● B2B offer — Property & lettings management
Tenant relations: every request handled, every tenancy respected
A tenant's request arrives by email, by phone, through the portal — and each one calls for an answer that takes the tenancy agreement into account. Your agent qualifies the request, finds the applicable clauses, prepares the reply and flags what calls for a management decision. Hosted in France: your tenants' contact details and circumstances stay with you. The property manager approves every reply sent.
Hosted in France
Tenant data protected
GDPR & AI Act: governed deployment
Human oversight
Updated on
Deployed in a few weeks
Tenant relations · hosted in France
A tenant reports a broken boiler. What does their tenancy agreement say?
Request qualified: equipment maintenance, property identified, tenancy agreement found.
The agreement sets out the allocation of maintenance costs — the clause is cited and referenced.
Reply prepared with a note of the usual response times, ready to read over.
🔗 Sourced · clause cited from the tenancy agreement
The agreement sets out the allocation of maintenance costs — the clause is cited and referenced.
Reply prepared with a note of the usual response times, ready to read over.
🔗 Sourced · clause cited from the tenancy agreement
And if the tenant asks for a rent reduction?
I gather the material: dates of the report, nature of the equipment, clauses of the agreement and the applicable framework.
What follows commits your company and the contract: it belongs to the manager, and the agent puts the file together so the decision can be quick and documented.
✎ Support · file gathered, management decision
What follows commits your company and the contract: it belongs to the manager, and the agent puts the file together so the decision can be quick and documented.
✎ Support · file gathered, management decision
Local inference · no data outside the EU
Tenancy agreements hosted in France
Sovereign by designLocal inference or hosting in France
GDPR & AI Act: governed deploymentTraceability & human oversight
TurnkeyDesigned, installed and operated for you
The officer decidesThe agent prepares, never rules
✦ In brief
A Blue Lemon Agent tenant relations agent qualifies the requests received, finds the clauses of the tenancy agreement concerned and prepares the reply, citing its source. It tells apart everyday replies from what calls for a management decision, and leaves the latter to the manager. It runs on local inference or is hosted in France: your tenants' contact details and circumstances stay with you, architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity. Your teams write to it from Microsoft Teams, Slack or their email, and your tenants reach it on WhatsApp Business, your website chat or email — with no account to create and nothing to install. These connections are included in every plan, at no extra cost, within the number of connections your level includes.
100%
hosted in France in the target architecture
0
transfer outside the EU in the target architecture
6
lettings management 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 the size of your portfolio of units is confirmed by a pilot.
The context
What does an AI agent bring to your tenant relations?
A quick, accurate answer grounded in the right agreement is what makes lettings management feel well run. It is also what takes the most time.
! The issue
Every request calls for two things: the right clause of the right agreement and a clear reply. Finding the first across a portfolio of several hundred units takes up most of the time; writing the second takes up the rest. The agent takes on the search and the drafting, systematically citing the clause it relies on.
✓ Our answer
The manager receives a qualified request, the applicable clauses cited and a reply ready to read over. Whatever commits the company — a goodwill gesture, works, litigation — is flagged as a management decision. Local inference or an isolated resource hosted in France: your tenants' contact details and circumstances, personal data under the GDPR, are entrusted to no third party.
The decisive point
Your tenants' contact details and circumstances: sovereignty & compliance
Your tenants' data falls under the GDPR and their personal circumstances call for particular discretion. Here is how the architecture of our agents protects them.
Local inference
The agent can run on a machine belonging to your organisation: no tenancy agreement and no tenant data leaves the network.
Hosting in France
Otherwise, a dedicated and isolated resource hosted in France, under French law — your agreements and your tenant requests: processing and access within the European Union targeted by the architecture.
Reduced extraterritorial exposure
For your tenants' contact details and circumstances, 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 practice and its portfolio.
A clause cited with every reply
Every reply prepared refers back to the clause of the agreement it rests on; encryption, role-based access and logging of every tenant file consulted.
AI Act: governed deployment
The agent is strictly in support; no reply is sent and no management decision is taken 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 files involving arrears, litigation or social vulnerability, 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.
Tenant relations · hosted in France
Four flags this morning across your portfolio.
· Forty-one per cent of requests come from eight per cent of the homes. That is not eight per cent of demanding tenants.
· Twenty-three homes have raised no request in two years. That is not necessarily good news.
