A consulting firm's actual product is judgment: the recommendation, the point of view, the answer to a question the client could not answer alone. Most of the week does not go to that. It goes to the research that has to happen before the judgment can, the deck that presents it, and the update that keeps a client informed between milestones. Rebuild the same competitor scan for a similar client, assemble the briefing the night before a call, chase down who agreed to what after a meeting nobody wrote up properly. None of it is the work clients are paying for, and all of it eats the hours a consultant actually has for thinking.
At Sanaf, an AI employee takes on exactly that supporting work, through four real jobs on the roster, while the recommendation stays with a person every time.
The problem is where the hours actually go
Ask a consultant what they did this week and the honest answer is rarely "advised clients." It is research, slide-building, scheduling and chasing action items, with the actual advising squeezed into whatever time is left. That ratio is backwards for a business that bills for judgment, and it is not a staffing problem so much as a task-allocation problem: the wrong hours are going to the wrong person.
What an AI employee actually does for a consulting firm
- The research analyst job researches companies, competitors, markets and suppliers, checks facts and sources against each other, writes short briefs with every source cited, and builds the comparison table or shortlist a first draft needs, instead of a consultant starting from a blank search bar.
- The executive assistant job manages the calendar around client calls, prepares the briefing before each one, summarizes what was discussed and agreed, and tracks the commitments so a promise made on a call does not get forgotten by the next one.
- The business analyst job monitors the metrics a client engagement is tracking, spots the trend or the anomaly, and builds the report from the real numbers instead of a person rebuilding the same spreadsheet every milestone.
- The administrative assistant job processes the notes into documents, keeps the engagement's files organized, and generates the recurring status report so the update goes out on schedule instead of whenever someone has an hour free.
That is the research pulled together, the briefing ready before the call, the numbers current, and the update sent, with the recommendation itself left for the consultant it is actually paid for.
Where it stops
The research job separates what it found from what it thinks, and says plainly when sources disagree or run out. It does not make the decision, and nothing it finds about a person is used beyond what was asked for. The business analyst lays out plainly what the numbers show and where it is unsure; it never dresses up a shaky number as a certainty, and every recommendation stays a person's to accept or ignore. Nothing here writes the point of view a client is paying for. That is the whole point: the hours that go back to the consultant are the hours spent on exactly the judgment call the firm exists to make.
How to tell if this is worth it before your next hire
- A junior consultant's week is mostly research and slide assembly, and the actual client-facing time is what is left over.
- The same competitor scan or market brief gets rebuilt from scratch for every similar engagement, because nothing from the last one got reused.
- Client updates slip because writing the status report always waits for a slower week.
- Commitments made on a call get forgotten because nobody wrote them down in a place anyone checks again.
If none of that sounds familiar, a well-run research library and a disciplined status cadence are probably still enough. This starts paying for itself once the research and reporting load outgrows what a junior consultant can absorb without cutting into billable client time.
Frequently asked questions
What does an AI employee actually do for a consulting firm? Researches companies, markets and competitors with sources cited, prepares the briefing before client calls, tracks commitments made on those calls, builds reports from real engagement metrics, and sends the recurring status update on schedule.
Does it write the actual recommendation or point of view? No. It separates what it found from what it thinks and never makes the call. The recommendation a client is paying for stays with a consultant every time.
Can it be wrong about a fact and I would not know? It cites every source and says plainly when sources disagree or the research runs out, rather than presenting a guess as settled.
Will it replace a junior consultant? It removes the research and slide-assembly hours a junior consultant's week is usually made of, so that time goes to client-facing work and the judgment calls that build a career, not to rebuilding a competitor scan.
What does this cost? Start free with Sanaf, no card, and pay for the research and reports as they happen, not a seat per consultant.
The bottom line
A consulting firm sells judgment, and judgment needs the research and the briefing to be ready before it can happen. An AI employee closes that gap, cites its sources, says when it is unsure, and leaves the actual recommendation exactly where a client is paying for it to come from: a person.
See the research analyst job on the full roster, how people and operations teams use the same reporting model, or get in touch to see it against your own engagement list.