We build Agents. You build the Future.
● B2B offer — Business servicesThe AI agent for service firms: an up-to-date CRM, time given back to selling
In a service firm, billable time drains away into administration: meeting notes, updating the CRM, follow-ups, quotations, repetitive replies. Your AI agent absorbs this coordination work. Hosted in France — on local inference or an isolated resource — your client data and your pipeline stay under control. Your teams keep the decision.
Updated on
Deployed in a few weeks
Sales assistant · hosted in France
Fictional simulation — not a client result
I have just come out of a meeting with Delcourt Conseil. Write up the notes, update the CRM and prepare the next actions.
Notes written up from your voice memos: a need for six months of support, a budget of €25k mentioned, decision-maker identified (the finance director), a competitor in the running.
CRM updated: the opportunity moved to "proposal", amount and expected date filled in. 3 actions proposed: send the quotation within 48 hours, rebook a review on the 15th, follow up if there is no reply by day 5.
⛓ Sourced · your meeting notes + the client record in the CRM
CRM updated: the opportunity moved to "proposal", amount and expected date filled in. 3 actions proposed: send the quotation within 48 hours, rebook a review on the 15th, follow up if there is no reply by day 5.
⛓ Sourced · your meeting notes + the client record in the CRM
How many quotations are due for a follow-up this week?
7 quotations unanswered for more than 5 days, of which 2 are priorities (high value, close deadline). For each one I have prepared a personalised follow-up picking up the context of the last exchange, professional in tone and not pushy.
The follow-ups are ready — you read them over and approve the send.
✎ Action · 7 follow-ups ready to read over — the salesperson approves
The follow-ups are ready — you read them over and approve the send.
✎ Action · 7 follow-ups ready to read over — the salesperson approves
Data hosted in France · sovereign model as an option
Data hosted in France
Hosted in FranceAI model called at its publisher, except with the Sovereign model option
Client data protected
GDPR & AI Act: governed deploymentTraceability & human oversight
TurnkeyDesigned, installed and operated for you
The decisions remain yoursThe agent carries out what you have authorised
For an organisation established outside the European Union, the law of its own country applies; at its request, we put equivalent safeguards in place (human review before any action, a log of the criteria, information for the people concerned).
✦ In brief
In a service firm, a Blue Lemon Agent agent automates sales and administrative coordination — meeting notes, updating the CRM, follow-ups, qualifying enquiries, client support — and frees up billable time. Data hosted in France; the AI model is called at its publisher, except with the Sovereign model option. With that option, your client records, pipeline and exchanges are never exposed to a foreign service, architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity. Time gained on administration is redirected to delivery and business development. Live within a few weeks.
100%
hosted in France in the target architecture
0
transfer outside the EU with the Sovereign model option (target architecture)
3
uses ready to deploy on this scope
0
decision with a legal or significant effect taken by an agent
Reference points describing our offer, not results measured at a client. The scale of the gain is confirmed by a pilot on your own scope.
Indicative estimate, to be confirmed by a pilot on your real files (general terms of sale, article 4).
Pricing
Three options, one agent
A sales agent (meeting notes, CRM, follow-ups), installed and operated for you. Choose according to how you work.
Agility
Setup + controlled subscription
2,490 € excl. VAT setup
then 269 € 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
- Hosting in France, a dedicated and isolated resource
The simplest
Serenity
All inclusive, no setup fee
404 € 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
- Hosting in France, managed end to end
100% Sovereign
On site, you own it
4,340 € excl. VAT setup
then 369 € 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.
- Hardware installed on your premises (you own it)
- French / European AI models run locally
- Secure remote maintenance (Pro support included)
Included in the package: hosting and AI consumption up to the volume of your capacity tier, in documents, exchanges, tokens and storage; beyond that, re-invoiced at real cost with no margin, tracked in your client area as usage statements are issued. 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 for the package chosen, in France by default, 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.
Use cases
The uses of AI in a service firm
Here are the capabilities included in the agent, then the neighbouring agents that can be added to it. The agent carries out by itself what you authorise it to do; you decide what it does on its own and what it submits to you.
Included in your agent
The 3 capabilities essential to this promise are included, at no extra cost.
From 269 € excl. VAT / month
Request qualification & explicit prioritisation
Taking up pending requests and files, qualifying them, and saying which ones come first — amount, deadline, age — rather than handling everything in order of arrival.
Structured meeting summary
From your notes: the need expressed, the budget mentioned, the decision-maker identified, the competitors in the running — a summary reviewed and approved by the salesperson. Keeping the CRM — opportunities, scoring, forecasts — is the work of a dedicated agent, ordered separately.
