You rated the stay a 6. What would have made it a 10?
D
Post-Stay Survey
Evaluator agent · 1.9K uses
Evaluator Agent
New Menu Item Customer Feedback
Strong first-week sales can still be a failure if it cannibalized something better. New menu items are judged on units sold in the first weeks, which rewards novelty and hides cannibalization. This conversation asks the three things that actually predict whether an item earns its place: whether it met expectations, whether they would order it again, and what they would otherwise have ordered.
Menu mix margin protected
Cannibalization made visible
Expectation gaps caught early
Used 1,272+ times
What's inside this template
Start from this conversation and adapt it to your team — change any question, add your own logic, and connect the tools you already use.
Information it collects
What they expected from the menu description or photoWhat they actually receivedWhether it met, missed or beat expectationsWhether they would order it againWhat they would have ordered insteadWhether it displaced a higher-margin item
Questions it always asks
The core fields every response captures.
Always ask what they expected before asking what they got
Always ask what they would have ordered if this item had not existed
How it adapts
Follow-ups that change based on what people say.
If it missed expectations, ask whether the menu description or the photo set them wrong
If they would not reorder, ask what they will order instead next time
Where it routes people
Different paths for different answers.
Send expectation gaps to the menu and content team
Flag cannibalization of higher-margin items to finance before rollout
Automations it can trigger
Actions that fire the moment a response comes in.
Post item-level test results to #culinary in Slack
Update the item record in HubSpot with reorder intent and cannibalization
Alert finance when a test item displaces a higher-margin dish
SOC 2 Type II and ISO 27001:2022 certified. Responses are encrypted in transit and at rest, and you own your data. View our Trust Center.
How this AI agent works
The conversation asks the diner what they expected when they ordered the new item, based on the menu description and any photo, then what they actually got. It asks directly whether they would order it again or go back to their usual, and what they would have ordered if the new item had not been there. That last question is what turns a sales number into a real menu decision.
Getting started
- 1
Trigger from a receipt link on checks containing the test item
- 2
Run it across enough locations and dayparts that the result is not one kitchen
- 3
Capture what the item replaced so cannibalization is visible
- 4
Feed the results into the menu decision before the full rollout
Template Details
- Agent Type
- Evaluator
- Category
- Hotels and hospitality
- Industry
- Hotels & Hospitality
- Business outcome
- Build the right product
- Job to be done
- Test concepts & creative
- Journey stage
- Feedback & evaluation
- Replaces
- Survey tools
- Integrations
- Slack, Hubspot, Email
- Times Used
- 1,272+
Forms collect fields. Conversations capture context.
Static forms force complex situations into rigid dropdowns. Perspective captures structured data and the reasoning behind it — so your team makes better decisions, faster.
The static form
yoursite.com/intake
Category *
Select...
Details
Describe your situation...
Submit
Result:Category: "Other"|Details: "It's complicated"
No context. No follow-up. No next step.
- Unit sales in the first weeks reward novelty rather than staying power, and nearly every new item sells well initially because regulars try it once.
- Sales data cannot see what the item replaced. An item that sells strongly by displacing a higher-margin dish can reduce total profit while appearing to succeed.
- Taste-test panels evaluate the food in isolation, without the menu description and photo that set the guest's expectation, which is where a large share of disappointment actually originates.
The AI conversation
"Tell me more about the timeline — when did this start, and is there a deadline your team is working against?"
Extracted & structured automatically
Category
High-priority
Urgency
Deadline: 2 weeks
Sentiment
Frustrated but hopeful
Next step
Route to senior team
Triggered: Slack alert sent| CRM updated
Right team. Full context. Instant action.
- The conversation captures expectation before delivery, which separates a food problem from a description or photography problem. Those have different owners and very different costs to fix.
- Asking whether they would order it again rather than whether they liked it is a far better predictor of whether the item earns a permanent place.
- The cannibalization question, what they would have ordered otherwise, is the single most valuable output and is invisible in every other data source.
What is menu testing?
It is structured feedback on a new or trial menu item from guests who actually ordered it in a live service environment. It differs from taste panels, which evaluate food under controlled conditions without the menu context, and from sales tracking, which measures units without explaining them. The purpose is to decide whether an item deserves a permanent listing.
FAQ
Frequently Asked Questions
Explore More Agents
More hospitality templates for upsell, takeout and restaurant win-back.
Did the way they handled it actually fix how you felt?
E
Service Recovery Survey
Evaluator agent · 2K uses
What did those first ten minutes do to the stay?
S
Check-In Experience Survey
Evaluator agent · 1.1K uses
Did you mention it to anyone at the time?
O
Room Quality Survey
Evaluator agent · 1.3K uses
Did the works actually affect your stay?
M
Renovation Impact Survey
Evaluator agent · 2.1K uses
Was the order right when you got home?
A
Drive-Thru Survey
Evaluator agent · 2K uses
Related Articles
How restaurant brands judge a new item beyond first-week units.
Best Hotel Guest Experience Software in 2026: 8 Platforms Compared
Eight guest experience platforms compared by what they actually capture.
Read article
Restaurant Customer Feedback in 2026: From Comment Cards to Conversations
Why restaurant feedback collapses into stars, and what a conversation recovers.
Read article
Toast's AI Strategy: How the Restaurant Platform Learns What Operators Need in 2026
How a hospitality platform uses conversations to understand its operators.
Read articleForms are costing you business
Replace drop-off, poor qualification, and missing context with AI conversations that capture structured data and real understanding. Set up in minutes.
No credit card required • Cancel anytime