Sector

AI leadership for professional services firms

Last updated

Justas Butkus is an AI expert and fractional AI officer based in Vilnius, Lithuania, working with mid-market companies and scale-ups across the UK, EU and US at two to four days a month. He builds and operates the production AI systems he advises on, rather than only advising on them.

Short answer

Firms that sell expertise by the hour face a specific tension: the most automatable work is also the work they bill for. The value usually sits in the unbilled surrounding activity rather than in the advice itself.

Where does AI actually pay for itself in professional services?

The processes worth automating in any sector share three properties: they repeat constantly, they consume qualified people, and nobody enjoys them. In this sector that usually means:

  • Intake and scoping – repetitive qualification before any billable work starts.
  • Document assembly and review – templated output that still takes qualified time.
  • Internal knowledge retrieval – the firm already knows the answer; finding it is the cost.
  • Client reporting – assembled manually, every cycle, by people who bill.
  • Proposal production – largely reassembly of previous proposals.

None of these are exciting. That is the point – the value comes from volume, not sophistication, and the exciting use case is usually the one that fails.

What makes this sector different

  • Client confidentiality constrains architecture, and it is the first question a client will ask. Where data goes matters more than which model is used.
  • Professional obligations do not transfer to a tool. Whoever signs the advice remains responsible for it, which makes human review a design requirement rather than a preference.
  • The billing model complicates the business case. Automating billable work reduces revenue unless capacity is redeployed, so the honest first candidates are the unbilled processes.

How an engagement here typically runs

  1. Inventory what is already runningIncluding the unapproved tools. In regulated environments this is usually the most uncomfortable and most valuable step.
  2. Map the repeating processes against costFrequency multiplied by fully loaded cost per occurrence. The ranking usually surprises people.
  3. Establish governance proportionate to the riskWho approves what, what happens when a system is wrong, and how a decision is reconstructed months later.
  4. Ship the unglamorous one firstProve the pipeline end to end before attempting the ambitious thing.

Frequently asked questions

Does AI cannibalise billable hours in professional services?

It can, which is why the strongest first candidates are unbilled processes: intake, internal knowledge retrieval, reporting and proposal assembly. Automating billable work only pays if the freed capacity is redeployed.

How do we handle client confidentiality?

It is an architecture question, decided before anything is built: where data is processed, what is retained, and whether it can be demonstrated. Clients ask this first and a general assurance does not survive the question.

What is the most common mistake here?

Starting with the advice itself because it is the most visible use case. The economics almost always favour the surrounding unbilled work.

If this describes your operation

The useful first conversation is about one specific process rather than about AI in general.

Get in touch