Report of files left unanswered
Quotes and requests past the agreed deadline come back with their context and their age; sending any follow-up stays the salesperson’s decision.
Controls and safeguards
These 3 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 oversight, costs, quality, incidents and safe stop
Prioritise opportunities on explicit criteria
Personalise without breaching consent or prospecting rules
What the agent must be connected to
These 2 connections are required for the agent to work. They concern your information system and are scoped during the audit.
Handoff contracts, least-privilege permissions and shared context
Prepare the next action and write back to the CRM without loss
Other needs our agents cover
Each card says how the matching agent can be obtained today: available, on quote, or not yet in the catalogue.
Customer support
Answer recurring questions, track requests and route the complex cases to the right person.
Available 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.
Outbound prospecting
Identifying targets, personalised approach sequences and tracking of replies to feed the pipeline.
Outbound prospecting agent (automated email, assisted LinkedIn) from 332 € excl. VAT / month Sales prospecting →Collection & chasing unpaid invoices
Chase outstanding quotations and invoices in multi-channel sequences, with the right tone and the right timing — you keep control.
Debt collection agent (multichannel sequences) from 307 € excl. VAT / month Debt recovery →Accounts & administration
Reading documents, preparing entries and tracking payments — proposed, for your finance team to approve.
Accounting agent (summaries, anomalies) from 406 € excl. VAT / month Accounting agent →Tender responses
Pre-fill a response file from your references and the invitation to tender, to read over and complete before submission.
Tender response agent (drafting) from 354 € excl. VAT / month Tender responses →Voice switchboard
Take calls, answer simple enquiries and transfer to the right department, with no call lost.
Voice agent (AI switchboard) from 456 € excl. VAT / month Voice switchboard →AI sales & CRM agent
Every salesperson knows it: the CRM is never really up to date.
Sales agent (meeting notes, follow-ups, CRM) from 307 € excl. VAT / month Discover the agent →AI agent for lead qualification and appointment booking
A lead who contacts you expects an answer straight away.
Lead qualification / appointment booking agent from 312 € excl. VAT / month Discover the agent →AI knowledge base agent
The information exists in your company — but it is scattered across procedures, contracts, an intranet and the memory of a few people.
Document agent (FAQ, knowledge base) from 353 € excl. VAT / month Discover the agent →
Does your need fall outside this?
Build your agent
In 30 minutes we identify the most relevant agent — without oversizing the project.
Demonstration
See the agent at work
5 everyday 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.
Sales assistant · hosted in France
Fictional simulation — not a client result
Four flags this morning, with no request from you.
· A time-and-materials engagement has passed its purchase order ceiling by 18%. 32 days booked against 27 ordered, and the work continues.
· A framework agreement renews in 14 days on 2022 rates. Yours have been revised twice since.
· A consultant is assigned to two engagements on the same days in the week of 17/08, at two different clients.
· Four accounts have had no contact for more than five months, including your second-largest by revenue. morning-watch_4-flags.pdf4 flags · amount and deadline
⛓ Source · purchase orders, time records, framework agreements, resourcing plan, CRM
· A time-and-materials engagement has passed its purchase order ceiling by 18%. 32 days booked against 27 ordered, and the work continues.
· A framework agreement renews in 14 days on 2022 rates. Yours have been revised twice since.
· A consultant is assigned to two engagements on the same days in the week of 17/08, at two different clients.
· Four accounts have had no contact for more than five months, including your second-largest by revenue. morning-watch_4-flags.pdf4 flags · amount and deadline
⛓ Source · purchase orders, time records, framework agreements, resourcing plan, CRM
The ceiling overrun — can it be recovered?
Yes, and here is the good news: the five days over are documented.
What I reconstructed: the purchase order covers 27 days. 32 have been booked by two consultants. The five extra days match three client requests, all in writing: two emails from their project director and a steering committee minute they approved.
What is missing, and it is worth five days: no variation was ever requested. The five days are real, the request is written, and nothing covers them contractually.
The variation request is ready, with the three documents, the dates recalled and the amount computed at the purchase order's rate. Sent today it covers days already delivered and evidenced; sent at closure it covers a surprise invoice — and that is the whole difference between a regularisation and a dispute.
What is left to negotiate, and it is yours: sending it. A variation is negotiated with a person across the table, it commits the firm to an amount and a date, and it carries a signature. You read it, you choose the moment — every day of waiting makes the document harder to defend. ceiling-overrun_variation-ready.pdf5 days · 3 written requests
⛓ Source · purchase order, time records, 3 written client requests
What I reconstructed: the purchase order covers 27 days. 32 have been booked by two consultants. The five extra days match three client requests, all in writing: two emails from their project director and a steering committee minute they approved.
