Client: Restaurant group
Interactive business intelligence tool (pilot)
The problem
Mission-critical data that restaurant managers use to analyze past performance and plan the weeks ahead was scattered across eight stand-alone systems that do not interact. Across the portfolio, managers spent roughly 400 hours a week pulling and analyzing it.
The solution and impact
Eight systems, one conversation: Eight systems, fragmentedEight systems, a single conversation
- A system on its own
- Stitched together by hand
- Connected in one place
- Outside signals: events, weather
- 8 stand-alone systems that don't interact
- 4 hours a week per manager spent pulling reports and planning
- 0 forward-looking views of next week
- 1 place to ask any question about the business
- 4 hours a week back for each manager
- 400 hours a week back across 100+ locations at full rollout
-
Map
We mapped every system the restaurant teams rely on, and every piece of data within each one.
-
Prototype
We built a chat interface on exported data from all eight systems. A scheduled agent logs into each system, exports the data and files it for the tool to read. No API integrations and no data warehouse: the goal was to demonstrate what's possible before the client commits to either.
-
Ask, or be told
Managers can ask any question in plain language, or press one button for what to focus on next week. Answers draw on history, bookings, events, the local calendar and the weather.
Where it stands
The pilot proved the value. Taking it to production means API integrations and a data warehouse, a larger investment the client has chosen to make at a later stage.
Read the diagram as text
Before: Eight systems, fragmented
- Reservations, point of sale, inventory, staff scheduling, private events management, accounting, payroll and training each sit in a separate system.
- A manager pulls a report from each system by hand.
- The reports are stitched together in spreadsheets.
- The result is a look back at what happened, never a look ahead.
After: Eight systems, a single conversation
- All eight systems feed one interface, together with local events and the weather forecast.
- A manager asks: “I need to reduce labour by $700 next week. How can I do that without impacting guest experience?”
- The tool answers: the weather looks gloomy next Tuesday and Thursday afternoons, so the patio will be slow. That's the best opportunity. Front of house, Jonathan and Rachel have the most flexibility in their schedules. Back of house, Kyle and Alex. Removing their 11AM to 3PM shifts and having them start late should save $704 next week, with minimal impact on guest experience.
- Every Monday, one button answers “What should I focus on next week?”: for example, downtown Chicago will be incredibly busy, with the HIMSS Global Health Conference & Exhibition expected to draw 25,000 attendees and the Bulls in the NBA play-in tournament adding roughly 20,000 fans, so prepare for a big week.
- Walk-ins are usually 25% of tables; expect 40% next week. Add three servers on the floor (Blair, Isaac and Paige are free in the evenings) and ask Chef to raise prep levels by about 20%.