· One hundred and four jobs are marked done without anybody asking the tenant whether the problem was.
· A November request mentioned a window that closes badly. It was filed under "joinery". morning-watch_4-flags.pdf41% of requests from 8% of homes · 23 silent homes
⛓ Source · 1,480 homes, 6,200 requests, works log
· Forty-one per cent of requests come from eight per cent of the homes. That is not eight per cent of demanding tenants.
· Twenty-three homes have raised no request in two years. That is not necessarily good news.
· One hundred and four jobs are marked done without anybody asking the tenant whether the problem was.
· A November request mentioned a window that closes badly. It was filed under "joinery". morning-watch_4-flags.pdf41% of requests from 8% of homes · 23 silent homes
⛓ Source · 1,480 homes, 6,200 requests, works log
Eight per cent of the homes — what does that mean exactly?
That the problem is in the home, and it was being looked for in the occupant.
What I record: 118 homes out of 1,480 account for 41% of requests. Of those 118, 73 changed occupant at least once in the period — and the volume of requests did not fall at the changeover.
What that establishes: the factor is not the person. When two successive occupants report the same family of problems, what stays constant is the home.
Why it was invisible: because requests are filed by requester and by trade, never by property. A home behind nineteen requests in three years appears nowhere as such: it appears as nineteen cases, each closed within the deadline.
What I do: I group by home, then by building, and I flag those exceeding their family — same floor area, same year of construction, same heating.
What that produced: of the 118, 29 come from a single design defect in two buildings. Handled one by one, those requests cost more than the rework, and they came back. 118-homes_29-one-defect.pdfWhat stays constant when the occupant changes is the home
⛓ Source · 118 homes, 41% of requests, 73 with a change of occupant
What I record: 118 homes out of 1,480 account for 41% of requests. Of those 118, 73 changed occupant at least once in the period — and the volume of requests did not fall at the changeover.
What that establishes: the factor is not the person. When two successive occupants report the same family of problems, what stays constant is the home.
Why it was invisible: because requests are filed by requester and by trade, never by property. A home behind nineteen requests in three years appears nowhere as such: it appears as nineteen cases, each closed within the deadline.
What I do: I group by home, then by building, and I flag those exceeding their family — same floor area, same year of construction, same heating.
What that produced: of the 118, 29 come from a single design defect in two buildings. Handled one by one, those requests cost more than the rework, and they came back. 118-homes_29-one-defect.pdfWhat stays constant when the occupant changes is the home
⛓ Source · 118 homes, 41% of requests, 73 with a change of occupant
How do you work, and where do you hand over?
I receive requests, qualify them, route them, and I keep nothing that could be turned against a tenant.
What I do: I acknowledge receipt with a real timescale, not a standard one, qualify the subject and the urgency, route to the right person, and come back to the tenant after the works to ask whether it is resolved.
Routing follows who can act: a safety emergency goes straight through, bypassing the queue; a home with repeated requests to the technical manager, grouped; a defect shared by several homes to whoever decides on works; a request outside the landlord's remit is redirected, and said so.
With a monthly summary: requests per home and per building, works with no tenant follow-up, real timescales by type, and homes with no request for a long time.
What this analysis has already given you: 118 homes out of 1,480 identified as the source of 41% of your requests, 29 of them traced to a single design defect in two buildings, 104 works marked done taken back to the tenant, and 23 homes silent for two years placed under check.
From tomorrow: one grouped repair instead of twenty-nine call-outs that kept coming back — cheaper than the sum of the visits, and the problem stops returning. Nothing a tenant reports goes into their file: not the number of their requests, not their tone, not what they say about their neighbours — and that is exactly what makes the figure usable, because I count per home, and it is the home that gets repaired. The timescales I give are the ones the technical team has confirmed, and the day you want me to book a slot myself, you give me the mandate: bounded to the types of work you set, dated, withdrawn on a word.
The next step is ready: the 118 homes are ranked by how far they sit from their peer group — same floor area, same build year, same heating —, and the two buildings concerned have their repair note. Tell me when to put it to whoever decides on works.
✎ Framework · nothing reaches the tenant's file · no timescale committed
What I do: I acknowledge receipt with a real timescale, not a standard one, qualify the subject and the urgency, route to the right person, and come back to the tenant after the works to ask whether it is resolved.
