We built a custom platform that consolidated every stage of the application review cycle. Then we kept building as the program grew.
Internal platform — not publicly accessible
“The Review app is a great tool for assessing applications — easy to use, with a strong UI and helpful features that meaningfully reduce manual effort.”
— Accel Atoms team
The platform consolidated the team’s entire review workflow under one roof. AI features, layered in across cohorts, made navigating and assessing applications faster for every analyst on the team.
Accel is one of the world’s most active early-stage investors. Their Atoms program runs multiple cohorts a year, each generating a high volume of applications that a small analyst team has to move through quickly.
Slow tools cost real decisions. Fragmented analytics meant no one had a clear picture of where the cohort stood. A confusing UI meant analysts spent time navigating the tool instead of evaluating companies.
The investment review cycle had friction at every stage.
Phase 01
~2 weeks
Before designing anything new, we stabilised the review queue. A rapid interim solution shipped in two weeks to keep analysts moving while we scoped the real platform. The goal: unblock the team, then understand what they actually needed.
Phase 02
~1 month
Multiple rounds of interviews with the Accel investment team — surfacing what worked in the old flow, what was actively hurting them, and what the right product needed to do differently. We combined those learnings with hands-on review of how the team actually worked, then designed a new platform from scratch: clean application views, a structured analyst workflow, and analytics built into the product itself.
Phase 03
Ongoing across cohorts
We shipped the platform and kept building. As each new Atoms cohort ran, new demands surfaced — AI search across the full application database, AI-generated company overviews, pitch deck summaries, website summaries, cohort-level data segregation, and user-level access management.
Find any company, founder, or detail across the full application database instantly.
AI-generated summaries surfaced inline on the listing page. Context at a glance, before opening a single application.
Key thesis, team, traction, and ask pulled automatically from uploaded decks.
Per-applicant website summaries on demand. No tab-switching to research companies mid-review.
Cohort-level data segregation with clean switching between programs
User-level access management
Pitch deck summaries on demand
Website summaries per applicant company

How to pronounce EPYC?
© 2026 EPYC