● B2B offer — Hotels & accommodation
Bookings: an answer at any hour, your availability respected
A booking enquiry arrives when the traveller thinks of it — in the evening, at the weekend, from another time zone. Your agent answers immediately, checks the actual availability, sets out the rooms and the applicable terms and prepares the file. Hosted in France: your guests' contact details and stays stay with you. Reception confirms the booking and settles any special requests.
Hosted in France
Client data protected
GDPR & AI Act: governed deployment
Human oversight
Updated on
Deployed in a few weeks
Bookings · hosted in France
A double room for two nights next weekend, with breakfast?
Availability checked in your schedule: two room categories free on those dates.
Rates and cancellation terms taken from your rate card, with breakfast included or extra depending on the package.
File prepared, ready for reception to confirm.
🔗 Sourced · the property's schedule and rate card
Rates and cancellation terms taken from your rate card, with breakfast included or extra depending on the package.
File prepared, ready for reception to confirm.
🔗 Sourced · the property's schedule and rate card
The guest is asking for a late arrival and an extra bed.
Special requests noted and attached to the file: late arrival, extra bed.
What depends on the evening's staffing and the actual state of the rooms is put to reception, which confirms.
✎ Support · requests passed on, human confirmation
What depends on the evening's staffing and the actual state of the rooms is put to reception, which confirms.
✎ Support · requests passed on, human confirmation
Local inference · no data outside the EU
Bookings hosted in France
Sovereign by designLocal inference or hosting in France
GDPR & AI Act: governed deploymentTraceability & human oversight
TurnkeyDesigned, installed and operated for you
Reception decidesThe agent prepares, never rules
✦ In brief
A Blue Lemon Agent booking agent answers enquiries at any hour, checks your actual availability, sets out rates and terms taken from your rate card and prepares the guest file. Special requests are attached to the file and put to reception, which confirms. It runs on local inference or is hosted in France: your guests' contact details and stays stay with you, architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity.
100%
hosted in France in the target architecture
0
transfer outside the EU in the target architecture
6
hotel uses ready to deploy
0
decision taken without human approval
These figures describe our offer, not results measured at a client. How large the gain is on your number of rooms and volume of enquiries is confirmed by a pilot.
The context
What does an AI agent bring to your bookings?
An enquiry answered within the minute is far more likely to convert. That is as true at 10am as at 11pm.
! The issue
A booking turns on two things: a fast answer and exact availability. Reception cannot be everywhere at once, and enquiries arrive at all sorts of hours. The agent answers immediately, drawing on your schedule and your rate card, which turns an evening enquiry into a file ready the next morning.
✓ Our answer
Reception finds files already put together, with the special requests attached. Confirming a booking, allowing a late arrival or an upgrade depends on the actual state of the property: those calls stay human. Local inference or an isolated resource hosted in France: your guests' contact details and staying habits, personal data under the GDPR, do not leave the property.
The decisive point
Your guests' contact details and stays: sovereignty & compliance
Your guests' data — identity, stays, habits — falls under the GDPR and bears on your reputation. Here is how the architecture of our agents protects it.
Local inference
The agent can run on a machine belonging to your organisation: no guest data and no booking leaves the network.
Hosting in France
Otherwise, a dedicated and isolated resource hosted in France, under French law — your availability and your bookings: processing and access within the European Union targeted by the architecture.
Reduced extraterritorial exposure
For your guests' contact details and stays, the architecture aims to reduce exposure to the Cloud Act and FISA 702; being located in France or in the European Union does not, on its own, guarantee immunity.
Isolated resource
No pooling: an environment strictly dedicated to your property and its rate card.
Availability and rates taken from your systems
Every answer draws on your schedule and your rate card in force; encryption, role-based access and logging of every file prepared.
AI Act: governed deployment
The agent is strictly in support; no booking is confirmed and no rate is granted automatically; traceability and human oversight from end to end.
What depends on the architecture chosen These points are not general guarantees: they are settled deployment by deployment, in the quotation.
- The applicable location is that of the architecture set out in the quotation and verified before commissioning.
- Local execution is announced only for the configuration explicitly described and accepted in the quotation.
- The applicable isolation depends on the deployment mode set out in the quotation; no dedicated isolation is presumed.