Routing follows who can act: a safety emergency goes straight through, bypassing the queue; a home with repeated requests to the technical manager, grouped; a defect shared by several homes to whoever decides on works; a request outside the landlord's remit is redirected, and said so.
With a monthly summary: requests per home and per building, works with no tenant follow-up, real timescales by type, and homes with no request for a long time.
What this analysis has already given you: 118 homes out of 1,480 identified as the source of 41% of your requests, 29 of them traced to a single design defect in two buildings, 104 works marked done taken back to the tenant, and 23 homes silent for two years placed under check.
From tomorrow: one grouped repair instead of twenty-nine call-outs that kept coming back — cheaper than the sum of the visits, and the problem stops returning. Nothing a tenant reports goes into their file: not the number of their requests, not their tone, not what they say about their neighbours — and that is exactly what makes the figure usable, because I count per home, and it is the home that gets repaired. The timescales I give are the ones the technical team has confirmed, and the day you want me to book a slot myself, you give me the mandate: bounded to the types of work you set, dated, withdrawn on a word.
The next step is ready: the 118 homes are ranked by how far they sit from their peer group — same floor area, same build year, same heating —, and the two buildings concerned have their repair note. Tell me when to put it to whoever decides on works.
✎ Framework · nothing reaches the tenant's file · no timescale committed
Twenty-three homes with no request — is that a good sign?
It is the signal I watch most closely, and the only one no dashboard produces.
What I record: 23 homes have raised no request in two years. Across a comparable portfolio the median is 2.1 requests per home per year.
What it may mean, and I do not settle it: the home is in good condition and the occupant satisfied; or the occupant does not know how to ask; or they stopped asking because an earlier request came to nothing; or they fear that asking will be held against them.
Why that last hypothesis must be named: because a tenant cannot leave the way a customer changes supplier. A dissatisfied customer's silence shows up as a departure; a tenant's shows up as nothing at all. That is what makes an absence of complaints uninterpretable.
What I do: I send the list to whoever manages the area, with no conclusion, and I contact nobody myself. An approach prompted by silence can be experienced as surveillance.
What that produced: of the 23, a courtesy visit was offered to those whose home had not been visited in over three years. Eleven accepted, and seven unreported problems came to light. 23-homes_7-unreported-problems.pdfA customer’s silence is a departure; a tenant’s is nothing
⛓ Source · 23 silent homes, median 2.1 requests/year, 7 problems found
What I record: 23 homes have raised no request in two years. Across a comparable portfolio the median is 2.1 requests per home per year.
What it may mean, and I do not settle it: the home is in good condition and the occupant satisfied; or the occupant does not know how to ask; or they stopped asking because an earlier request came to nothing; or they fear that asking will be held against them.
Why that last hypothesis must be named: because a tenant cannot leave the way a customer changes supplier. A dissatisfied customer's silence shows up as a departure; a tenant's shows up as nothing at all. That is what makes an absence of complaints uninterpretable.
What I do: I send the list to whoever manages the area, with no conclusion, and I contact nobody myself. An approach prompted by silence can be experienced as surveillance.
What that produced: of the 23, a courtesy visit was offered to those whose home had not been visited in over three years. Eleven accepted, and seven unreported problems came to light. 23-homes_7-unreported-problems.pdfA customer’s silence is a departure; a tenant’s is nothing
⛓ Source · 23 silent homes, median 2.1 requests/year, 7 problems found
And the hundred and four jobs with no follow-up?
They are closed by whoever did the work, never by whoever had the problem.
What I record: 104 jobs are marked "completed". The tenant was contacted in none of them. That is the normal arrangement: the contractor declares it done, the case closes.
What it hides: a job can be done and not resolve anything. Across a sample of 40 of those cases contacted afterwards, 9 problems persisted — and in 6 the tenant had not raised a new request, taking it that the work was done and that it would be the same again.
What I do: I ask one question, only one, a few days after the works: "is it resolved?". No satisfaction scale, no score. A score measures a mood; a question measures the problem.
What I do with the answer: if it is no, I reopen the original case rather than create a new one — otherwise the home shows as having two requests instead of one unresolved problem.
