Your open call, actually read.
You publicly invite founders to apply. Then the median submission gets 150 seconds and silence. EvalLens reads every submitted deck in full against your investment dimensions and hands your team a comparable, evidence-linked view of the whole batch — every deck with flags, founder questions, and a quote behind every finding.
Either read it properly — or accept the call is optics
- 2:31minutes a cold deck actually gets read, vs 4:18 for a warm introDocSend engagement research
- ~60%of reading attention lands on the first four slides. Most of every submission is never weighedDocSend
- 3,000 → 9decks an analyst sees per year vs deals the fund closes. The skim is a rational answer to that mathfunds' own math
- “Too early”what most founders hear back — if anything. Every ghosted founder is your program's NPS and your scout network's credibilityfounder surveys
Built as the aide. Never the arbiter.
Yes — this is the axis Google Ventures tested for you.
GV built “The Machine” to score deals, watched it drift from diligence aide into de facto investment committee, and shelved it in 2022 — Axios’ epitaph: data went back to being “aide, rather than arbiter.” We built on the right side of that line on purpose: six AI reviewer roles — not six people; six adversarial reads that can’t anchor on each other — prepare the evidence. The scores are advisory. The shortlist is built from your partners’ decisions, and there is no mode where the machine says yes or no to a deal. Conviction was never the part that needed automating. Reading was.
Application window → pipeline meeting, in five steps
EvalLens prepares the comparable first read. The shortlist is your partners' — always.
- 01
Frame the program
Open call, scout batch, fellowship or demo day — your investment dimensions and thesis-fit criteria become the shared rubric, locked before the window opens. Your Typeform/Airtable intake stays; we ingest the batch. You get: one standard every reader shares.
- 02
Every deck read in full — as it lands
Every page, coverage logged; reports ready as submissions arrive. Signal engines can't see a pre-seed founder with no web footprint. A full read can. You get: deck #1 and #400 on the same standard.
- 03
A brief per deck: flags, questions, quotes
Red flags, the three questions you'd ask the founder, and a page-referenced quote behind every finding. Where reviewer roles disagree, the gap goes to your team as an open question — never averaged away. You get: screening calls that stop being first reads.
- 04
One ranked view. Partners decide.
The batch becomes a single comparable board; analysts review the borderline; partners confirm or override, and their calls build the shortlist. The evidence pastes straight into your own memo. You get: an answer for every rank, meeting-ready.
- 05
Pass with feedback — under your control
Substantive, evidence-based pass-feedback for every founder — reviewed and editable by your team before anything is sent, tone configured to your house style, opt-out per call. The rare funds that do this treat it as a sourcing edge; for you it's a byproduct with a safety catch. You get: founder goodwill without the letterhead risk.
“Why is this deck ranked #4?”
Every rank traces to findings your team can check — and challenge — before anything reaches the partnership.
Quotes are verified against the deck before a finding stands. No quote, no finding.
Feedback founders remember — without the screenshot risk.
AI-generated feedback going out under a fund’s name is one hallucinated critique away from a thread on X. That’s why the feedback engine ships with a safety catch, not a fire hose.
Priced against the skim, not a fantasy. A 400-deck call gets roughly 17 hours of skim today — 60% of it on the first four slides. EvalLens reads every page of all 400, flat per program, inside a platform budget: no partner approval, no IC, no seat licences. Failed runs never billed. See pricing or book a call.
What your skeptical GP will ask
- “Isn't this the thing Google shelved?”
- It's the opposite side of that experiment. GV's “Machine” drifted from aide to arbiter and was shelved for it (Axios, Sep 2022 — worth the read). EvalLens is architecturally the aide: it reads and prepares evidence; ranking is built from partner decisions; there is no autonomous yes/no. The GV story isn't our risk — it's our design spec.
- Our real pipeline is warm intros. Why bother?
- This isn't for the warm pipeline. It's for the programs you run for coverage — open calls, scout batches, demo days — where inbound currently gets 150 seconds a deck. Either read it properly or accept the call is optics; we make reading it properly cost less than a week of associate time.
- We already run Harmonic / Specter / Affinity.
- Keep them — different job. Signal engines discover companies via external data; a pre-seed founder who just applied often has no signal footprint at all. Nothing in that stack reads submitted decks against your criteria. Complementary, not competitive.
- What happens when the AI misquotes a deck?
- Quotes are verified against the source document before any finding stands — no quote, no finding. Where evidence is thin, scores move down, never up, and the gap is flagged to your team rather than papered over. And nothing founder-facing leaves without your sign-off.
- What are founders told?
- Template disclosure for your application form: AI assists the first read; partners make every decision; decks are processed only for this program, never used for training. Transparency here is a sourcing asset, not a risk.
- Who else runs this?
- Founding cohort stage — which is why the first step is a parallel run on your own batch: your analysts screen as usual, EvalLens reads the same decks, and you compare agreement and catches before committing anything.
Run it in parallel on your next call.
Under NDA, we demo on your own decks first. Then one batch, side by side: where the ranked view agreed with your analysts — and what the skim missed. The first run is free through August 31, for batches up to 10 decks.