Putting AI to Work in Your Firm: when to start, what to weigh, and the mistakes to avoid
A practical guide for M&A advisory firms, investment banks and business brokers. Where AI is paying off in deal work today, how to think about timing when the technology changes every few months, what to settle before you build, and what we have learned building AI workflows for M&A teams.
Mitchel Campbell, Analyst3 · Published September 2026 · Updated October 2026 · 5 min read
Key takeaways
- 1
Start with the work, not the tool.
Map where your team's hours actually go before choosing anything. The best first project is usually repetitive and unglamorous – screening, intake, follow-ups – not the showcase demo.
- 2
Invest in the layer that lasts.
Models and apps will keep changing. Your mapped workflows, clean deal data and review steps will not, and they are what make each new model useful to your firm.
- 3
Keep advisers in the loop.
Agents research, draft, screen and prepare. People review, decide and own the relationship. Work the team can see and check is work the team will use.
- 4
Choose a partner who evolves with you.
The value is not a single build. It is keeping your workflows current as AI improves, without your deal team having to follow the market themselves.
Where is AI paying off in M&A advisory firms today?
The gains are in the repetitive work around a deal – research, drafting, screening and record-keeping – across four areas of the firm.
Origination and BD
- Target and buyer lists:screen a market, enrich contacts and rank by fit.
- Outreach research:a researched reason to contact each owner or buyer.
- Referral partner coverage:know which accountants and lawyers have a reason to hear from you.
Sell-side execution
- CIM and teaser first drafts:built from the data room, edited by the deal team.
- Buyer universe screening:every plausible buyer group, with the rationale for each.
- Seller education material:market-specific guides that position the firm early.
Buyer process
- Buyer responses:NDA requests and questions answered with deal context.
- Buyer intake and follow-up:NDAs logged, duplicates caught, follow-ups drafted on time.
- Process tracking:who has what, who has gone quiet, who needs a call today.
Firm operations
- CRM upkeep:calls, emails and meetings logged without anyone typing them in.
- Inbox and call triage:what needs a reply, drafted with the deal context attached.
- Deal knowledge:past deals, buyers and comps searchable in plain English.
When should an M&A firm start using AI?
When you have repeatable work, usable data and someone to own it – not when the tools stop changing. The tools will keep changing; the groundwork carries forward.
Tools expire. The groundwork doesn't.
The most common reason firms hold back is reasonable: whatever they buy now could be out of date in six months. It helps to separate what changes from what carries forward.
"The software changes every few months. If we invest now, we will be paying for something that is obsolete by next year."
Half right. The model and the app you use today may be replaced. What you build around them will not.
Changes every few months
- The model – a better, cheaper one arrives regularly
- The app or tool – features, vendors and pricing shift
- Prompts tuned to one model – need re-work as models improve
- What is possible – tasks that failed last year now work
Carries forward
- Mapped workflows – agreed steps for how your firm runs a deal
- Clean deal and CRM data – buyers, contacts and history you can trust
- Review steps and controls – who checks what before it leaves the firm
- A team used to working with agents – the habit is the hardest part to build
Firms that start now capture value on live mandates today, and every model upgrade lands on groundwork they already have.
Waiting is a reasonable choice. It is not a free one.
Firms across the market are already putting AI to work on live deals. There are three broad ways to respond, and each has a cost.
Wait and watch
What you get
- No spend and no disruption
- A clearer market to choose from later
What it costs
- Hours keep going to repetitive work
- Data and process stay unstructured, so you start from zero later
- Peers build capacity and experience in the meantime
Leave it to individuals
What you get
- Quick wins for the people who try it
- Low cost, no project to run
What it costs
- Gains stay with individuals and do not add up across the firm
- Client data goes into tools nobody reviewed
- Nothing connects to your CRM or deal process
Build the groundwork now
What you get
- Hours back on live mandates now
- Workflows, data and controls that carry forward
- Model upgrades build on what you have
What it costs
- Internal time to map the work and review output
- A budget line
- Success depends on choosing the right partner
Doing nothing for now is a legitimate decision. It just helps to make it knowingly.
Readiness is less about size than about having repeatable work, usable data and someone to own it.
See where your firm stands
The checklist tells you how ready you are. The readiness check looks at where your team's time actually goes – about three minutes – and gives you your position, where we'd look first and a tailored first 90 days.
Get the full guide
The full guide goes further: the ten-question readiness checklist, six things to settle before you build, the mistakes we see most often, and a sensible first 90 days.