03 · Intelligence and decisions
LiveHIRO Brain
Your team does not need another dashboard. It needs an answer
The Analyst
The Brain reads the house memory and returns the play: who is missing, who is slipping away, which table is oversold, which dish is not returning food cost, who to call before Friday. Every answer carries a citation to the data behind it, so it can be argued with. A dashboard shows you a number. This tells you what to do with it.
Working today with real users.
Her usual Riesling. Allergy: tree nuts.
No booking since March. Birthday on Saturday.
A dashboard shows you a number. An employee tells you what to do with it.
The work
What it does, exactly.
Every line carries its own status. If something is in beta, it says so on the line, not in a footnote.
- Reads real guest and menu behaviourLive
- Detects patterns, drops and opportunitiesLive
- Daily digest with the moves of the dayLive
- Proposes the next action, cited to the dataLive
- Ask it in plain language about your houseBeta
What it does not do yet
The natural-language assistant is in beta, available on the Pro and Connect plans.
Margin per dish requires food cost loaded or an integrated POS.
It does not forecast demand: that is HIRO Revenue, on the roadmap.
Real product
No makeup. This is how it works inside.
Clips recorded on the real product, with a demo restaurant and synthetic data. In production, this is your house.
Today's brief
Your floor walks into service with the work done.
Bookings, covers, VIPs, birthdays and allergies for the service. Generated automatically every day at 3pm and emailed to the owner.
Segments
Every guest in its segment. Every segment with its play.
Six RFM segments recalculated every night from your bookings, with the suggested action already written.
Guest profile
The whole guest, before you pick up the phone.
Visits, spend, RFM score and what HIRO sees today, with a recommendation cited to the data.
Born in AI
Not a CRM with a chatbot bolted on. Without AI, half of what it does would not exist.
Cited to the data
Every claim links to the data behind it. Arguable, which is what makes it useful.
Before, not after
The report that lands on Monday is already useless. This lands before service.
GDPR, no training
Anthropic with enterprise no-training policies. Your house data does not feed models.
Job 01
Customer intelligence
Six actionable segments updated automatically over the house memory. Every guest walks in with their play already written, before the first chair.
Each segment arrives with its playbook already written.
Job 02
Menu intelligence
Your menu turned into data: what gets ordered with what, which dish over-rotates and which does not return food cost. Requires food cost loaded or an integrated POS.
→ Push the tartare tomorrow. Retire the rice before month-end.
Job 03
Operations intelligence
Service coverage before doors open, dormant account alerts and the three moves of the day on the owner's desk. Information on time, not after service.
The 3 moves of the day, before the first coffee.
Job 04
Ask the house
In beta: ask in plain language what to do with a guest, a table or a dish. Anthropic Claude engine, Voyage embeddings. Your data does not train models.
Every answer with a clickable citation to the underlying data.
HIRO Brain · The Analyst
Every recommendation with a clickable citation to the data. If it cannot be cited, it is not claimed.
What it works with
It plugs into what you already have.
Anthropic Claude
LiveThe reasoning engine behind every employee. A provider with a no-training policy on your data.
Voyage
LiveThe embeddings that let the house memory be searched by meaning instead of by exact word.
CoverManager
LiveReal availability, creation, modification and cancellation of the booking. The employee does not leave a note for someone to key in later, the booking is made.
TPV · Revo
BetaReading the closed ticket to know what was actually consumed, not just who booked.
Actionable segments
Reasoning engine
Embeddings
Data used for training
When it goes on duty
The hospitality year is not flat.
The periods where this employee carries the most weight.
January, after the storm
December left you data. January decides what you do with it.
September and October, the group diary opens
Group season starts long before Christmas.
Quotes from paying customers will sit here once we sign the consent. In the meantime, book a demo and we will show it working.
Coming soon: testimonials from production customers
The vacant post
What happens without it
Generic CRMs that do not speak restaurant
Reports that arrive on Monday, when service is already over
Decisions made by gut because the data sits in five places
Opaque AI assistants that do not show their work
What changes
With it on the team
The owner sees the three moves of the day, not forty charts
Every recommendation can be audited down to the source data
Accounts going cold surface before they are fully cold
The team argues over the same data, not three versions of it
Do you want it on your digital team?
Tell us what your team still does by hand and we will show you how this employee works in a house like yours.
Tell us what your team does by hand