A junior analyst at an M&A boutique spends most of their first two years assembling Information Memorandums. The job is repetitive, document-heavy, and unforgiving — a misstated EBITDA in a sell-side IM is a credibility hit that travels.
Most partners we talk to estimate 20 to 80 hours of junior time per IM. The range is wide because deals vary, but the work inside that range is remarkably consistent. Breaking it down is the first step to understanding why AI changes the economics — and what specifically gets faster.
Where the 80 hours actually go
We’ve mapped this with several boutiques. The numbers below are mid-range for a typical mid-market sell-side mandate; the shape of the breakdown holds across firms.
| Activity | Hours | Share |
|---|---|---|
| Reading & indexing the data room | 8–14 | ~18% |
| Extracting financials into the firm’s working model | 10–16 | ~22% |
| Cross-validating figures across sources | 6–10 | ~14% |
| Writing first draft of qualitative sections | 12–20 | ~28% |
| Formatting, branding, version control | 4–8 | ~10% |
| Partner review iterations | 4–8 | ~8% |
The first four categories — roughly 80% of the time — are structural work. They’re the same on every deal: read, extract, validate, write a first pass. The remaining 20% is partner judgment and editorial polish.
The structural work is exactly what AI does well now. The judgment work is not.
What “under one hour” actually means
When we say NaS_OS produces an Information Memorandum in under an hour, we’re talking about a specific thing: the structural 80% — read, extract, validate, draft — done in roughly 40 minutes of compute time, plus 20 minutes of human review.
What still requires a human:
- Strategic positioning of the target (“how do we tell this story to buyers?”)
- Sensitivity to confidential context the data room doesn’t capture
- Final tone calibration for the specific buyer universe
- The partner’s signature on the final document
What doesn’t:
- Every figure on every page of every PDF, extracted and cross-checked
- A clean working table of historical financials, with footnotes traced to source pages
- A first-draft Business Overview, Market Section, and Financial Analysis written in the firm’s house style
- Every red flag, gap, and inconsistency the data room reveals — surfaced, not buried
A junior who used to do this in a week now reviews it in an afternoon.
Why this matters for a boutique’s economics
For a 6-person boutique running 8 mandates a year, the math is straightforward. Assume the structural work is 60 hours per IM at €120/hour fully-loaded analyst cost — that’s €3,600 in time per document, or roughly €28,000 a year tied up in structural IM production.
That’s not the interesting number, though. The interesting number is what those 480 hours could do instead:
- More deal origination calls per partner
- Faster turnaround on inbound mandate requests (the firm that responds in 48 hours vs. 2 weeks wins disproportionately often)
- More buyer outreach per active mandate
- Better-targeted teasers, run in parallel
Boutiques don’t lose mandates because they’re bad at producing IMs. They lose mandates because they can’t pitch fast enough, can’t run enough parallel buyer outreach, and can’t take on a new mandate when their current pipeline is already capacity-bound. Freeing up 480 analyst-hours a year directly attacks that capacity ceiling.
The “but our work is special” objection
We hear this a lot. It’s a reasonable concern — every boutique believes (correctly) that their IMs reflect a specific voice, sector specialization, and standard of analysis.
It would be wrong if AI tried to produce a one-size-fits-all output. It doesn’t. The firm’s frameworks, tone, sector conventions, and house style get encoded into the system once, then applied to every subsequent deal. The output reads like your firm wrote it, not like a generic AI tool produced it.
A horizontal AI tool — built for hedge fund analysts or in-house legal teams — can’t do this, because it doesn’t model the specific conventions of sell-side M&A in the first place. That’s why specialized matters. A junior analyst at a boutique spends six months learning how their specific firm writes an IM; a specialized M&A AI starts from that baseline.
What happens to junior analysts
The honest answer: the role changes shape. The 60 hours per IM that used to be a junior’s primary deliverable becomes 20 minutes of review and 2 hours of judgment work — the interesting part of the job they would have done eventually anyway. The grunt work doesn’t transfer to someone else; it stops being grunt work.
The boutiques moving fastest on this are the ones whose juniors want the change. They didn’t sign up for an analyst program to format Excel tables for two years.
The cost of waiting
A specific way to think about it: every quarter a boutique waits to adopt AI for structural IM work is roughly 120 partner-equivalent hours of lost capacity — capacity that goes to origination, buyer outreach, and taking on the marginal mandate.
In a competitive market for sell-side mandates, that compounds. The boutique whose 6 partners each have 50% more capacity to pitch and run deals isn’t 50% bigger — they’re winning the bake-offs the slower firm doesn’t even get invited to.
NaS_OS is the AI deal analyst built specifically for this. If you run an M&A boutique and want to see what under-one-hour Information Memorandums actually look like on your own documents, apply for early access.