Menu & pricing for quick-service restaurants

In quick-service the menu is a screen, not paper. Digital menu boards are on 62% / 82% QSR of fast-food and QSR signage, the highest adoption of any format, which means Menu engineering stops being a once-a-year print job and becomes a live control surface. You can move your highest-margin combo to the top-left where a Texas A&M Human Behavior Lab eye-tracking study found a +20% profit profit lift from repositioning, shrink a low-margin loss-leader, and 86 (to 86 an item) a sold-out item across every board the instant the walk-in runs dry.

AI does the ranking a manager never has time for: it reads your Point of sale (POS) history, scores each item by contribution margin and velocity, and tells you which four or five products deserve the board's best real estate. Platforms like Toast and Square for Restaurants hold the sales data and drive the boards, so the engineering and the display live in one system rather than a spreadsheet nobody updates.

The drive-thru is where a QSR earns most of its living: nearly ~75% of restaurant traffic is now off-premises, and 58% LS / 41% FS for limited-service. Drive-thru AI adds a tireless upsell on every car, and it makes about 4x more context-specific attempts than a rushed crew member. That is a direct lift to Average check and a way to steer mix toward the combos you actually want to sell. Presto Automation reports up to up to +6% revenue, 25+ states/provinces with reported accuracy gains, and voice ordering now runs at scale (890+ restaurants, and ConverseNow deployed chain-wide at 60 locations).

Be honest about the ceiling. In a mystery-shopper study of confirmed voice-AI drive-thru visits, employees still had to step in on 22% employee intervention rate of interactions, and 2.5-year test ended with IBM after two and a half years. This is production-grade but supervised technology. For a single-unit operator, the win today is the upsell and order-taking on peak lanes with a human on standby, not a lights-out replacement for the crew. Vendors worth a look include SoundHound AI (for Restaurants) and Hi Auto alongside the ones above.

Menu engineering fails if the featured item is 86'd at noon or if prep for it becomes trash at close. Demand forecasting closes that loop by predicting orders per item per daypart. Lineup.ai markets its sales forecasting as 35% more accurate than traditional methods, 7shifts runs across 55,000 restaurants / 1.5M workers and markets POS-integrated forecasts as 95% accurate, and an item-level model built for Blue Bottle improved ordering accuracy by 8% accuracy gain in its first month. A sharper forecast lets you thaw, prep, and staff to the actual curve rather than a manager's memory of last Tuesday.

That same forecast is your labor and waste dial. Limited-service labor sat at 31.7% of sales of sales in 2024, so scheduling to the predicted rush protects the second-largest cost on the P&L, and touchless waste tracking like Winnow reports 2-8% food-cost savings. Pair it with real-time food costing from MarginEdge or Restaurant365 so every board decision is checked against what the ingredients actually cost this week.

Dynamic pricing is real in QSR, but a value guest punishes anything that smells like surge. The safer move is structural: use AI to price combos and bundles so the highest-margin path is also the most attractive, run gentle daypart pricing on the digital board, and reprice individual items only when invoice costs actually move. With restaurant prices up +4.1% in 2025 and 40% profitability goal / 48% raise prices, the difference between a smart mix shift and an across-the-board hike is whether you keep the traffic.

Because a QSR board is programmable, you can test a price on a subset of items, watch the mix and check average for a few days, and roll it back with no reprint if velocity drops. Let the tools that hold your food cost flag the drift for you: when a supplier raises a key ingredient, the system that sees your Food cost percentage in real time (MarginEdge, Restaurant365, or 5-Out) should be what tells you a combo went underwater, not a surprise at month-end. Explore what fits your stack on the /tools page, learn the terms on the /glossary, and see the general playbook at menu engineering and pricing.