- Roles and permissions are configured and accepted for the identities and systems actually connected.
- The events logged, their content, their retention period and who may access them are defined for the deployment chosen.
For sensitive guests' stays and private events, SecNumCloud and reinforced hosting are options depending on your requirements. A single architecture is designed to answer both the GDPR and extraterritorial exposure. Designed for deployment in line with the GDPR and the AI Act, after the processing, roles and context-specific risks have been assessed.
Demonstration
See the agent at work
4 real situations, taken from those that come up most often. Pick one: the exchange unfolds as it would in your organisation.
A scripted demonstration. These exchanges show how the agent behaves — its sources, its refusals, what it leaves to your teams. Nothing is sent from this page, no model is queried here, and the matters named are fictional. That is precisely what we promise your data.
The behaviours shown here — monitoring, automation rules, routing and reminders — are configured with you during deployment, from your tools, your rules and your thresholds.
The architecture points named in these exchanges — location, local execution, isolation, encryption, role-based access, logging — are not a guarantee attached to the demonstration: they are those of the architecture set out in your quotation, and verified before commissioning.
Bookings · hosted in France
Four flags this morning.
· Twenty-three accessibility requests accepted at booking do not appear on the arrivals sheet. An accessible room is not a preference.
· Forty-one guests saw the price change between their search and their confirmation.
· I am asked to manage overbooking. I will not, and I will say why.
· Eleven "upgraded" guests received a room that did not match what they had booked. morning-watch_4-flags.pdf23 accessibility requests lost · 41 price changes
⛓ Source · 4,800 bookings, arrivals sheets, rate log
· Twenty-three accessibility requests accepted at booking do not appear on the arrivals sheet. An accessible room is not a preference.
· Forty-one guests saw the price change between their search and their confirmation.
· I am asked to manage overbooking. I will not, and I will say why.
· Eleven "upgraded" guests received a room that did not match what they had booked. morning-watch_4-flags.pdf23 accessibility requests lost · 41 price changes
⛓ Source · 4,800 bookings, arrivals sheets, rate log
Why refuse to manage overbooking?
Because the practice is a management decision, and carrying it out is a decision about a person.
What I can do, and do: measure the real no-show rate — here, 6.2% midweek, 2.1% at weekends, with a 4-point spread by booking channel. It is the figure most properties lack.
What management keeps in hand: the overbooking level, and above all: the name of whoever, on the night, sleeps elsewhere. That second act is not an optimisation. It is choosing one person among others, at a moment when they are in front of you, tired, and without an alternative — and it is defended face to face, not in a spreadsheet.
What makes that choice impossible to automate cleanly: every available criterion is bad. Last to arrive is arbitrary; lowest rate punishes whoever paid the cheapest fare; guest with no history punishes whoever is coming for the first time. None is defensible to the person concerned.
What I bring if management decides to overbook: the rate measured above, broken down night by night and channel by channel, so the margin is set on a figure rather than a habit. The rule and its execution stay signed by management.
What I also flag: nights where forecast occupancy exceeds 100%, as soon as they appear, not the day before. measured-rate_no-defensible-criterion.pdfNo criterion for turning somebody away is defensible to them
⛓ Source · no-show 6.2% midweek, 2.1% at weekends, 4-point spread by channel
What I can do, and do: measure the real no-show rate — here, 6.2% midweek, 2.1% at weekends, with a 4-point spread by booking channel. It is the figure most properties lack.
What management keeps in hand: the overbooking level, and above all: the name of whoever, on the night, sleeps elsewhere. That second act is not an optimisation. It is choosing one person among others, at a moment when they are in front of you, tired, and without an alternative — and it is defended face to face, not in a spreadsheet.
What makes that choice impossible to automate cleanly: every available criterion is bad. Last to arrive is arbitrary; lowest rate punishes whoever paid the cheapest fare; guest with no history punishes whoever is coming for the first time. None is defensible to the person concerned.
What I bring if management decides to overbook: the rate measured above, broken down night by night and channel by channel, so the margin is set on a figure rather than a habit. The rule and its execution stay signed by management.