What is missing, and it is worth five days: no variation was ever requested. The five days are real, the request is written, and nothing covers them contractually.
The variation request is ready, with the three documents, the dates recalled and the amount computed at the purchase order's rate. Sent today it covers days already delivered and evidenced; sent at closure it covers a surprise invoice — and that is the whole difference between a regularisation and a dispute.
What is left to negotiate, and it is yours: sending it. A variation is negotiated with a person across the table, it commits the firm to an amount and a date, and it carries a signature. You read it, you choose the moment — every day of waiting makes the document harder to defend. ceiling-overrun_variation-ready.pdf5 days · 3 written requests
⛓ Source · purchase order, time records, 3 written client requests
How did you spot all that? And who do you tell?
I watch, continuously, what you have opened to me: your purchase orders against time booked, your framework agreements and their rates, the resourcing plan, and the date of last contact on each account. A ceiling crossed, a renewal approaching, two engagements overlapping, an account going quiet — I say so, I do not wait to be asked.
Routing follows your organisation: the overrun to the engagement director and the account partner, because a variation is negotiated by two; the renewal to the partner; the overlap to resourcing, with both clients named; the quiet accounts to their lead partner.
With a chase: 24 h on the overrun and the overlap — one grows daily, the other has a date; 72 h on the rest. Then a weekly summary: by account, never by consultant.
And I do not stop at the flag: all four files are built. The variation is drafted with its three documents and its amount; the renewal comes back with the framework agreement's rate card and the gap between the contractual rate and the average rate actually invoiced; the overlap arrives with two possible re-plans, each costed in days moved; the quiet accounts with the date and subject of the last exchange. That is half a partner-day handed back every week.
What is left to you, and what justifies it: sending a variation, invoicing, changing a resourcing plan, contacting a client. All four commit the firm towards a third party or towards a consultant — that is what makes them enforceable, and why they carry a name.
And I read what you have opened to me: an agreement made verbally at the client is invisible to me, and saying so is more use than keeping quiet about it. Every access is logged and withdrawable on a word.
✎ Proposal · watch and chases to be configured — you set the thresholds
Routing follows your organisation: the overrun to the engagement director and the account partner, because a variation is negotiated by two; the renewal to the partner; the overlap to resourcing, with both clients named; the quiet accounts to their lead partner.
With a chase: 24 h on the overrun and the overlap — one grows daily, the other has a date; 72 h on the rest. Then a weekly summary: by account, never by consultant.
And I do not stop at the flag: all four files are built. The variation is drafted with its three documents and its amount; the renewal comes back with the framework agreement's rate card and the gap between the contractual rate and the average rate actually invoiced; the overlap arrives with two possible re-plans, each costed in days moved; the quiet accounts with the date and subject of the last exchange. That is half a partner-day handed back every week.
What is left to you, and what justifies it: sending a variation, invoicing, changing a resourcing plan, contacting a client. All four commit the firm towards a third party or towards a consultant — that is what makes them enforceable, and why they carry a name.
And I read what you have opened to me: an agreement made verbally at the client is invisible to me, and saying so is more use than keeping quiet about it. Every access is logged and withdrawable on a word.
✎ Proposal · watch and chases to be configured — you set the thresholds
We handle everything in the order it arrives and we lose track. Can you say what goes first?
63 requests and files pending, gone through one by one: qualification first, prioritisation second, and it is explicit.
Qualification sorts each file by what it calls for, not by the inbox it came from: 21 quotes to draw up or chase, 17 requests for information, 14 engagement files waiting on a document from the client, 8 complaints, 3 framework agreements to renew.
Prioritisation applies three criteria, and you see the arithmetic: the amount at stake, the deadline running, the age of the request. Every file carries its rank and the sentence that explains it — "2nd: €46,000, answer due in 4 days, waiting 9 days". A rank with no reason cannot be challenged, so it never gets corrected.
What order of arrival was costing you, measured over the last quarter: the 21 quotes are worth €214,000, and the 5 largest — €118,000 between them — sat at ranks 12 to 19 in the pile, behind ten-minute information requests.
What I propose and you alone decide: the 17 information requests go out today, because they cost ten minutes each and clear the pile; the 3 framework renewals go ahead of the quotes, because a framework not renewed closes the door on the quotes that follow.