What I do with reopenings, and it is the opposite of a scoreboard: I attach them to the home first, because some come back when the cause lies elsewhere — water tracking through a floor is not bad workmanship. So I bring back, contractor by contractor, the reopenings whose cause really is the same, with the home, the date, the reason and the original job. That file is discussed and signed: a remark to a contractor engages a contractual relationship, and it is yours to make. Across the 40 cases checked, 9 problems persisted: all nine are already worked up and ready to reopen. 104-jobs_9-of-40.pdfA score measures a mood, a question measures the problem
⛓ Source · 104 jobs with no follow-up, 9 persisting problems out of 40 checked
What I record: 104 jobs are marked "completed". The tenant was contacted in none of them. That is the normal arrangement: the contractor declares it done, the case closes.
What it hides: a job can be done and not resolve anything. Across a sample of 40 of those cases contacted afterwards, 9 problems persisted — and in 6 the tenant had not raised a new request, taking it that the work was done and that it would be the same again.
What I do: I ask one question, only one, a few days after the works: "is it resolved?". No satisfaction scale, no score. A score measures a mood; a question measures the problem.
What I do with the answer: if it is no, I reopen the original case rather than create a new one — otherwise the home shows as having two requests instead of one unresolved problem.
What I do with reopenings, and it is the opposite of a scoreboard: I attach them to the home first, because some come back when the cause lies elsewhere — water tracking through a floor is not bad workmanship. So I bring back, contractor by contractor, the reopenings whose cause really is the same, with the home, the date, the reason and the original job. That file is discussed and signed: a remark to a contractor engages a contractual relationship, and it is yours to make. Across the 40 cases checked, 9 problems persisted: all nine are already worked up and ready to reopen. 104-jobs_9-of-40.pdfA score measures a mood, a question measures the problem
⛓ Source · 104 jobs with no follow-up, 9 persisting problems out of 40 checked
The window that closes badly — was it serious?
Filed as "joinery", handled in six weeks. In November, in an electrically heated home, it is a heating and damp problem.
What I record: the subject of the request was accurate and so was the classification. The timescale applied was joinery's — six weeks, which is reasonable for a window.
What was missing: the season, the heating, the aspect. A window that closes badly in June is a nuisance; the same one in November is a bill, cold, and condensation that damages the home.
What I do: I cross the subject with the period and with what you know of the property — heating, year of construction, floor. I do not requalify the urgency by myself: I flag that the standard classification may be understated, and I say why.
What I took out of the cross-check, and I measured it before taking it out: the tone of the message. Across the 6,200 requests it lifted 210 files and pushed 178 down — 41 of which concerned a heating fault in December. A tenant who writes curtly does not have a more urgent problem than a tenant who apologises for disturbing you — and the second is almost always the one who has waited longest. So the cross-check keeps only facts: subject, period, heating, year, floor, and each one is shown in the flag.
What that gave: of 6,200 requests, 147 warranted a faster look under that cross-check. 38 were reclassified by the technical team. 147-requests_38-reclassified.pdfThe second is almost always the one who has waited longest
⛓ Source · 6,200 requests, 147 flagged, 38 reclassified by the team
What I record: the subject of the request was accurate and so was the classification. The timescale applied was joinery's — six weeks, which is reasonable for a window.
What was missing: the season, the heating, the aspect. A window that closes badly in June is a nuisance; the same one in November is a bill, cold, and condensation that damages the home.
What I do: I cross the subject with the period and with what you know of the property — heating, year of construction, floor. I do not requalify the urgency by myself: I flag that the standard classification may be understated, and I say why.
What I took out of the cross-check, and I measured it before taking it out: the tone of the message. Across the 6,200 requests it lifted 210 files and pushed 178 down — 41 of which concerned a heating fault in December. A tenant who writes curtly does not have a more urgent problem than a tenant who apologises for disturbing you — and the second is almost always the one who has waited longest. So the cross-check keeps only facts: subject, period, heating, year, floor, and each one is shown in the flag.
What that gave: of 6,200 requests, 147 warranted a faster look under that cross-check. 38 were reclassified by the technical team. 147-requests_38-reclassified.pdfThe second is almost always the one who has waited longest
⛓ Source · 6,200 requests, 147 flagged, 38 reclassified by the team
Can you reply to tenants directly?
On facts, yes. On anything that commits them, never.
What I can write: the acknowledgement with the real timescale observed for that type of request, the progress, the works date once the contractor confirms it, and the follow-up question after the visit.