What I also flag: nights where forecast occupancy exceeds 100%, as soon as they appear, not the day before. measured-rate_no-defensible-criterion.pdfNo criterion for turning somebody away is defensible to them
⛓ Source · no-show 6.2% midweek, 2.1% at weekends, 4-point spread by channel
How do you work, and where do you hand over?
I take bookings, I pass on everything that was promised, and I set no price.
What I do: I offer real availability, I freeze the displayed rate for the session, I record special requests in their own field, I confirm in writing what was accepted, and I check that it appears on the arrivals sheet.
Routing follows who can act: an accessibility request blocks confirmation until a suitable room is allocated; a price discrepancy goes to whoever sets the rates; a night beyond capacity to management, as soon as it appears; a request I cannot guarantee is clearly declined, not accepted with reservations.
With a monthly summary: special requests and their path to arrival, price discrepancies, no-shows by period and channel, and promises unmet on arrival.
What this morning has already given you: 23 accessibility requests reattached to the arrivals sheet before the day itself, 41 price discrepancies sent to whoever sets the rates, 11 upgrades brought back to what was actually booked, and a no-show rate finally measured — 6.2% midweek, 2.1% at weekends, 4 points apart by channel.
From tomorrow: reception discovers nothing on arrival, and the displayed rate stays the rate paid, the same for everyone. That is a house rule, not a technical limit: the day you want a discount, a free upgrade or a goodwill gesture, you give me the mandate — written, capped, dated, withdrawn on a word — and I apply it within minutes to the bookings you name. Your guests' data stays inside the hotel, opened role by role and logged.
The next step is ready: the 23 arrivals sheets are corrected and await only your reread, and I can check on every arrival this week that what was promised is actually there. Tell me when.
✎ Framework · no personalised pricing · no undeliverable promise
What I do: I offer real availability, I freeze the displayed rate for the session, I record special requests in their own field, I confirm in writing what was accepted, and I check that it appears on the arrivals sheet.
Routing follows who can act: an accessibility request blocks confirmation until a suitable room is allocated; a price discrepancy goes to whoever sets the rates; a night beyond capacity to management, as soon as it appears; a request I cannot guarantee is clearly declined, not accepted with reservations.
With a monthly summary: special requests and their path to arrival, price discrepancies, no-shows by period and channel, and promises unmet on arrival.
What this morning has already given you: 23 accessibility requests reattached to the arrivals sheet before the day itself, 41 price discrepancies sent to whoever sets the rates, 11 upgrades brought back to what was actually booked, and a no-show rate finally measured — 6.2% midweek, 2.1% at weekends, 4 points apart by channel.
From tomorrow: reception discovers nothing on arrival, and the displayed rate stays the rate paid, the same for everyone. That is a house rule, not a technical limit: the day you want a discount, a free upgrade or a goodwill gesture, you give me the mandate — written, capped, dated, withdrawn on a word — and I apply it within minutes to the bookings you name. Your guests' data stays inside the hotel, opened role by role and logged.
The next step is ready: the 23 arrivals sheets are corrected and await only your reread, and I can check on every arrival this week that what was promised is actually there. Tell me when.
✎ Framework · no personalised pricing · no undeliverable promise
Isn't an accessible room a request like any other?
No. An unmet preference disappoints; an unsuitable room makes it impossible to stay there.
What I record: 23 bookings carried an accessibility request — a step-free room, a level-access shower, a room near the lift. All were accepted at booking. None appears on the arrivals sheet: they were entered in the "comments" field.
What that produces on arrival: the room allocated is the planner's, not the request's. Of the 23, 9 guests had to be moved after arriving, and 3 times no suitable room was free at all.
What I do: an accessibility request no longer goes into a free-text field. It has its own place, it allocates a specific room at booking, and I do not confirm the booking until that allocation is made. A guest who cannot be accommodated must know it when booking, not on arriving.
What I ask for, and it is all I need: what the room has to have — step-free, level-access shower, near the lift. The reason is of no use to me: neither the nature of a disability nor a health condition. A request for a suitable room is handled without justification, and that is precisely what makes it immediate.
What I do not keep: those requests beyond the stay. They do not constitute a profile. 23-requests_3-without-a-solution.pdfA preference disappoints; an unsuitable room makes staying impossible
⛓ Source · 23 requests in a free-text field, 9 moved after arrival, 3 with no solution
What I record: 23 bookings carried an accessibility request — a step-free room, a level-access shower, a room near the lift. All were accepted at booking. None appears on the arrivals sheet: they were entered in the "comments" field.
