The applications are obvious; the economics are not
Research synthesis, first drafts, proposal assembly and knowledge retrieval are all things AI does well, and professional firms do a great deal of all four.
The harder question is commercial. A firm billing by the hour that halves the hours on a deliverable has reduced its own revenue, which is why adoption in these firms is more complicated than the technology suggests.
Knowledge retrieval first
Every firm has previous work that is relevant to the current engagement and unfindable. Retrieval over past deliverables, research and precedents is the highest-value and lowest-risk starting point.
The value is not saving typing. It is that a junior consultant finds the analysis a colleague did two years ago, instead of redoing it. That is margin, and it is invisible in any time sheet.
Research and synthesis
Gathering, summarising and reconciling material from many sources, with citations, for a person to verify and build on. The output must always be checkable back to source, because the firm's credibility depends on the facts being right.
Never let synthesised research go to a client without verification. Invented citations are the most damaging failure mode in this sector.
Proposal production
- Assembly from previous proposals and approved boilerplate
- Tailoring to the client's language and sector
- Consistency of pricing and terms with firm standards
- A record of what was proposed, for later comparison against what was delivered
Proposals are high-volume, deadline-driven and largely assembly, which makes them one of the better returns available.
The pricing conversation
If deliverables take less time, hourly billing shrinks. Firms respond in three ways: absorb it and compete on price, shift to fixed or value pricing, or use the capacity to take on more work.
The third is the easiest transition and depends on demand existing. The second is where the sector is heading and requires a different conversation with clients.
What clients will start asking
Clients increasingly ask whether AI was used and whether their data went into it. Have an answer: what tools, what data, what human review, what the contract with the provider says.
Being able to answer clearly is becoming a competitive advantage rather than a compliance chore.