AI-Powered P&L Coaching System

The Challenge

A multi-brand restaurant corporation needed a way to coach general managers on financial performance, consistently, every week, across 70+ locations. The existing process was manual and inconsistent: district leaders would review P&L reports when they had time, and the coaching that happened was often generic rather than tailored to each store’s specific situation.

The critical gap was connecting financial data to actionable coaching. Managers had access to dashboards, but interpreting trends, spotting problems early, and knowing what to do about them required expertise that wasn’t available at every location every week. District leaders, meanwhile, were spending hours pulling reports and preparing for coaching conversations instead of actually coaching.

The Approach

I built an AI coaching system that generates personalized weekly performance reports for every location, delivered by email every Tuesday morning.

The system pulls data from multiple sources via custom MCP servers:

• Financial metrics — period-to-date and weekly sales, labor cost %, food cost %, guest counts, and comp percentages, all compared against budget and district/brand benchmarks
• Peer comparison — each store’s metrics benchmarked against district and brand averages, so managers can see exactly where they stand relative to their peers
• Customer satisfaction — OSAT survey scores and verbatim customer comments, including specific employee mentions
• Goal tracking — the prior week’s AI-assigned goal is automatically evaluated against actual performance

Each week, the AI coach analyzes the data and produces a structured meeting outline for the general manager:

1. Period-to-date summary with interpretation
2. Last week’s performance with local market context (weather, events, promotions)
3. Cost breakdown (labor, food, packaging) with status indicators
4. Customer satisfaction trends with highlighted comments
5. Bonus qualification status: exactly where they stand and what’s at stake
6. Evaluation of last week’s goal (pass/fail based on actual metric movement)
7. A new goal for the coming week, dynamically selected by the AI based on where the store has the most room to improve
8. Specific action items with implementation steps, measurement criteria, and expected impact

The goal system creates a closed loop: the AI identifies each store’s biggest opportunity, assigns a targeted goal, and the following week’s report evaluates whether the needle moved, creating accountability that adapts to each location’s evolving situation.

Key technical decisions:

• Multi-model architecture — Claude for analysis and writing (with prompt caching for cost efficiency), GPT-4 for structured goal extraction, Tavily for local market context
• Prefect orchestration — the weekly flow processes all locations concurrently
• MCP-based data access — financial metrics and configuration data are accessed through FastMCP servers deployed on Vercel, keeping the coaching system decoupled from the data sources

The Results

Managers and district leaders each save roughly two hours per week. Managers no longer have to manually pull and interpret reports, and district leaders no longer have to prepare coaching materials for every location.

More importantly, managers now have a consistent, data-driven picture of their store’s performance every single week. They see how they compare to their own historical trends and to peer locations across the district and brand, context that was previously only available in quarterly business reviews, if at all.

The AI-assigned goals mean that every location is working on its highest-impact opportunity at any given time, not a generic corporate initiative. A store with strong sales but rising labor costs gets a labor-focused goal. A store with declining guest counts gets a traffic-driving goal. The coaching adapts to the data, not the other way around.

The system runs autonomously every Tuesday with no human intervention required to generate, personalize, or deliver the reports.