What gets signed, and the work I do before it: a timescale commitment, how a cost splits between landlord and tenant, a position on a dispute, a refusal to take a case on. Take the second one: who pays for a repair depends on the lease, the age of the equipment and the cause of the damage — and an automatic answer there is a legal position taken without a file, in writing, therefore enforceable.
So I build the whole file: I find and quote the applicable clause of the lease concerned, I attach the year the equipment was fitted, the home's history and the cause recorded by the contractor, and I draft the answer to the tenant, ready to be read over. Whoever manages the lease decides and signs in two minutes, instead of reopening the lease and searching: the tenant has the answer the same day, and it is reasoned.
And a works date goes out only once the contractor confirms it, never "estimated": an announced date is waited for — a tenant who takes a half-day off for nothing does not do it twice, and they stop believing the dates that follow. what-is-never-written.pdfAn announced date is waited for
✎ Framework · no position on costs, no disputes, no estimated dates
What I can write: the acknowledgement with the real timescale observed for that type of request, the progress, the works date once the contractor confirms it, and the follow-up question after the visit.
What gets signed, and the work I do before it: a timescale commitment, how a cost splits between landlord and tenant, a position on a dispute, a refusal to take a case on. Take the second one: who pays for a repair depends on the lease, the age of the equipment and the cause of the damage — and an automatic answer there is a legal position taken without a file, in writing, therefore enforceable.
So I build the whole file: I find and quote the applicable clause of the lease concerned, I attach the year the equipment was fitted, the home's history and the cause recorded by the contractor, and I draft the answer to the tenant, ready to be read over. Whoever manages the lease decides and signs in two minutes, instead of reopening the lease and searching: the tenant has the answer the same day, and it is reasoned.
And a works date goes out only once the contractor confirms it, never "estimated": an announced date is waited for — a tenant who takes a half-day off for nothing does not do it twice, and they stop believing the dates that follow. what-is-never-written.pdfAn announced date is waited for
✎ Framework · no position on costs, no disputes, no estimated dates
On the requests that come back all the time, what exactly do you prepare?
I sorted your 6,200 requests by subject: nine families cover 71 % of the volume.
The prepared routine replies, as the property manager receives them: for each of the nine families — no hot water, cold radiator, door entryphone, leak under the sink, bulky waste, lost fob, landing light, noisy ventilation, forced letterbox — I write the full text, with the delay actually observed for that type of job in that building, the lease clause where one applies, and the contractor holding the trade package. 4,402 requests out of 6,200 go back out with a reply ready to read over and send.
What that gives back: a hand-written reply takes 11 minutes; reading over a prepared one takes 2. Across 4,402 requests, more than 660 hours given back over the year — more than eighteen 35-hour weeks, rounded down.
Dated information goes out by SMS, because an e-mail arrives after the visit: water shut-off announced the evening before at 6 p.m., contractor's visit confirmed that morning, appointment moved. The SMS carries those three only, and never a message that commits the landlord.
The figure that does not flatter me: of the first 300 prepared replies, 27 were rewritten by your managers — 9 %. 22 of the 27 were in the same family, the entryphone: my text announced an average delay of 8 days, when in the two buildings fitted with the older model it is 21 days for want of parts. The average was correct and useless — it mixed two estates. I split the family in two by installed model; over the next 900 replies, rewriting fell to 2 %.
⛓ Sourced · 9 families for 71 % of volume, 4,402 prepared replies, 27 rewritten out of 300
The prepared routine replies, as the property manager receives them: for each of the nine families — no hot water, cold radiator, door entryphone, leak under the sink, bulky waste, lost fob, landing light, noisy ventilation, forced letterbox — I write the full text, with the delay actually observed for that type of job in that building, the lease clause where one applies, and the contractor holding the trade package. 4,402 requests out of 6,200 go back out with a reply ready to read over and send.
What that gives back: a hand-written reply takes 11 minutes; reading over a prepared one takes 2. Across 4,402 requests, more than 660 hours given back over the year — more than eighteen 35-hour weeks, rounded down.
Dated information goes out by SMS, because an e-mail arrives after the visit: water shut-off announced the evening before at 6 p.m., contractor's visit confirmed that morning, appointment moved. The SMS carries those three only, and never a message that commits the landlord.