One weighting got it wrong, and I am publishing it: my first version put age ahead of deadline. Three quotes due in 5 days ended up at rank 4, and two were submitted late. Deadline now comes ahead of age as soon as it is under 7 days; over the following six weeks, nothing submitted late across 38 deadlines.
✎ Framework · qualification by nature, prioritisation on three stated criteria
Qualification sorts each file by what it calls for, not by the inbox it came from: 21 quotes to draw up or chase, 17 requests for information, 14 engagement files waiting on a document from the client, 8 complaints, 3 framework agreements to renew.
Prioritisation applies three criteria, and you see the arithmetic: the amount at stake, the deadline running, the age of the request. Every file carries its rank and the sentence that explains it — "2nd: €46,000, answer due in 4 days, waiting 9 days". A rank with no reason cannot be challenged, so it never gets corrected.
What order of arrival was costing you, measured over the last quarter: the 21 quotes are worth €214,000, and the 5 largest — €118,000 between them — sat at ranks 12 to 19 in the pile, behind ten-minute information requests.
What I propose and you alone decide: the 17 information requests go out today, because they cost ten minutes each and clear the pile; the 3 framework renewals go ahead of the quotes, because a framework not renewed closes the door on the quotes that follow.
One weighting got it wrong, and I am publishing it: my first version put age ahead of deadline. Three quotes due in 5 days ended up at rank 4, and two were submitted late. Deadline now comes ahead of age as soon as it is under 7 days; over the following six weeks, nothing submitted late across 38 deadlines.
✎ Framework · qualification by nature, prioritisation on three stated criteria
Fictional simulation — not a client result
I bought a list of 4,000 finance director contacts. Run a three-message sequence on it.
Before sending, one point that governs everything else: I do not have the source of these contacts.
A bought list says neither where the addresses came from nor what the people agreed to. Without that, your firm cannot answer the first question it will be asked — and that is not a theoretical question: it is the one an irritated recipient asks, and the one the list vendor gets when you pass it on.
What I did instead, and it gives a better starting point: I matched the 4,000 contacts against what you already hold.
· 410 are already in your CRM, 63 of them with an exchange in the last twelve months. Contacting them as strangers would be the weakest message you could send.
· 78 are former clients.
· 140 match your tight target — size, sector, and a recent signal of a role opening or a funding round.
The sequence on those 140 is ready, personalised from what you know about them. list-of-4000_matched-to-CRM.pdf410 already known · 78 former clients · 140 targeted
⛓ Source · CRM, client history, list supplied
A bought list says neither where the addresses came from nor what the people agreed to. Without that, your firm cannot answer the first question it will be asked — and that is not a theoretical question: it is the one an irritated recipient asks, and the one the list vendor gets when you pass it on.
What I did instead, and it gives a better starting point: I matched the 4,000 contacts against what you already hold.
· 410 are already in your CRM, 63 of them with an exchange in the last twelve months. Contacting them as strangers would be the weakest message you could send.
· 78 are former clients.
· 140 match your tight target — size, sector, and a recent signal of a role opening or a funding round.
The sequence on those 140 is ready, personalised from what you know about them. list-of-4000_matched-to-CRM.pdf410 already known · 78 former clients · 140 targeted
⛓ Source · CRM, client history, list supplied
And the other 3,460 — do we bin them?
No — we first ask the vendor what we bought, and that question has immediate commercial value.
What I prepared for them: four short questions — where the data came from, on what basis it was collected, when it was last updated, and the bounce rate seen by their other clients. A serious vendor answers in two days; a vendor who does not answer has just told you something.
What I measured on the sample meanwhile: of the 410 contacts you already hold, 97 have an address that differs from yours and 31 have changed roles according to your own notes. On a 10% sample, that suggests roughly a quarter of the list is stale.
What that changes in practice: sending to 4,000 addresses of which a quarter bounce damages your sending domain's reputation, and the 140 good contacts land in spam with the rest.
My proposal is one sentence: the 140 today, the 3,460 once the vendor has answered — and a systematic check of every new list against the CRM before the first send. On this one, it would have avoided writing to 63 people who were talking to you six months ago. vendor-questions_list.pdf4 questions · sample measurement
✎ Action · vendor questions ready · 140 usable contacts today
What I prepared for them: four short questions — where the data came from, on what basis it was collected, when it was last updated, and the bounce rate seen by their other clients. A serious vendor answers in two days; a vendor who does not answer has just told you something.
What I measured on the sample meanwhile: of the 410 contacts you already hold, 97 have an address that differs from yours and 31 have changed roles according to your own notes. On a 10% sample, that suggests roughly a quarter of the list is stale.
