Insights
What M&A advisors should automate first with AI in 2026
Mitchel Campbell, Co-founder and CEO
· 6 min read
The short answer
Automate the work that takes the most hours and the least judgement. For almost every M&A advisory firm that is buyer and target list research, not CIM drafting – which is where most firms start, and which is the part of the process where the judgement actually lives.
The ranking below is not about what AI can do. It is about where a 2 to 20 person advisory firm's hours actually go, and which of those hours a machine can take back without taking anything you are paid for.
The test
Three questions, in this order.
How many hours a month does it take? If you cannot answer, that is your first finding and the first week of work. Most firms are wrong about this by a factor of two, usually in the direction of underestimating research.
How repeatable is it? Same inputs, same steps, same shape of output every time. A process that is different on every deal is a process you are still designing.
How much judgement is in it? Not how much skill – how much judgement. Formatting a bank book takes skill. Deciding what the business is worth takes judgement. Only one of those is safe to hand over.
High hours, high repeatability, low judgement is the quadrant. Everything else waits.
Where the hours go
Rough shape for a five-person advisory firm running six to eight live mandates. Your numbers will differ; the ordering usually does not.
| Activity | Share of analyst time | Repeatable | Judgement |
|---|---|---|---|
| Buyer and target research, list building | 30-40% | High | Low |
| Outreach, follow-up, scheduling | 15-25% | High | Low |
| Diligence request and Q&A handling | 10-15% | Medium | Medium |
| CIM and teaser drafting | 10-20% | Medium | High |
| Pipeline and CRM upkeep | 5-10% | High | Low |
| Valuation and analysis | 10-15% | Low | High |
The top two rows are over half the analyst time in most firms and almost none of the fee logic. That is the answer to the question.
The five candidates, ranked
1. Buyer and target list research. The highest-hour, lowest-judgement work in the building. An agent reads the mandate, screens a universe against the criteria, pulls the firmographics, and produces a list with the reasoning attached for each name. The human decides who is actually on it. The list is usually better too, because the machine does not get bored at name 400.
Start here unless you have a specific reason not to.
2. Outreach and follow-up. Second because the hours are real and the work is genuinely repetitive, but it carries a reputational edge that list building does not. Personalised outreach written in the firm's voice, sequenced, with replies handled and meetings booked. The constraint is the approval gate: anything that reaches an owner, buyer or seller should be seen by a person first. Firms that skip that gate save an hour and spend a year regretting it.
3. Pipeline and CRM upkeep. Low glamour, high compliance. An agent that reads your email, calls and calendar and keeps deal records current removes a task nobody does well and everyone resents. Small hours, near-zero risk, and it makes everything downstream more reliable.
4. Diligence request and Q&A triage. Sorting incoming questions, routing them, finding the document that answers them, drafting the response for review. Real time back, but it sits closer to the client, so it needs the approval gate and a tighter permission model.
5. CIM and teaser first drafts. Last, despite being first on most firms' lists. A model can assemble a structured first draft from the data pack and it will look impressive. The problem is that the value of a CIM is in the positioning – what to lead with, what to disclose when, how to frame a bad year – and that is the judgement you are paid for. Automate the assembly, never the argument.
What not to automate, in 2026 or later
- The valuation conclusion
- The negotiation
- Anything that goes to a client or LP without a person approving it
- The relationship
A provider who does not volunteer a list like this has not thought about yours.
Find your own number in a week
You do not need a consultant to do this part.
- Pick the two activities from the table above that feel heaviest in your firm
- Have every analyst log hours against them for one week – a tally in a spreadsheet, nothing fancier
- Multiply by four and compare to the fee that work supports
- Pick whichever of the two has more hours and less judgement
- Scope a first build on that one process only, with the baseline measured before anything is built
The last step is the one firms skip, and it is the one that makes the result arguable six months later. If you cannot say what the number was before, you cannot say what changed.
Frequently asked questions
What should an M&A advisory firm automate first with AI? Buyer and target list research. It is typically 30 to 40 percent of analyst time, highly repeatable, and carries almost no judgement – the human still decides who is on the final list.
Why not start with CIM drafting? Because the value of a CIM is the positioning, not the assembly. A model can produce a structured first draft, but what to lead with and how to frame a weak year is the judgement clients pay for. Automate the assembly, not the argument.
How long does the first workflow take to get live? Weeks rather than months when it is scoped to one process. A build that is quoted before anyone has measured the baseline is a guess, however long it is quoted for.
Do we need to hire someone technical first? No, and firms that do often build the wrong thing well. What you need is a measured baseline and one process worth fixing.
How do we know it worked? Agree the number in advance – hours returned, turnaround time, coverage or completeness – measured against where you started. A demo is not a measurement.
How we do it, for completeness
Analyst3 starts with a strategy and workflow audit that maps where a firm's hours actually go and ranks what to automate first, quantified in hours and dollars. The highest-value workflow goes first, run by hand on live deals until the team signs off on the output, then wired into the systems the firm already uses. Clients own the code, the data and the workflows, running in their own accounts.
We rank buyer list research first because that is what the audits keep finding, not because it is what we like building.
If this helps you pick the right first workflow and you build it with someone else, the guide did its job.