Helm

Operating Layer for AI-First Funds

Helm is the cross-tool operating layer for AI-first funds and the portcos they back. It reads across Linear, Notion, Figma, Slack, and GitHub, surfaces hygiene drift, infers ownership, calibrates estimates, and generates partner-grade briefings.

The Problem

Funds investing in AI-first companies juggle five-plus shared tools per portco — and signal of "is this on track?" is buried in tickets, message threads, design files, and PR backlogs. Manual cross-tool synthesis doesn't scale past two or three deals.

The Solution

A cross-tool intelligence layer that:

• Reads, never writes (by default) — Observes Linear, Notion, Figma, Slack, GitHub via API; never replaces underlying tools
• Drift detection — Surfaces cross-tool hygiene drift (stale Linear tickets, orphan Figma files, GitHub PRs without ownership)
• Ownership inference — Cluster signals to infer who actually owns what, even when explicit assignment is missing
• Estimate calibration — Compares stated estimates against historical execution to surface miscalibration patterns
• Partner-grade briefings — Generates per-portco summaries the partner can actually act on, with concrete drift items and trend analysis

The Posture

Observes first, drafts second, schedules enforcement third. Augments existing tools — never replaces them.