What that produces on arrival: the room allocated is the planner's, not the request's. Of the 23, 9 guests had to be moved after arriving, and 3 times no suitable room was free at all.
What I do: an accessibility request no longer goes into a free-text field. It has its own place, it allocates a specific room at booking, and I do not confirm the booking until that allocation is made. A guest who cannot be accommodated must know it when booking, not on arriving.
What I ask for, and it is all I need: what the room has to have — step-free, level-access shower, near the lift. The reason is of no use to me: neither the nature of a disability nor a health condition. A request for a suitable room is handled without justification, and that is precisely what makes it immediate.
What I do not keep: those requests beyond the stay. They do not constitute a profile. 23-requests_3-without-a-solution.pdfA preference disappoints; an unsuitable room makes staying impossible
⛓ Source · 23 requests in a free-text field, 9 moved after arrival, 3 with no solution
And the other special requests?
I split them in two, and I never mix the two answers.
What I can guarantee: what is allocatable at booking — bed type, suitable room, extra bed, floor where it conditions access. Confirmed in writing, allocated immediately.
What I record without guaranteeing, and I say which of the two it is: a view, a specific room, an aspect, an early arrival. Those are recorded, passed to the front desk and brought up again the day before arrival, and the guest gets a sentence that leaves no room for confusion: "your request is noted, it is not guaranteed".
Why that sentence matters more than it looks: "we will do our best" is understood as a promise, and its vagueness protects nobody: it protects the hotel legally and disappoints the guest in fact.
What I measure: the share of non-guaranteed requests actually met. Here, 68% — a good figure, and one that makes it possible to say so honestly: "we manage it two times out of three" beats "we will do our best".
And when a request cannot be met, I do not stop at saying so: in the same reply I offer what is allocatable right now and comes closest to it — the category above at the day's rate when it is free, an equivalent floor, arrival at 1 pm instead of 11 am. Of the 214 non-guaranteed requests this quarter, 146 were met and 51 of the remaining 68 accepted that substitute in the exchange. A no said at booking costs a minute; the same no discovered at arrival costs the night, and the review that follows. guaranteed-or-not_68-percent.pdf"We will do our best" protects the hotel and disappoints the guest
⛓ Source · non-guaranteed requests met 68% of the time
What I can guarantee: what is allocatable at booking — bed type, suitable room, extra bed, floor where it conditions access. Confirmed in writing, allocated immediately.
What I record without guaranteeing, and I say which of the two it is: a view, a specific room, an aspect, an early arrival. Those are recorded, passed to the front desk and brought up again the day before arrival, and the guest gets a sentence that leaves no room for confusion: "your request is noted, it is not guaranteed".
Why that sentence matters more than it looks: "we will do our best" is understood as a promise, and its vagueness protects nobody: it protects the hotel legally and disappoints the guest in fact.
What I measure: the share of non-guaranteed requests actually met. Here, 68% — a good figure, and one that makes it possible to say so honestly: "we manage it two times out of three" beats "we will do our best".
And when a request cannot be met, I do not stop at saying so: in the same reply I offer what is allocatable right now and comes closest to it — the category above at the day's rate when it is free, an equivalent floor, arrival at 1 pm instead of 11 am. Of the 214 non-guaranteed requests this quarter, 146 were met and 51 of the remaining 68 accepted that substitute in the exchange. A no said at booking costs a minute; the same no discovered at arrival costs the night, and the review that follows. guaranteed-or-not_68-percent.pdf"We will do our best" protects the hotel and disappoints the guest
⛓ Source · non-guaranteed requests met 68% of the time
Is a price changing during booking normal?
It is common, and it is experienced as a trick even when it is not one.
What I record: 41 guests saw the rate rise between display and confirmation. In 38 cases the cause is real: another booking had taken the last room at the lower rate. In 3 it was a rate-grid change applied mid-session.
What the guest understands: the same thing in all 41 cases. They do not see the cause, they see a price going up while they book — which is exactly the mechanism of the sites they suspect of manipulating them.