What that changes in practice: sending to 4,000 addresses of which a quarter bounce damages your sending domain's reputation, and the 140 good contacts land in spam with the rest.
My proposal is one sentence: the 140 today, the 3,460 once the vendor has answered — and a systematic check of every new list against the CRM before the first send. On this one, it would have avoided writing to 63 people who were talking to you six months ago. vendor-questions_list.pdf4 questions · sample measurement
✎ Action · vendor questions ready · 140 usable contacts today
Fictional simulation — not a client result
Give me the profitability of this quarter's engagements.
14 engagements, and I can calculate 11.
The eleven measurable ones: complete time records, expenses attached, billing rates applied. Gross margin averaging 34%, from 12% to 51% across engagements.
The other three, and what each is missing:
· Delcourt engagement — 41 days booked, but three consultants have not booked time for six weeks. The real time is higher, by an unknown amount.
· Verne engagement — travel expenses sit in the general account, not attached to the engagement. Of twelve trips, only four carry the engagement code.
· Aubel engagement — pre-sales time appears nowhere. Six meetings and a pitch, before signature.
What I did to make them calculable, rather than estimating them: for Delcourt, the named chaser to the three consultants is written, with the exact weeks to complete; for Verne, the eight uncoded trips are found in the general account, amounts and dates in hand, ready to attach; for Aubel, the six meetings and the pitch are dated from the diaries, only the durations remain to be confirmed. Three acts, and all 14 margins are measured next week.
Why I do not estimate them meanwhile: an estimated margin shown beside eleven measured ones reads as a measured margin — and that is the one the committee will remember. quarter-margins_11-of-14.pdf11 measured · 3 unmeasurable, reason given
⛓ Source · time records, expense claims, billing rates
The eleven measurable ones: complete time records, expenses attached, billing rates applied. Gross margin averaging 34%, from 12% to 51% across engagements.
The other three, and what each is missing:
· Delcourt engagement — 41 days booked, but three consultants have not booked time for six weeks. The real time is higher, by an unknown amount.
· Verne engagement — travel expenses sit in the general account, not attached to the engagement. Of twelve trips, only four carry the engagement code.
· Aubel engagement — pre-sales time appears nowhere. Six meetings and a pitch, before signature.
What I did to make them calculable, rather than estimating them: for Delcourt, the named chaser to the three consultants is written, with the exact weeks to complete; for Verne, the eight uncoded trips are found in the general account, amounts and dates in hand, ready to attach; for Aubel, the six meetings and the pitch are dated from the diaries, only the durations remain to be confirmed. Three acts, and all 14 margins are measured next week.
Why I do not estimate them meanwhile: an estimated margin shown beside eleven measured ones reads as a measured margin — and that is the one the committee will remember. quarter-margins_11-of-14.pdf11 measured · 3 unmeasurable, reason given
⛓ Source · time records, expense claims, billing rates
We have never counted pre-sales. Is that a problem?
It is not a problem and it is instructive — because that is exactly where the gap sits between your best engagements and the rest.
What I could reconstruct, from diaries rather than time sheets: across the eleven measurable engagements, pre-sales runs from 2 to 14 days. The two least profitable engagements of the quarter are also the two with the longest pre-sales.
What that correlation does not prove: that long pre-sales causes low margin. It may equally be the symptom of an undecided client, who stays undecided during delivery.
What I can say with certainty: none of your go / no-go decisions has that figure today, although it sits in your diaries.
A proposal, if you approve it: automatically attach to an opportunity the diary entries that name it, and offer the consultant a one-click confirmation. Nothing is booked on their behalf — it is a suggestion they approve or correct.
And for the three unmeasurable engagements: a booking reminder at day 7 to the consultant, at day 14 to the engagement director. By engagement, never by consultant — a list of late time sheets by name gets you time sheets, not accurate ones. pre-sales_reconstructed.pdf11 engagements · 2 to 14 days
✎ Proposal · attachment offered, never imposed
What I could reconstruct, from diaries rather than time sheets: across the eleven measurable engagements, pre-sales runs from 2 to 14 days. The two least profitable engagements of the quarter are also the two with the longest pre-sales.
What that correlation does not prove: that long pre-sales causes low margin. It may equally be the symptom of an undecided client, who stays undecided during delivery.
What I can say with certainty: none of your go / no-go decisions has that figure today, although it sits in your diaries.
A proposal, if you approve it: automatically attach to an opportunity the diary entries that name it, and offer the consultant a one-click confirmation. Nothing is booked on their behalf — it is a suggestion they approve or correct.
