Keystone

You run multiple units and manage them the way you managed one — by staying close to everything. Every Monday you pull reports from three different systems, reconcile numbers that never quite match, and spend the first hour of your week figuring out which location is underwater before you can act on any of it. Your regional managers send updates in different formats. Your comp data lives in a spreadsheet someone else built. You are the integration layer between every system in your portfolio, and it is costing you decision time you do not have. Keystone is a franchise operations and multi-unit business intelligence persona built for operators running 3–50 locations who need system-level visibility without adding headcount. Every workflow was extracted from real multi-unit operational patterns — not franchise consulting theory. The unit health scoring system tracks 14 operational variables per location on a daily cadence, built specifically because a single underperforming unit spent 11 weeks dragging a 12-location portfolio's comp numbers before it surfaced in a standard report. The anomaly detection protocol exists because of that gap. Unlike a generic operations assistant, Keystone never aggregates metrics without flagging the outlier — because portfolio-level averages are how bad units hide. That rule exists because "average performance" across eight locations can mask one unit in freefall and three at risk. Keystone surfaces the distribution, not the mean, and attaches a specific recommended action to every anomaly it identifies. It will not hand you a dashboard. It will hand you a ranked list of what to fix and in what order. What you get: SOUL.md — Keystone's operational identity and decision-making framework. IDENTITY.md — role definition, escalation protocols, and communication standards for multi-unit context. UNIT_HEALTH.md — 14-variable location scoring system with daily update cadence and threshold logic. ANOMALY_PROTOCOL.md — detection and response workflow for surfacing performance outliers before weekly reporting. REPORTING_CADENCE.md — weekly, monthly, and quarterly report structures across all units. Requires access to unit-level POS, scheduling, and P&L data sources.