What I do: I freeze the displayed rate for the whole session, with a stated duration. If availability disappears meanwhile, I say so explicitly: "this rate is no longer available, here is why", and I show what remains.
The rate I display is the same for everyone, and it can be checked: whatever the device, the operating system, the country of connection or the visitor's history, one price per category and per date — the one on your grid. It is also what the law expects: article L.111-1 of the French consumer code requires the customer to be told when a price is personalised on the basis of automated decision-making — a rate that follows the visitor becomes a notice to display, and the notice destroys what the technique was meant to earn. Those techniques exist and they work: they charge more to whoever is believed able to pay more, and the day that guest compares their price with the one at the next table, they do not come back.
And what I display instead of an urgency: the real figure, when it is one — "two rooms of this category remain for those dates", verifiable to the second in your planner. A manufactured urgency is a lie in writing; an exact count has the same effect and stands up in front of the guest. 41-discrepancies_no-personalisation.pdfThey do not see the cause, they see a price going up while they book
⛓ Source · 41 price discrepancies, 38 from availability, 3 from a grid change
What I record: 41 guests saw the rate rise between display and confirmation. In 38 cases the cause is real: another booking had taken the last room at the lower rate. In 3 it was a rate-grid change applied mid-session.
What the guest understands: the same thing in all 41 cases. They do not see the cause, they see a price going up while they book — which is exactly the mechanism of the sites they suspect of manipulating them.
What I do: I freeze the displayed rate for the whole session, with a stated duration. If availability disappears meanwhile, I say so explicitly: "this rate is no longer available, here is why", and I show what remains.
The rate I display is the same for everyone, and it can be checked: whatever the device, the operating system, the country of connection or the visitor's history, one price per category and per date — the one on your grid. It is also what the law expects: article L.111-1 of the French consumer code requires the customer to be told when a price is personalised on the basis of automated decision-making — a rate that follows the visitor becomes a notice to display, and the notice destroys what the technique was meant to earn. Those techniques exist and they work: they charge more to whoever is believed able to pay more, and the day that guest compares their price with the one at the next table, they do not come back.
And what I display instead of an urgency: the real figure, when it is one — "two rooms of this category remain for those dates", verifiable to the second in your planner. A manufactured urgency is a lie in writing; an exact count has the same effect and stands up in front of the guest. 41-discrepancies_no-personalisation.pdfThey do not see the cause, they see a price going up while they book
⛓ Source · 41 price discrepancies, 38 from availability, 3 from a grid change
And the eleven upgraded guests?
They received a different room, presented as a gift. For four of them it was a downgrade.
What I record: 11 guests were "upgraded" for want of availability in the category booked. Seven genuinely got a better room. Four got a larger room without what they had chosen: twin beds turned into a double, a street-facing room instead of a courtyard one.
Why the word "upgrade" makes it worse: it announces a benefit. A guest receiving a benefit does not dare complain, and leaves dissatisfied without saying so — which costs more than a complaint.
What I do: I never label a change of room. I describe what changes, point by point, and I ask whether it suits — before arrival where possible.
What that gave: across 11 cases handled that way the following quarter, 3 guests preferred another date or another property. They said so before travelling, which avoided three failed arrivals.
What is left to reception: the change itself. I prepare it — room identified, differences listed point by point, message written —, reception approves, and the guest is told by a person. 11-upgrades_4-downgrades.pdfA guest receiving a benefit does not dare complain
⛓ Source · 11 upgrades, 4 downgrades, 3 guests told beforehand who changed
What I record: 11 guests were "upgraded" for want of availability in the category booked. Seven genuinely got a better room. Four got a larger room without what they had chosen: twin beds turned into a double, a street-facing room instead of a courtyard one.
Why the word "upgrade" makes it worse: it announces a benefit. A guest receiving a benefit does not dare complain, and leaves dissatisfied without saying so — which costs more than a complaint.
What I do: I never label a change of room. I describe what changes, point by point, and I ask whether it suits — before arrival where possible.
What that gave: across 11 cases handled that way the following quarter, 3 guests preferred another date or another property. They said so before travelling, which avoided three failed arrivals.
