Toast
AI voice order-taking agent with menu upsell (basket increase)
Toast launched ToastIQ and the Sous Chef AI agent in 2025 to suggest add-on items to the customer at the moment of ordering; a pilot restaurant saw its average order volume rise 6 percent, and ToastIQ was used by more than half of the 164,000 locations within four months of launch, with more than 8 million queries.
Key points
- Toast has an AI (Sous Chef, the ToastIQ ecosystem) suggest add-on items at the moment of ordering to increase the basket.
- A pilot restaurant saw its average order volume rise 6 percent thanks to the menu upsell tool.
- ToastIQ was used by more than half of locations within four months, with more than 8 million queries.
- At scale: 164,000 restaurants and ARR above 2 billion dollars (+26 percent) at end of 2025; voice agent for kiosks and drive-thru rolling out in 2026.
Objective
Increase the value of each order at the moment the customer orders, by having an AI suggest relevant add-on items (upsell), and make the platform more useful to the operator through an AI assistant grounded in its sales data. All across a base of 164,000 restaurants in the United States, where staff time is the main constraint.
The deployment
Toast, the point-of-sale and management platform for US restaurants, brings AI into the moment of ordering. In May 2025, the company launched ToastIQ, an intelligence layer that acts as a co-pilot for the operator (menu analysis, marketing, inventory) and houses customer-facing functions. The one most directly tied to revenue is a menu upsell tool attached to the Sous Chef AI agent: as the order is being built, the AI suggests the relevant add-on items to lift the basket. In the pilot program, one restaurant saw its average order volume rise 6 percent thanks to this tool (source PYMNTS, May 2025). This is a pilot result at a single location, not a generalized average across the base. The ecosystem itself is already widely adopted: in the fourth quarter of 2025, Toast reports that more than half of its locations used ToastIQ within four months of launch, with more than 8 million queries sent. The company also announced the rollout of AI voice agents for kiosks and drive-thru, with a multi-lane drive-thru product planned for 2026: this is where voice upsell most directly touches the end customer's experience. At scale: 164,000 locations at end of 2025, 30,000 net added during the year, and ARR that surpasses 2 billion dollars, up 26 percent.
Results Proof B
Toast's AI ecosystem is quantified in the Q4 2025 earnings call (ToastIQ adoption by more than half of locations within four months, 8 million queries, 164,000 locations, ARR of 2 billion dollars up 26 percent): this is financial-results-grade evidence for scale. But the upsell figure that carries the case, +6 percent in average order volume, is a pilot result at a single restaurant reported by the press (PYMNTS, May 2025), not a base-wide average. Hence a B level and not A: the scale evidence is solid, the evidence of impact on conversion/monetization remains at the pilot stage.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- plateforme ToastIQ Toast's intelligence layer launched in May 2025, which turns the restaurant operating system into a conversational co-pilot for the operator (menu analysis, marketing, inventory) and houses customer-facing AI functions such as upsell.
- outil Sous Chef Toast's AI agent and assistant, a conversational and voice lineage integrated into the ToastIQ ecosystem; it is the component credited with a menu upsell that increases average order volume.
How it runs, concretely
For ops teams-
1The customer starts their order customer
The customer interacts with the kiosk, the drive-thru, or voice order-taking, or the team enters the order on the Toast POS.
-
2AI upsell suggestion AI
As the basket is being built, the Sous Chef agent suggests the relevant add-on items based on the menu and the order context.
-
3Add to basket and checkout customer
The customer accepts the suggestion or not; the item is added to the order and flows to the POS for checkout.
-
4Operator steering marketing
The operator queries ToastIQ (menu analysis, marketing, inventory) to adjust the menu and offers, and tracks the effect on average order volume.
The structured menu (items, options, add-ons) and the order history. Without a clean, up-to-date menu, the AI does not know what to offer as upsell and the suggestions lose their relevance.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- A structured, up-to-date menu (items, options, add-ons, prices)
- First-party order history to refine the suggestions
- Instrumented order points (kiosk, drive-thru, POS) able to display or speak a suggestion
Org prerequisites
- A point-of-sale system able to host an AI upsell layer
- Business rules on what can be offered (margins, allergens, availability)
- A transparency and compliance framework if the upsell goes through a voice (AI Act, GDPR on the voice and the history)
Possible stack
- A voice or conversational agent connected to the order system
- An add-on recommendation engine grounded in the menu and the history
- An operator assistant to steer menu, marketing, and inventory
- A POS or kiosk as the display and checkout point
The plan, step by step
- Step 1Structure and clean up the menu (items, options, add-ons, margins) so the AI knows what to offer.Deliverable: A menu usable by a recommendation engine, with suggestion rules.
- Step 2Connect an upsell agent to a pilot order channel (drive-thru or kiosk) and define the suggestion moments.Deliverable: AI upsell active on one channel, at one or a few locations.
- Step 3Measure the effect on average order volume and the suggestion acceptance rate, and adjust the rules.Deliverable: A quantified reading of basket impact and the most-accepted suggestions.
- Step 4Extend to other channels and locations, and give the operator an assistant to steer menu and offers.Deliverable: AI upsell generalized plus operator assistant in production.
First step: Pick a high-volume order channel (drive-thru or kiosk) and test an AI upsell there on a selection of high-margin add-on items, measuring the effect on average order volume before extending.
The same pattern, proven in other industries
Grab
AI assistant for merchant advice and activation
Ping An
AI conversational agent for service and sales support on the customer base
Sources
- S1 Toast (TOST) Q4 2025 Earnings Call Transcript Established press archive pending
- S2 Toast: AI-Powered Menu Upsell Tool Boosts Restaurant's Average Order Volume 6% Established press archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.