The figure that does not flatter me: of the first 300 prepared replies, 27 were rewritten by your managers — 9 %. 22 of the 27 were in the same family, the entryphone: my text announced an average delay of 8 days, when in the two buildings fitted with the older model it is 21 days for want of parts. The average was correct and useless — it mixed two estates. I split the family in two by installed model; over the next 900 replies, rewriting fell to 2 %.
⛓ Sourced · 9 families for 71 % of volume, 4,402 prepared replies, 27 rewritten out of 300
Do you keep the request history per tenant?
Requests are attached to the DWELLING, not the occupant — and that choice earns you more than it costs.
What it revealed: 118 dwellings out of 1,480 account for 41 % of requests. That is not 8 % of difficult tenants: it is 8 % of failing dwellings, and the problem was being sought in the occupant. A per-tenant history would have produced a list of names; a per-dwelling history produces a list of works.
What I keep in no case: the number of requests per tenant, nor the tone of their messages, nor their tenancy length set against their reports. Those are data that serve only to characterise a person, and they resurface at the worst moment — at lease renewal, before a committee.
If you want that view anyway: it requires attaching requests to the person, which means informing tenants and declaring a purpose. I will tell you so, and I will hand you back the 118 dwellings — because in ten cases out of ten, that is the figure that fixes the problem.
The signal I watch most closely, and that no dashboard produces: 23 dwellings have made no request at all in two years. That is not necessarily good news. the-dwelling-rather-than-the-occupant.pdfWhat it revealed · what is not kept · the 23 silent ones
⛓ Source · 118 dwellings of 1,480 for 41 % of requests, 23 silent dwellings
What it revealed: 118 dwellings out of 1,480 account for 41 % of requests. That is not 8 % of difficult tenants: it is 8 % of failing dwellings, and the problem was being sought in the occupant. A per-tenant history would have produced a list of names; a per-dwelling history produces a list of works.
What I keep in no case: the number of requests per tenant, nor the tone of their messages, nor their tenancy length set against their reports. Those are data that serve only to characterise a person, and they resurface at the worst moment — at lease renewal, before a committee.
If you want that view anyway: it requires attaching requests to the person, which means informing tenants and declaring a purpose. I will tell you so, and I will hand you back the 118 dwellings — because in ten cases out of ten, that is the figure that fixes the problem.
The signal I watch most closely, and that no dashboard produces: 23 dwellings have made no request at all in two years. That is not necessarily good news. the-dwelling-rather-than-the-occupant.pdfWhat it revealed · what is not kept · the 23 silent ones
⛓ Source · 118 dwellings of 1,480 for 41 % of requests, 23 silent dwellings
So what do you keep?
What describes the estate — and what a request does not say about itself.
What is kept: requests attached to the dwelling, their subject and nature, real lead times by type of works, the interventions and their closure, the dwellings with no requests, and the urgency reclassifications.
What a request does not say about itself: a window that closes badly was classed "joinery" and handled in six weeks. In November, in an electrically heated dwelling, it is a heating and damp problem. I do not reclassify the urgency on my own: I flag that the standard classification may be understated — by season and heating type — and I say why.
Closure, the blind spot: 104 interventions are marked "completed" with no feedback from the occupant. They are closed by whoever carried out the work, never by whoever had the problem. I now request a confirmation, and no reply is not counted as a success.
What I can write directly to the tenant: the acknowledgement with the real observed lead time for that type of request — not a theoretical one —, the progress status, the agreed intervention date. What binds the landlord stays signed by the landlord. what-you-keep_tenants.pdf5 items kept · closure that does not belong to whoever intervened
⛓ Source · 104 interventions closed with no occupant feedback, seasonal reclassification
What is kept: requests attached to the dwelling, their subject and nature, real lead times by type of works, the interventions and their closure, the dwellings with no requests, and the urgency reclassifications.
What a request does not say about itself: a window that closes badly was classed "joinery" and handled in six weeks. In November, in an electrically heated dwelling, it is a heating and damp problem. I do not reclassify the urgency on my own: I flag that the standard classification may be understated — by season and heating type — and I say why.
Closure, the blind spot: 104 interventions are marked "completed" with no feedback from the occupant. They are closed by whoever carried out the work, never by whoever had the problem. I now request a confirmation, and no reply is not counted as a success.