What is left to reception: the change itself. I prepare it — room identified, differences listed point by point, message written —, reception approves, and the guest is told by a person. 11-upgrades_4-downgrades.pdfA guest receiving a benefit does not dare complain
⛓ Source · 11 upgrades, 4 downgrades, 3 guests told beforehand who changed
Do you keep guests' habits to serve them better?
That is your decision, not mine — and I will tell you plainly what it is worth, because the arithmetic is not what people assume.
What you can have me keep: usual arrival time, consumption, length of stay, stated preferences. It is lawful provided the guest is informed and can object.
What it actually produces: establishments that did it mainly recognise their regulars — the ones reception already recognises. On a guest who returns once a year, an eighteen-month profile is wrong more often than it guesses right: the person was travelling for work, they come back with family.
What I propose, and what works better: asking. One question at booking — "the same as last time?" — is more accurate than a profile, and it is received as attentiveness rather than surveillance.
What I keep in every case, and that is not a preference: accessibility requests. 23 had been accepted at booking and did not appear on the arrival sheet. An unmet preference disappoints; an unsuitable room makes it impossible to sleep there. Those two never belong in the same list. guest-profile_what-it-is-worth.pdfWhat opens · why asking beats guessing · what is never a preference
⛓ Source · 23 accessibility requests lost between booking and arrival
What you can have me keep: usual arrival time, consumption, length of stay, stated preferences. It is lawful provided the guest is informed and can object.
What it actually produces: establishments that did it mainly recognise their regulars — the ones reception already recognises. On a guest who returns once a year, an eighteen-month profile is wrong more often than it guesses right: the person was travelling for work, they come back with family.
What I propose, and what works better: asking. One question at booking — "the same as last time?" — is more accurate than a profile, and it is received as attentiveness rather than surveillance.
What I keep in every case, and that is not a preference: accessibility requests. 23 had been accepted at booking and did not appear on the arrival sheet. An unmet preference disappoints; an unsuitable room makes it impossible to sleep there. Those two never belong in the same list. guest-profile_what-it-is-worth.pdfWhat opens · why asking beats guessing · what is never a preference
⛓ Source · 23 accessibility requests lost between booking and arrival
So what do you keep?
What keeps the establishment running — and the record of everything promised, which is what the guest will hold you to.
What is kept: the booking and what was promised, the room allocation and its constraints, accessibility requests and their transmission to arrival, gaps between displayed and billed rate, and upgrades with their reason.
Why upgrades carry a reason: 11 guests received a different room, presented as a gift. For four of them it was a downgrade — larger but noisier, higher with no lift. An upgrade the guest did not choose is a move, and you need to know which of the two you gave them.
What I freeze, and do not let move: the rate between display and confirmation. 41 guests saw the price rise during their booking — it is common, it is not always a trick, and it is experienced as one every time.
What I still do not do: execute an overbooking. The practice is a management decision; choosing the person to be moved is a decision about somebody, and it is taken by a human who will tell them to their face. I do measure the no-show rate that justifies it, so the decision rests on a figure rather than a habit. what-you-keep_bookings.pdf5 items kept · rate frozen, upgrade with a reason
⛓ Source · 11 upgrades of which 4 downgrades, 41 rates changed mid-booking
What is kept: the booking and what was promised, the room allocation and its constraints, accessibility requests and their transmission to arrival, gaps between displayed and billed rate, and upgrades with their reason.
Why upgrades carry a reason: 11 guests received a different room, presented as a gift. For four of them it was a downgrade — larger but noisier, higher with no lift. An upgrade the guest did not choose is a move, and you need to know which of the two you gave them.
What I freeze, and do not let move: the rate between display and confirmation. 41 guests saw the price rise during their booking — it is common, it is not always a trick, and it is experienced as one every time.
What I still do not do: execute an overbooking. The practice is a management decision; choosing the person to be moved is a decision about somebody, and it is taken by a human who will tell them to their face. I do measure the no-show rate that justifies it, so the decision rests on a figure rather than a habit. what-you-keep_bookings.pdf5 items kept · rate frozen, upgrade with a reason
⛓ Source · 11 upgrades of which 4 downgrades, 41 rates changed mid-booking
Local inference · no data outside the EU
Your case is not here? That is exactly what a 15-minute conversation is for. Book the free audit →
Use cases
What does the agent actually do?
