1. Connect
One OAuth flow per source. Stripe, Shopify, Klaviyo, Google Ads, Meta Ads, GA4, Search Console, HubSpot, QuickBooks, Zendesk — 10 production integrations on day one.
Connect your stack. We diagnose where revenue is leaking against vertical benchmarks, project the dollar lift of fixing each, and — when you want it — execute the fix. White-label tier for agencies.
Deterministic numbers. AI-grounded narration. No hallucinated dollars.
One OAuth flow per source. Stripe, Shopify, Klaviyo, Google Ads, Meta Ads, GA4, Search Console, HubSpot, QuickBooks, Zendesk — 10 production integrations on day one.
Eight named revenue metrics extracted directly from your synced data. Compared against vertical benchmarks sourced from public reports (Klaviyo, Shopify, Baymard, WordStream, FirstPageSage), every claim is citable.
Each leak gets a per-formula dollar projection. The LLM only narrates — every numeric token is whitelisted before it reaches you, with retry-then-fallback to deterministic templates.
Done-for-you fulfillment from $1.5K Focus engagements to $5K/mo Partner retainers, or hand the report to your in-house team and execute internally. We don't charge for the report itself.
A multi-client diagnosis platform. Your brand, your clients, our infrastructure.
One agency account, unlimited client workspaces. Per-client source connections, per-client diagnoses, per-client share links. Designed for fractional teams and growth marketing shops.
Custom logo, subdomain, and color tokens on the agency tier. Share read-only reports under your brand, not ours.
The architectural difference between a credible diagnostic tool and a ChatGPT wrapper.
Customer metric minus benchmark p50, multiplied by your traffic or revenue, with the formula visible to anyone who clicks. The LLM never produces a number.
124 rows across 5 verticals, sourced from named public reports with material corrections logged. Low-confidence benchmarks (n<100) are flagged in the UI before they reach a customer.
Each connector's column contract is verified before extraction runs. If a vendor adds, renames, or removes a field, the diagnosis aborts rather than producing a wrong dollar projection.
Diagnosis output is content-addressed by metric values + benchmark cohort. Two agencies running the same diagnosis on the same workspace at the same time get the same report — no LLM nondeterminism leaks through.
Backend is feature-complete and smoke-tested. UI is in active development. Pilot agencies (free for the first 5) get founder-led onboarding and a free DFY engagement.
First-5 pilot agencies: free year of the Agency tier + one DFY engagement on us.