How VCs screen inbound pitch decks

Screening is the least glamorous work in venture capital and the most consequential. Every fund runs some version of the same funnel, whether it admits to a process or not. Here is what that funnel looks like in numbers, what the screen actually checks, and where software has started to change the economics.

The funnel is roughly 100 to 1

The largest academic survey of institutional VCs, by Gompers, Gornall, Kaplan and Strebulaev (885 funds), found that for each deal a fund closes it considers about 100 opportunities. Roughly 1 in 4 pitches gets a management meeting, and a median fund closes about 4 deals a year. Multiply it out and a small fund is reading hundreds of decks a year to make a handful of investments. The screen is where 75 percent of those decks exit the process.

The first read lasts under two minutes

DocSend's investor behavior research measured the average time investors spend on a seed deck at 1 minute 56 seconds, the first year ever under two minutes, down from 3 minutes 18 seconds in 2021. That is not carelessness; it is triage. A reader with two hundred decks in the queue gives each one exactly enough attention to answer a single question: does this deserve a meeting?

What the screen actually checks

Talk to enough screeners and the checklist converges. In rough order of how fast each can kill a deal:

• Fit: stage, sector, geography and check size against the fund's mandate. The fastest pass there is.
• Team: in the same NBER survey, 95 percent of VCs rate the team as an important factor and 47 percent call it the single most important one.
• Coherence: does the ask match the financial plan, do the numbers agree with each other, is there a believable use of funds.
• Evidence: traction appropriate to stage. Not revenue at pre seed, but demand evidence the deck can substantiate.
• Risk signals: regulated market with no compliance story, hockey stick plans that ignore the sales cycle, adjectives where unit economics should be.

The economics of the screen are broken

A partner hour is the most expensive resource a small fund has, and the screen consumes it on decks that mostly exit the funnel. The common workarounds all cost something: interns and analysts screen inconsistently, "warm intro only" policies outsource the screen to the network and miss outliers, and unanswered inbound quietly damages a fund's reputation with the founders it will want to see again a round later.

Where software fits

Deal flow CRMs organized the pipeline: who sent what, which stage it sits in. What they never did is read the deck. The newer layer is AI screening, where the deck content itself is analyzed: a summary, scores against the fund's own criteria, and flagged risks, produced in minutes for every inbound deck rather than for the few a human gets to. The screen stays the fund's judgement; the software makes sure that judgement is applied to every deck, consistently, and that the obvious passes never reach a partner's queue at all.

That is the layer Bizznote builds: AI deal flow screening on your criteria, with a one click reply so every founder gets an answer.