Procurement and inventory automation
Kitchen ordering across 22 restaurants without the 6am phone calls
Twenty-two kitchens ordered from forty suppliers by phone, text and supplier apps, with chefs guessing quantities from memory. Food cost varied five points between sites. We built an ordering workflow that proposes each day’s order from sales forecasts and stock counts, sends it to the right supplier in their format, and reconciles deliveries and invoices automatically.
- Client
- Casual dining restaurant group
- Region
- United States
- Scale
- 22 locations, 1,400 employees
- Timeline
- 8 weeks, four pilot sites then rollout
Representative engagement. Client details are anonymised and figures reflect typical outcomes from comparable deployments.
What the client got
0.0 pts
Reduction in food cost percentage
Across the group, driven by forecast ordering and caught variances.
0 min
Chef time saved per day per site
Ordering is a two-minute approval instead of an hour of calls.
0%
Invoices matched automatically
The rest are exceptions with the discrepancy already identified.
The bottlenecks
Bottleneck 1
Ordering was a chef’s morning
Head chefs spent the first hour of every day calling and texting suppliers. Orders were based on what looked low in the walk-in.
Bottleneck 2
Invoices did not match deliveries
Short deliveries, substitutions and price changes were discovered weeks later in accounts payable, if at all.
Bottleneck 3
Food cost was a monthly surprise
With no link between sales, orders and stock, food cost percentages appeared at month end and nobody could explain the variance.
What we built
Step 1
Forecast-driven order proposals
Each site’s POS sales, reservations, weather and local events feed a demand forecast. Combined with par levels and last stock count, the workflow proposes the day’s order per supplier for the chef to adjust and approve on a phone.
Step 2
Supplier-format dispatch
Approved orders go out in whatever each supplier accepts: EDI, portal, email or a formatted text message. Confirmations are read back and discrepancies flagged.
Step 3
Delivery and invoice matching
Delivery notes are photographed and read by an LLM, matched against the order, and the invoice is matched against both. Shorts, substitutions and price variances are raised with the supplier automatically.
Step 4
Food cost in real time
Managers see theoretical versus actual food cost by site and category daily, with the top variances explained.
The automation workflow
Forecast to reconciled invoice
Daily ordering cycle per site
Questions clients ask about this engagement
Can chefs override the proposed order?
Always. The proposal is a starting point. Overrides are tracked so the forecast learns from them.
What about suppliers with no digital ordering?
They receive a formatted text or email and reply to confirm. The workflow reads the reply. Two suppliers still take phone orders, which the site logs manually.
More in Hospitality & Restaurants
Industry overview
AI reservation and guest messaging
An AI guest agent that answers, upsells and books across five hotels
Reservation enquiries came through email, WhatsApp, Instagram and the booking engine, and the small reservations team could not keep up outside office hours. We built an AI guest agent that answers in seconds, checks live availability, and completes or hands off bookings.
70%
Enquiries resolved without staff
Reputation and review automation
Every review answered within hours across 40 restaurant locations
Reviews on Google, TripAdvisor and delivery apps went unanswered for days, and head office only learned about problems from monthly reports. We built a workflow that drafts on-brand responses, routes serious issues to the right manager, and turns review themes into a weekly operations brief.
6 hrs
Median response time
Meetings and events sales automation
Group and event proposals out in 90 minutes instead of three days
Corporate event planners send the same request to six hotels and book the one that answers first with a clear price. This group’s sales teams took three days. We built a workflow that reads the RFP, checks space and room availability, prices against the revenue management rules, and drafts a proposal the sales manager can send after a short review.
90 min
Median RFP response time
B2B lead generation for extended stay
Corporate travel accounts for a serviced apartment operator
An extended-stay operator filled units through online travel agencies at heavy commission and wanted direct corporate accounts: consultancies, project-based engineering firms and relocation programmes. We built outbound to travel managers, HR mobility leads and project directors, and the operator replaced commission-heavy bookings with negotiated corporate rates.
34
Corporate accounts signed
Have a similar bottleneck?
Tell us how the work happens today. We will show you the workflow we would build and prototype it on your own process.