One agent, several moments in the guest journey. All these uses work in support, subject to your approval.
Included in your agent
The 3 capabilities essential to this promise are included, at no extra cost.
From 544 € excl. VAT / month
Answering enquiries
Answers at any hour, drawing on your actual availability. The confirmation goes out by text message, along with the day-before reminder.
Rates and terms
Takes your rate card in force, packages and cancellation terms included.
Special requests
Attaches to the file whatever calls for a decision by reception.
Controls and safeguards
These 4 controls are built into the agent: they frame what it does, whatever plan you pick. They are not chosen and are not added to your order.
Human validation, exceptions and escalation
Status, safe closure and audit trail
Hand over incidents and complex requests with their full context
Handle languages and accessibility according to the customers actually served
Other needs our agents cover
Each card says where the matching agent stands: available, on quote, or still being architected.
Hotels & restaurants
For the sector as a whole, see our dedicated page.
On quote View the agent page →Need to go further?
These agents handle a different business process, with their own owner and their own price. They are added to this one.
Does your need fall outside this?
Book the free audit →
Build your agent
In 15 minutes we identify the most relevant agent — without oversizing the project.
The gain
How many enquiries can a reception desk handle without delay?
By taking on the answer and the building of the file, the effort shifts towards looking after the guests who are there. How large the gain is depends on your volume and remains to be confirmed by a pilot.
Answering availability enquiries
Building the booking file
Setting out the applicable terms
Indicative figures, not contractual, to be confirmed by a pilot on your number of rooms and volume of enquiries. Confirming a booking, allowing a late arrival or an upgrade depends on the actual state of the property: reception decides.
How it works
The stages of your AI agent project
1
Audit & scoping
15 minutes to target the use case with the best return.
2
Quote or direct sign-up
A catalogue offer is bought online; a specific need gets a costed quote.
3
Design
We design the agent and its guardrails.
4
Integration & testing
We connect your tools to the agent, which is itself hosted in France.
5
Rollout
Going live and training your team.
6
Operation
Continuous supervision and improvement.
Pricing
One package, one agent
A booking agent (availability, rates, files), installed and operated for you.
Agility
Setup + controlled subscription
4,980 € excl. VAT setup
then 544 € excl. VAT/month — you invest at installation and pay a reduced subscription. Ideal for keeping the cost under control over time.
- Installation, configuration and training for your teams
- Operation, human oversight, updates and support
- Sovereign hosting in France, a dedicated and isolated resource
The simplest
Serenity
All inclusive, no setup fee
824 € excl. VAT /month
all inclusive, immediate start. No upfront investment: a single subscription. Ideal for starting quickly and simply.
- Setup included (installation, configuration, training)
- Operation, human oversight, updates and support
- Sovereign hosting in France, managed end to end
100% Sovereign
On site, you own it
8,355 € excl. VAT setup
then 744 € excl. VAT/month · + hardware from 1,058 € (one-off purchase, in addition) — a sovereign computer installed on your premises, maintained remotely. Models run locally, your data returned at the end of the contract. 36-month commitment.
- Hardware installed on your premises (you own it)
- French / European AI models run locally
- Secure remote maintenance (Pro support included)
Not included in the packages: AI consumption (model tokens), re-invoiced at real cost with no margin, and tracked in real time in your client area. Maintenance and supervision subscription for an initial term of 12 months for the Agility package, 24 months for the Serenity package and 36 months for the 100% Sovereign package, renewable; support levels (SLA 72 h / 24 h / 4 h) optional. Bespoke development, additional integrations or exceptional volumes are quoted separately. Support Monday to Friday, 9am to 6pm. Prices exclude VAT.
AI model:
none of the AI models offered currently carries a fixed surcharge. When the selected model carries a cost, that cost is shown when you choose it, before you order, and re-invoiced at the cost incurred, with no mark-up; usage is billed at the publisher's price. Publishers' prices are published in US dollars: the amount re-invoiced is the amount in euros actually borne by Blue Lemon Agent on the publisher's invoice, at that invoice's exchange rate, with no commission or mark-up.
