Insights
AI for independent sponsors: screening inbound deals without a team
Mitchel Campbell, Co-founder and CEO
· 6 min read
The short answer
An independent sponsor's binding constraint is attention, not deal flow. The highest-return thing to build is a consistent first-pass screen that reads every inbound teaser within a day and sorts it, so the 80 percent that will never fit die quickly and the rest reach you with the work already done.
That is a narrower claim than it sounds. The screen does not decide. It reads, scores against criteria you wrote down, and routes. You still say yes and no.
The problem, stated honestly
A sponsor with one or two people sees more deals than they can read properly. Bankers know this, which is why teasers arrive in volume and why a thin response rate is tolerated.
What happens in practice is not that bad deals get through. It is that good deals get a slow reply. A teaser that lands on a Tuesday in a busy week gets opened on Sunday, and the banker has already had three conversations. The cost of being under-resourced is not poor judgement; it is being late and being inconsistent – different criteria applied on a quiet week than on a loud one.
Those two failures are exactly what a screen fixes, and they are the only two it should be asked to fix.
What a first-pass screen actually does
Three steps, and it is worth being precise because the middle one is where most attempts go wrong.
Extract. Pull the facts out of the teaser or CIM into a consistent shape: sector, revenue, EBITDA, margin, geography, customer concentration if disclosed, ownership situation, process timing, banker. Teasers are PDFs with inconsistent layouts, so this step is unglamorous and it has to be reliable before anything downstream matters.
Score. Compare the extracted facts against your stated criteria and produce a ranked result with the reasoning attached. Not a number on its own – a number and the sentence that explains it, so you can disagree with it in five seconds.
Route. Three buckets. Clear fit goes to you with a short brief. Clear miss gets a polite decline drafted for your approval, which is the part that preserves banker relationships most sponsors quietly damage by not replying. Everything in between goes to a queue you look at once a day.
What it must not do
It must not send the decline on its own. Not because the drafting will be bad, but because a banker who gets an obviously automated no from a sponsor remembers it, and your deal flow is a relationship business with a very small number of counterparties.
Keep a person on the send button. The time saving is in the reading and the drafting, not the clicking.
The part most sponsors have not done
Here is the uncomfortable finding from doing this work: most independent sponsors' investment criteria are not specific enough to screen on.
"Lower middle market, founder-owned, 2 to 8 million of EBITDA, business services, North America" is a positioning statement. It is not a screen. Run it against a hundred teasers and it will pass thirty, which is not a filter.
A screen needs the next layer down, and it is usually in the sponsor's head rather than on paper:
- Which sub-sectors inside business services, and which are hard noes
- What customer concentration is disqualifying rather than merely noted
- Whether you will look at a carve-out, a turnaround, a roll-up platform, an add-on
- Minimum revenue as well as minimum EBITDA, and whether a margin threshold applies
- What process dynamics you avoid – broad auctions, a banker you have had a bad experience with, a seller staying on
- Whether you have capital partners with their own constraints, and what those are
Writing this down takes an afternoon and it is the highest-value afternoon in the project. It also tends to surface disagreements between partners that have been running quietly, which is uncomfortable and useful in that order.
The 30-day version
You can test this without committing to a platform.
- Take the last 50 teasers you received and the decisions you made on them
- Write the criteria down at the level of detail above
- Score those 50 against the written criteria by hand and compare to what you actually did
- Where the screen disagrees with your decision, work out which was right – often the criteria are wrong, not the decision
- Only then build it
Step 4 is the whole exercise. A screen that reproduces your past decisions is worth having. A screen built on criteria you have never tested is a confident way to miss things.
What it costs you in attention
Running, roughly fifteen minutes a day: reading the briefs on the clear fits, glancing at the middle queue, approving declines. That compares to the two to four hours a week most sponsors spend opening PDFs to discover they are not interested.
The build is a scoped first workflow, not a platform rollout. If someone proposes a platform to an independent sponsor, they have not understood the firm.
Frequently asked questions
Can AI screen inbound deals for an independent sponsor? Yes, for the first pass. It extracts the facts from a teaser or CIM, scores them against written criteria, and routes the result. The investment decision stays with the sponsor.
How accurate is the extraction from teasers? Good on disclosed figures, unreliable on anything inferred. Treat extracted revenue and EBITDA as a starting point to confirm, not as a fact, and keep the source page reference so it can be checked in seconds.
Will it replace an associate? No. It replaces the first thirty minutes of reading on deals that were never going to work, which is the least valuable thirty minutes an associate spends.
What if my criteria change? They should, and the screen should be a document you edit rather than code someone else maintains. If changing a threshold requires a vendor ticket, the design is wrong.
Do I need a data room integration? Not for screening. First-pass screening works on the teaser, which arrives by email. Diligence automation is a separate and later question.
How we do it, for completeness
Analyst3 builds screening and sourcing workflows that run in a sponsor's own accounts, on the systems they already use. We start with the criteria exercise above, because a screen built on vague criteria is worse than no screen. Nothing reaches a banker or a seller until someone at the firm approves it, and the sponsor owns the code, the data and the workflow.
We spend more time on the criteria than on the build, which is not a good sales pitch and is nonetheless the right order.
If you run the 30-day version yourself and never call us, the guide did its job.