What I can write directly to the tenant: the acknowledgement with the real observed lead time for that type of request — not a theoretical one —, the progress status, the agreed intervention date. What binds the landlord stays signed by the landlord. what-you-keep_tenants.pdf5 items kept · closure that does not belong to whoever intervened
⛓ Source · 104 interventions closed with no occupant feedback, seasonal reclassification
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 tenant relationship. 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 566 € excl. VAT / month
Qualifying enquiries
Identifies the unit, the agreement and the nature of the request as soon as it comes in.
Finding the clause
Finds the applicable clause in the agreement concerned and cites it.
Everyday replies prepared
Drafts the reply to recurring requests, ready to read over and send. Dated information — a visit, a shut-off, an appointment — goes out by text message.
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
Sources, access rights and handling of questions with no answer
Data protection, rights and human oversight
What the agent must be connected to
This connection is required for the agent to work. It concerns your information system and is scoped during the audit.
Integration with existing tools without double entry
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.
Debt collection agent (multichannel sequences)
This agent qualifies and answers. An unpaid rent, once established, belongs to the collections agent.
Debt collection agent (multichannel sequences) from 587 € excl. VAT / month Discover the agent →End-to-end contract management agent
The lease itself — drafting, amendments, deadlines — is handled by a contract agent.
End-to-end contract management agent from 702 € excl. VAT / month Discover the agent →
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 much time can a manager give back to tenants?
By taking on the clause search and the everyday drafting, the effort shifts towards the situations that call for a real judgement. How large the gain is depends on your volume and remains to be confirmed by a pilot.
Finding the applicable clause
Drafting the everyday replies
Qualifying incoming requests
Indicative figures, not contractual, to be confirmed by a pilot on the size of your portfolio of units. Goodwill gestures, works and litigation commit your company and the contract: those decisions belong to the manager.
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 tenant relations agent (qualification, clauses, prepared replies), installed and operated for you.
Agility
Setup + controlled subscription
5,065 € excl. VAT setup
then 566 € 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
846 € 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
8,400 € excl. VAT setup
then 766 € excl. VAT/month · + hardware from 2,491 € (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 lettings management
Your tenants' data stays with youLocal inference or an isolated resource hosted in France; no tenant contact details or circumstances entrusted to a third party, no data used to train a model.
Data in France, under French lawYour tenants' contact details and circumstances: minimisation and location in France, architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity.
The officer keeps the decisionThe agent produces prepared replies and qualified requests, which can be checked and altered; no approval is automated.
Human oversight & traceabilityOn the size of your portfolio of units: systematic logging and tracking, in line with the AI Act.
Going further
Related resources
Frequently asked questions
Your questions, our answers
Does the agent reply to the tenant directly?
The reply is prepared and put to the manager, who sends it. A written reply commits the landlord: it goes through a human reading.
What does the agent rely on to answer?
On the tenancy agreement for the unit concerned and your internal procedures. Every reply cites the clause it rests on, so the reading over is immediate.
Does it handle arrears situations?
It gathers the material in the file — dates, amounts, clauses, exchanges — so the decision can be quick and documented. What follows remains a management decision.
Is our tenants' 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.
Does it connect to our lettings software?
Yes, to the common tools on the market, reading the agreements and the files. The integration is defined at the design stage.
How long does it take to deploy this agent?
A few weeks as a rule, depending on the size of the portfolio and the integration with your lettings software, after a free audit then a phase of design, integration and testing.
Which tools can people use to talk to the agent?
The ones you already use. Internally, your teams write to it from Microsoft Teams, Slack or by email, just as they would to a colleague: nothing to install, nothing to learn. On the tenant side, the agent answers on WhatsApp Business, your website chat and email — your tenants reach you where they already write, with no account to create. You keep control from a web dashboard. These connectors rely on open standards, including the MCP protocol; they are included in every plan, at no extra cost, within the number of connections your level includes, and the catalogue grows at no extra cost. Only the fees charged by the platforms themselves — WhatsApp Business bills per conversation, and the gateway bills the text messages sent — are passed on at actual cost, with no margin, outside the subscription.
Can the agent notify tenants by text message?
Yes, as an option. The agent uses a text message to warn of a technical visit, a water shut-off in the building or an inventory appointment. 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 property management
Let's talk
Let's size up the potential in your lettings management
15 minutes to frame your portfolio and your agreements — hosted in France, supervised, with no commitment.