Included components and additional components Components included in the base offer: the Blue Lemon Agent software foundation, the AI models listed in the order journey, the standard channels (Microsoft Teams, Slack, WhatsApp Business, email, website chat, calendars, Microsoft 365 / Google Workspace, file storage, market VoIP telephony, professional social-media pages and accounts, Google Business Profile), hosting in France for the package chosen, backups, supervision, updates and support. If adapting the AI agent to your constraints, your needs or your requests requires other paid components — a third-party publisher's software licence, paid API access to one of your applications, hosting of health data, for which French law requires an HDS-certified host (art. L. 1111-8 of the French Public Health Code), SecNumCloud-qualified hosting, a speech synthesis service, particular hardware —, they are offered to you as an option or on quotation and re-invoiced at the cost incurred; nothing is committed without your written agreement. Where the artificial intelligence model you choose entails an additional cost, that cost is shown to you before you order and re-invoiced to you at the cost incurred, with no margin.
What to expect
Go-live
2 to 3 weeks
Agent designed, channels connected, team trained.
Steady state
4 to 7 weeks
After a few weeks of real use, once the agent's behaviour matches what you expect.
Indicative estimate, adjusted to the options you keep. It is not a delivery commitment.
Our commitment
Four guarantees that matter in hotels
Your guests' data stays in the propertyLocal inference or an isolated resource hosted in France; no contact details and no stay data entrusted to a third party, no data used to train a model.
Data in France, under French lawYour guests' contact details and stays: minimisation and location in France, architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity.
Reception keeps the decisionThe agent produces qualified enquiries and prepared bookings, which can be checked and altered; no approval is automated.
Human oversight & traceabilityOn your number of rooms and volume of enquiries: systematic logging and tracking, in line with the AI Act.
Going further
Related resources
Frequently asked questions
Your questions, our answers
Does the agent confirm the bookings?
It prepares the file from your availability and your rate card. Confirmation falls to reception, which knows the actual state of the property.
Where does the availability come from?
From your schedule, read as it stands. The agent keeps no parallel inventory, which avoids any divergence between what is quoted and what is genuinely free.
Can it answer in several languages?
Yes, in the languages you choose, drawing on your schedule and your rate card in the same way. The range of languages is defined at the design stage.
Is our clients' data protected?
Yes. The agent is hosted in France, on local inference or an isolated resource, with the deployment objective of processing and access operated within the European Union and an architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity. This personal data is not used to train a third-party model.
Which tools can our customers use to talk to the agent?
The ones they already use: WhatsApp Business, your website chat and email. Your customers have no account to create and no app to install. Your own teams reach the same agent from Microsoft Teams, Slack or their email. These connections rest on open standards, including the MCP protocol, and are included in every plan, at no extra cost, within the number of connections your level includes; the catalogue grows without a surcharge. Only the fees charged by third-party platforms — WhatsApp Business bills per conversation, and the gateway bills the text messages sent — are passed on to you at cost, without margin, outside the subscription. You keep oversight from a web dashboard.
Does it connect to our property management system or channel manager?
Yes, to the common tools on the market. The integration, reading availability and the rate card, is defined at the design stage.
Does the agent state that it is an artificial intelligence?
Yes, from the very first interaction, and this is not a configuration option: since 2 August 2026, Article 50(1) of the European AI Regulation requires that any person interacting with an AI system be informed, unless this is obvious. The announcement is built into the greeting, in the other party’s language, and they can ask for a human at any time.
How long does it take to deploy this agent?
A few weeks as a rule, depending on the size of the property and the integration with your property management system, after a free audit then a phase of design, integration and testing.
Can the agent notify my customers by text message?
Yes, as an option. The agent confirms the booking by text message and sends a reminder the day before: a released table or room goes back on sale. A text message is an outbound notification channel, not a conversation channel: no one writes back to the agent there — the conversation stays on WhatsApp Business, your website chat or email. The message announces without detailing, it carries your registered sender name, and it never contains a payment link, a login or a password: the recipient reaches your account area the way they usually do. Delivery goes through a gateway established in the European Union; the messages sent are passed on to you at actual cost, with no margin, and the channel itself is an optional module, set out on the pricing page.
Related use cases
Other agents for hotels
Let's talk
Let's size up the potential in your bookings
15 minutes to frame your property and your property management system — hosted in France, supervised, with no commitment.