And for the three unmeasurable engagements: a booking reminder at day 7 to the consultant, at day 14 to the engagement director. By engagement, never by consultant — a list of late time sheets by name gets you time sheets, not accurate ones. pre-sales_reconstructed.pdf11 engagements · 2 to 14 days
✎ Proposal · attachment offered, never imposed
Fictional simulation — not a client result
We are answering a tender on transforming a customer service function. Have we done this before?
Four times, and the most useful is not the most recent.
· Mareuil Distribution (2024) — closest scope: same size of function, same volume problem. Won, and usable as a reference.
· Groupe Aubel (2023) — larger, across three sites. Useful for methodology, not for pricing.
· Loiret Services (2022) — tender lost. The submission exists, and so does the debrief note: the client had found the proposal "too industrial".
· Verne Habitat (2025) — live engagement, so not citable as a reference without their agreement.
What I have prepared: the response outline from Mareuil, the methodology from Aubel, and the Loiret debrief at the front — because today's tender comes from a client of comparable size.
One document is missing: Mareuil's completion certificate is not on file. It can be requested, and it takes a week. references_4-comparable-engagements.pdf1 close win · 1 instructive loss
⛓ Source · 4 engagements, submissions, debrief notes
· Mareuil Distribution (2024) — closest scope: same size of function, same volume problem. Won, and usable as a reference.
· Groupe Aubel (2023) — larger, across three sites. Useful for methodology, not for pricing.
· Loiret Services (2022) — tender lost. The submission exists, and so does the debrief note: the client had found the proposal "too industrial".
· Verne Habitat (2025) — live engagement, so not citable as a reference without their agreement.
What I have prepared: the response outline from Mareuil, the methodology from Aubel, and the Loiret debrief at the front — because today's tender comes from a client of comparable size.
One document is missing: Mareuil's completion certificate is not on file. It can be requested, and it takes a week. references_4-comparable-engagements.pdf1 close win · 1 instructive loss
⛓ Source · 4 engagements, submissions, debrief notes
The Loiret debrief — what exactly did it say?
It runs to three lines, and they are usable as they stand.
The client had chosen a competitor, and agreed to tell you why: "a proposal that is too industrial, and does not show who will do the work". Two specific points were cited: no consultant named in the proposal, and a methodology identical to one they had received from another firm.
What that is worth today: the live tender comes from a client of comparable size, and your current outline still presents the methodology before the team.
What I have written, so you can compare: the same proposal in both orders — the named team first, methodology second, and the reverse — with strictly identical content. Across your fifteen tenders since 2022, the four you won named a consultant on the first page; of the eleven you lost, only one did.
What remains a choice, and it is one of positioning: which of the two outlines becomes yours. Two of your four engagements were won with the current one, and that fact counts as much as the other seven.
An observation beyond this tender: you hold eleven debrief notes from lost tenders since 2022, all written and none reread since. Seven mention the same criticism. That is no longer client feedback, it is a pattern.
What I can do, on your approval: on every new submission, surface the debriefs of lost tenders on a comparable scope, before drafting rather than after the result. lost-tender-debriefs_11.pdf11 tenders · 7 with the same criticism
✎ Support · 7 converging debriefs — the outline stays the firm's choice
The client had chosen a competitor, and agreed to tell you why: "a proposal that is too industrial, and does not show who will do the work". Two specific points were cited: no consultant named in the proposal, and a methodology identical to one they had received from another firm.
What that is worth today: the live tender comes from a client of comparable size, and your current outline still presents the methodology before the team.
What I have written, so you can compare: the same proposal in both orders — the named team first, methodology second, and the reverse — with strictly identical content. Across your fifteen tenders since 2022, the four you won named a consultant on the first page; of the eleven you lost, only one did.
What remains a choice, and it is one of positioning: which of the two outlines becomes yours. Two of your four engagements were won with the current one, and that fact counts as much as the other seven.
An observation beyond this tender: you hold eleven debrief notes from lost tenders since 2022, all written and none reread since. Seven mention the same criticism. That is no longer client feedback, it is a pattern.
What I can do, on your approval: on every new submission, surface the debriefs of lost tenders on a comparable scope, before drafting rather than after the result. lost-tender-debriefs_11.pdf11 tenders · 7 with the same criticism
✎ Support · 7 converging debriefs — the outline stays the firm's choice
Fictional simulation — not a client result
One of the quarter's engagements is both fixed-price and time-and-materials, and your two margin methods cannot read it together.
