Enterprise tier
At eight-figure GMV, reconciliation is infrastructure
At eight-figure GMV, "running an audit" is the wrong mental model — audits are events, and events have gaps the size of quarters. What the scale demands is infrastructure: event-driven detection wired into operations, recovery governed by policy rather than heroics, reporting that feeds controllers and boards without bespoke archaeology, and evidence ready for the diligence or audit that scale eventually attracts. Here is the programme spec, component by component.
Component one: the real-time pipeline
Webhook receiver with signature verification and durable claiming; queue-based processing with retry semantics; race protection via delay-plus-per-charge-locking; detectors running on every relevant event within seconds. This is exactly the architecture FeeGuard operates in production — the blueprint exists whether you adopt or mirror it.
Component two: governed recovery
Automation behind gates — realised losses only, risk-scored, floored and ceiled, notice-period-respecting, circuit-broken. Human queues own judgement calls with SLAs anchored to payout math. Governance documentation matters at this scale: who may approve what, logged immutably, reviewable quarterly. Surprise clawbacks reaching sellers before policies reach them is how enterprise programmes die publicly.
Component three: reporting surfaces
Controllers need journal-ready exports matching their ERP structures. Finance ops needs queue metrics trending weekly. The board needs one slide: exposure, baseline, recovered, causes-closed. Three audiences, one dataset, three views — designed once rather than hand-assembled every cycle.
Component four: diligence readiness
Growth-stage platforms meet diligence eventually: M&A, debt raises, audits. Leakage programmes generate the artefacts those processes request — measured baselines, recovery history, control descriptions with evidence trails — converting what would otherwise be a discovered liability into a demonstrated competency. The best time to start was before the data room opened.
Build-vs-buy at this scale
Dedicated engineering capacity changes the calculus — but maintenance reality persists regardless: schema drift, edge-case corpus growth, alert-fatigue tuning. Adopting buys the maintained stack today with CSV/API integration into sovereign ledgers; building buys control plus the roadmap. Price both honestly including the tail.
Enterprise shape
Multiple Stripe platform accounts, SSO/SCIM, signed DPA, dedicated success management — configured per agreement, never promised generically. Recovery fees negotiate at volume with floors held internally; procurement conversations welcome security packets early.
Common questions
SOC 2 available?
Report available under Enterprise agreements. We provide evidence layers; certification claims belong to auditors, and we never imply otherwise.
We run multiple Stripe platform accounts — supported?
Yes, under Enterprise configuration. Consolidated views with per-entity policies; details shaped per agreement.