.
This is my case. Today, I use AI tools to prepare my accounting: organising documents, pre-categorising transactions, identifying items that need attention. I arrive with preparatory work that AI helps me produce faster and more accurately than before. What I hand to my accountant is the final review, interpretation of fiscal grey areas, and the professional signature that engages his legal responsibility.
This profile — the AI-augmented client — is still relatively rare. It will become mainstream. The WEF Future of Jobs Report 2025 projects that 39% of key job-market skills will change by 2030, with technological skills led by AI and data literacy projected to grow faster than any other category. This movement does not only concern organisations: it transforms how individuals relate to every professional service they consume.
The information asymmetry that historically underpinned professional authority is narrowing. In law, the client who has used an AI assistant to analyse their contract arrives with questions their solicitor did not anticipate. In medicine, the patient who has cross-referenced their symptoms with clinical data arrives with a hypothesis their doctor must engage with seriously. In accountancy, the client who has pre-structured their data reduces the workload they hand over.
McKinsey's global leader of technology and AI, Kate Smaje, has stated plainly: "This is a moment where many of the fundamentals of the professional services model are coming under challenge."
The uncomfortable question for service providers is this: if a firm benefits both from its own automation tools and from the preparatory work of the AI-augmented client, what exactly is it billing for?
What remains legitimate — and what no longer does
Both cases converge on the same demand for clarity: distinguishing genuine irreplaceable value from pricing inertia.
What remains fully legitimate in professional fees:
- Final review and expert validation
- Legal and professional liability
- Judgement in areas of genuine complexity and uncertainty
- Strategic advisory relationships built over time
What is increasingly difficult to justify:
- Charging the client for the recovery of the firm's technology investment, without any reduction in their burden
- Maintaining uniform pricing regardless of the client's actual contribution
- Presenting automation as an added service rather than a reduction in production cost
2031: A realistic projection
The structural pressure that is building today will accelerate significantly over the next five years. The direction of travel is clear, and the evidence is already visible.
The billable hour is under terminal pressure. McKinsey has disclosed that approximately a quarter of its global fees now come from outcome-based pricing arrangements, with clients increasingly arriving not with a defined scope but with a desired outcome, asking the firm to price against actually delivering it. What is happening at the top of the consulting market will cascade downwards to accounting, law, and medical services.
The automation threshold is accelerating faster than expected. In November 2025, McKinsey Global Institute released a report finding that 57% of US work hours could be automated with technologies that already exist today — nearly double their 2023 estimate of 30% potential by 2030. The report specifically identifies routine accounting processes among the highly automatable cognitive skills facing the greatest near-term disruption.
AI-augmented clients will be the norm, not the exception. As AI and data literacy spread to the general population, the proportion of clients arriving with pre-prepared, AI-assisted work will grow substantially. Firms that price as if every client is a passive recipient of their expertise will face growing resistance.
A new pricing grammar is emerging. Interest in AI-enabled billing models is already growing, with 26% of tax professionals considering subscription-based services and 20% exploring bundled offerings that combine AI outputs with expert review. Within five years, these will likely be the dominant model for small and medium client relationships, replacing legacy structures that no longer reflect how work is actually produced.
The disintermediation risk is real for those who do not adapt. McKinsey has deployed around 12,000 AI agents internally to support consultants and enable leaner project teams, while Accenture has reorganised major divisions specifically to help clients overhaul their operations with AI. Smaller professional service firms that fail to rethink their value proposition face pressure not only from informed clients, but from platform-based alternatives offering expert review at a fraction of traditional costs.
I still work with an accountant. Not for the classification — I do that better myself now, with AI. I pay for what only he can provide: final control, interpretation of fiscal complexity, and the signature that puts his professional reputation on the line.
That is real value. I respect it. But it is not worth €1,000 for a file I have largely prepared myself.
My experience is modest. But it illustrates something significant. The AI transformation of professional services will not spare those who confuse the protection of their pricing model with the delivery of genuine value. The informed client will keep asking the same question: what exactly am I paying for?
AI does not eliminate professional value. It recalibrates it. And that recalibration concerns clients just as much as providers.
This article was written in collaboration with Claude (Anthropic), used as a research, sourcing, and structuring tool. All ideas, experience, and analysis are the author's own.