EXIT83 · AI Due Diligence
Find the insights
your analysts overlook.
The signal that changes the IC call is rarely on the deck. It’s three layers down: a founder’s prior interview, an adjacent acquirer’s hiring, a competitor’s quiet cap-table move, a watch-out buried in slide 47 of a 2022 talk. Your analysts run out of hours before they get there. AI doesn’t.
- The signal that flips a pass into a lead.
- The watch-out that saves you from an over-priced round.
- The pattern that grades against your win log, not the room’s.
/ what you get back
A score, a fit verdict, and three scenarios.
Every memo answers the three questions an IC actually asks, then backs them up with a four-pillar health report your partners can defend.
01 · The score
A single deal score, defensible to the IC.
One number out of 100, scored against your codified rubric. Same dimensions, same weights, every deal. Compare across partners, vintages, and verticals without recalibrating the read.
example82/100
02 · The fit
A clear yes, no, or pass-for-now against your thesis.
Not a vibe. We codify your investment thesis once, then test every deal against it line by line. Off-thesis deals get killed fast. On-thesis deals come back with the case already made.
verdictStrong fit
03 · The scenarios
Bull, base, and bear, with probabilities attached.
Three forecasts on a 36-month horizon with the assumptions made explicit. Each scenario names what has to break, what has to land, and the MOIC range that follows.
base case2.8× MOIC
04 · The health report
Four pillars. Every deal.
The verdict above is built from a structured health check across financial, cultural, technical, and competitive risk. Each pillar comes with its own score, the questions we asked to get there, and the watch-outs the partner should bring to founder follow-ups.
Financial
9 questions
Are the unit economics real?
CAC payback, NRR, GRR, gross margin, burn multiple, cohort behavior, and pricing power. Sourced from the data room and stress-tested against your historical comps.
Cultural
7 questions
Will this team make it?
Founder-market fit, prior outcomes, co-founder dynamics, equity split, hiring trajectory, Glassdoor signal, and pattern match against the operators who’ve shipped this category before.
Technical
8 questions
Is the moat actually built?
Architecture sniff-test, AI exposure, dependency risk, data network effects, integration depth, and the questions a CTO would ask before signing the term sheet.
Competitive
6 questions
Who eats this lunch?
Incumbent threat, adjacent entrants, switching costs, distribution chokepoints, and a frank read on whether the category is winner-take-most or fragmented for a decade.
/ signals in the noise
Six signals the room can’t see.
Every deal runs against six things the consensus can’t reach. The herd grades the deck and the room it walked through. We grade the signals underneath. Two reads from the same week below: the deal six firms were chasing, and the deal eleven firms passed on.
/ the six
Where the gap actually comes from.
01Founder pattern match. Against your historical winners and losers, not just deals that got funded.
02Off-radar talent moves. LinkedIn, GitHub, paper authorship, and team-departure patterns at adjacent acquirers.
03Geography & category bias. Where consensus pricing is artificially suppressed by where a deal was built.
04Quiet-money signal. Adjacent investors and operator angels who are already in but haven’t tweeted.
05Your firm’s loss log. Watchlist hits against the deals you wish you hadn’t done, encoded so you don’t repeat them.
06Cohort-vintage drift. When metrics that looked good in 2022 grade differently against 2026 efficiency screens.
Two reads, same week
Crowded trade
Helio Logistics
Series B · $24M @ $180M post · 6 firms competing
Market consensus
9.1/10
tracked across 14 funds, last 30 days · “hot, oversubscribed”
Your thesis says
64/100
unit economics fail your post-2023 efficiency screen
The herd is paying for growth. Your rubric grades the unit economics the same as Mythic 2022, the comp that ate a 60% mark down at the next round. CAC payback dragged from 11mo to 14mo and no one’s pricing that in.
RecommendationPass, or pre-empt 30% below the round.
Diamond in the rough
Vault Optics
Series A · $8M @ $32M post · passed by 11 firms
Market consensus
3.4/10
tracked across 11 firms · flagged on geography & category
Your thesis says
81/100
3-of-3 hit rate on the founder profile in your win log
Rejected on geography. Your three best returns were Bucharest-built B2B companies with the same go-to-market shape. Founder ran ML at a competitor that was acquired for $410M, talent signal not yet on Crunchbase. The herd can’t see this without your data.
RecommendationLead. Window closes if a tier-one notices.
Built for partners
The report is a readable document, not a dashboard. Skim it on the train before the IC. Drill in for an hour the night before the partner call. Hand it to an LP who wants to know how you got conviction. No login. No tooling. Just the read.
/ the engagement
Pay per assessment.
Commit to an assessment, receive a second one half off. Or you can choose to pay for the service for a period of time.
01 · AI Due Diligence
Per memo, or by retainer.
Senior partner reviewer · Per-memo or quarterly retainer · Pricing on request
What’s included
- Thesis kickoff: we codify your investment criteria once
- 30-minute founder call captured and synthesized
- Scored memo against your thesis, in 5 days
- Risk register with mitigable / monitorable / dealbreaker tags
- Senior EXIT83 partner reviews every memo before it ships
- Memo can be white labeled for your firm
/ faq
Questions, answered.
No. We give them leverage. Your associates still run the relationships, ask the deep questions, and own the decision. We take the 30-hour structured write-up off their plate so they can do that on more deals.
We start with a thesis kickoff. We codify your investment criteria, your scoring rubric, your watch-outs and your historical pattern matches into a private skill. Every memo runs against that skill, so the output reads like your firm wrote it.
The founder call transcript, the data room you share, and public sources: SEC filings, Crunchbase, LinkedIn, news, customer review sites. We can plug into your PitchBook, Affinity or Crunchbase Pro seat under your contract; we do not resell their data.
Under signed NDA per deal. Data room contents are processed in an isolated workspace, scoped to that engagement, and deleted after delivery on request. No training on founder data, ever.
An EXIT83 partner with operating and investing background. They sign off on every memo before it ships. The AI does the volume, a human owns the recommendation.
Per-memo has no commitment. Quarterly retainers cancel with 30 days notice, no surprise renewals. We would rather have a short engagement that helped than a long one that did not.
· 30 minute intro call ·
Stop passing on deals you didn’t have time to read.
Book a 30-minute call. We will walk through your thesis, run a memo on a live deal, and tell you honestly if we are a fit.