The engagement: Verne Industries — a fixed-price scoping phase at €14,800 and, in the same contract, a support phase billed by the day. One client, one contract, two billing modes.
The point that had not yet been defined in your method: the margin on a fixed price is calculated on time spent against price sold; that of time-and-materials is calculated on the day rate against the consultant's cost. Nothing says how the margin of an engagement carrying both is calculated. I did not settle it on my own — and the quarter is not waiting: times are complete on both phases, expenses are attached, and the margin is calculated under all three readings.
Three ways of doing it. A single margin on the whole contract: one figure per client, simple to present, but a comfortable time-and-materials phase masks a fixed price that overruns. Two separate margins, one per phase: each mode reads for what it is, at the price of a table with twice as many lines as engagements. Two separate margins, and a contract margin that consolidates them: you see the client first, then where the result comes from, at the price of one more column on the dashboard.
I recommend the third. It is the one that answers the two questions you actually ask — is this client profitable, and which of the two phases explains it — and your table already carries one line per engagement phase: the consolidation reads with no re-entry.
You can also give me another one — for instance: one margin per phase, and an alert threshold specific to fixed price, which you set.
Shall I remember it for the next ones? 4 of your 14 engagements this quarter carry two billing modes, and 11 out of the year's 38. If you approve the rule, I will calculate the next ones this way without bringing them to you one by one, and I will give you the count every month. engagement-fixed-price-and-time_three-ways.pdf1 engagement · 3 ways of doing it · 11 comparable cases
⛓ Sourced · Verne Industries contract, the firm's margin calculation method, the year's 38 engagements of which 11 with two modes
The engagement: Verne Industries — a fixed-price scoping phase at €14,800 and, in the same contract, a support phase billed by the day. One client, one contract, two billing modes.
The point that had not yet been defined in your method: the margin on a fixed price is calculated on time spent against price sold; that of time-and-materials is calculated on the day rate against the consultant's cost. Nothing says how the margin of an engagement carrying both is calculated. I did not settle it on my own — and the quarter is not waiting: times are complete on both phases, expenses are attached, and the margin is calculated under all three readings.
Three ways of doing it. A single margin on the whole contract: one figure per client, simple to present, but a comfortable time-and-materials phase masks a fixed price that overruns. Two separate margins, one per phase: each mode reads for what it is, at the price of a table with twice as many lines as engagements. Two separate margins, and a contract margin that consolidates them: you see the client first, then where the result comes from, at the price of one more column on the dashboard.
I recommend the third. It is the one that answers the two questions you actually ask — is this client profitable, and which of the two phases explains it — and your table already carries one line per engagement phase: the consolidation reads with no re-entry.
You can also give me another one — for instance: one margin per phase, and an alert threshold specific to fixed price, which you set.
Shall I remember it for the next ones? 4 of your 14 engagements this quarter carry two billing modes, and 11 out of the year's 38. If you approve the rule, I will calculate the next ones this way without bringing them to you one by one, and I will give you the count every month. engagement-fixed-price-and-time_three-ways.pdf1 engagement · 3 ways of doing it · 11 comparable cases
⛓ Sourced · Verne Industries contract, the firm's margin calculation method, the year's 38 engagements of which 11 with two modes
The third one. But the contract margin must carry the non-rechargeable expenses at contract level, not split by hand between the two phases. And yes, keep it.
Written as you have just said it, and it carries your correction. And re-reading the year's contracts, there is more to revisit than a calculation method.
The rule, in one sentence: an engagement with two modes carries one margin per phase and a contract margin that consolidates them; non-rechargeable expenses are carried at contract level, never split by hand between the phases.
What I saw re-reading your framework contracts, and it is worth more than today's case: 4 of your 11 framework contracts renew automatically with no rate review. Your rates have been revised twice since they were signed; those four contracts still bill the original schedule.
What it is worth, with your own figures: those 4 contracts carried 214 days billed over twelve months, at €640 instead of €720 — €17,120 of difference over the year, and the difference will repeat identically at the next renewal.
What I propose, and it is one line to change at your end: a written rate review 90 days before each automatic renewal date. I prepare the file — days billed, schedule applied, schedule in force, difference — and it is the partner who decides whether to open the discussion: a rate is renegotiated with a client, not with a table. The four files are ready, the nearest falls due in 14 days.
What lets you go back on today's rule: it is entered in your margin method under its name, with its date, the partner who approved it and the contract that gave rise to it. You remove it with one line.
✎ Rule approved and proposal · margin method, 11 framework contracts, 214 days billed on the old schedule — opening the discussion belongs to the partner
The rule, in one sentence: an engagement with two modes carries one margin per phase and a contract margin that consolidates them; non-rechargeable expenses are carried at contract level, never split by hand between the phases.
What I saw re-reading your framework contracts, and it is worth more than today's case: 4 of your 11 framework contracts renew automatically with no rate review. Your rates have been revised twice since they were signed; those four contracts still bill the original schedule.
What it is worth, with your own figures: those 4 contracts carried 214 days billed over twelve months, at €640 instead of €720 — €17,120 of difference over the year, and the difference will repeat identically at the next renewal.
What I propose, and it is one line to change at your end: a written rate review 90 days before each automatic renewal date. I prepare the file — days billed, schedule applied, schedule in force, difference — and it is the partner who decides whether to open the discussion: a rate is renegotiated with a client, not with a table. The four files are ready, the nearest falls due in 14 days.
What lets you go back on today's rule: it is entered in your margin method under its name, with its date, the partner who approved it and the contract that gave rise to it. You remove it with one line.
✎ Rule approved and proposal · margin method, 11 framework contracts, 214 days billed on the old schedule — opening the discussion belongs to the partner
Data hosted in France · sovereign model as an option
Your case is not here? That is exactly what a 30-minute conversation is for. Discuss your project →
How it works
The stages of your AI agent project
1
Audit & scoping
30 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.
The context
Why AI matters to service firms — and why they hesitate
What a service firm is worth is its experts' time and the quality of its client relationships. But that time is mechanically eaten into by data entry, coordination and follow-ups — and the data involved is the sensitive data of its whole portfolio.
! The issue
The service firm is caught between clients who expect responsiveness and flawless follow-up, and an administrative load that weighs on billable time (CRM, meeting notes, quotations, follow-ups, support). Yet most consumer AI tools amount to entrusting the client database, the sales pipeline, the exchanges and the pricing terms to a third party, often hosted outside Europe and subject to the Cloud Act.
✓ Our answer
AI is only of interest to a service firm if it is sovereign and confidential by design. Local inference or an isolated resource hosted in France, systematic human oversight, the decision reserved for your teams: time gained on administration is never paid for in lost confidentiality. The aim is not to replace your salespeople or your consultants, but to give them back useful time.
The decisive point
Confidentiality of your portfolio: sovereignty & compliance
Your client database and your pipeline are at the heart of your business. Here is how the architecture of our agents protects them, opportunity by opportunity.
Local inference
With the 100% Sovereign package, the agent runs on a machine belonging to the company: no customer data leaves the network, nothing passes through a cloud.
Hosting in France
Otherwise, with the Agility and Serenity packages, a dedicated, isolated resource hosted in France, under French law: data hosted in France; the AI model is called at its publisher, except with the Sovereign model option.
Reduced extraterritorial exposure
Architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity for your client database: our architecture depends on a subcontracting chain and remote access documented for the configuration chosen.
A resource isolated per firm
No pooling of your client database: an environment strictly dedicated to your company.
Encryption & controlled access
Encryption in transit and at rest, role-based access (RBAC), strong authentication and logging.
AI Act: governed deployment
The agent carries out what you have authorised; nothing is sent 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.
- The encryption mechanisms in transit and at rest, their components and key management are those documented for the architecture chosen.
- 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 the most sensitive data, SecNumCloud and HDS options are available 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.
The gain
How much time can a service firm recover?
By automating meeting notes, CRM updates and follow-ups, a service firm can aim for a reduction by half in salespeople's administrative time — reinvested in selling and in billable delivery.
Meeting notes + updating the CRM
Preparing a batch of follow-ups
Answering a recurring customer question
Qualitative, non-contractual comparison: the proportions shown illustrate the shift of the work towards review, they represent no measurement. Every output of the agent is reviewed and approved by a competent person.
Indicative estimate, to be confirmed by a pilot on your real files (general terms of sale, article 4).
Our commitment
Four guarantees that matter to a service firm
With 100% Sovereign, your client database never leaves the companyData hosted in France on an isolated resource; the AI model is called at its publisher, except with the Sovereign model option. With that option, no sales data entrusted to a foreign third party.
Data in France, under French lawNative location and minimisation for your client database; architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity.
Your teams keep the decisionThe agent produces meeting notes, follow-ups and summaries that can be checked; nothing is sent automatically without approval.
Human oversight & traceabilityMonitoring and logging frame the CRM & meeting notes.
Frequently asked questions
Your questions, our answers
Can AI really update my CRM on its own?
To a large extent, yes. From your meeting notes, emails or calls, the agent writes up the notes, qualifies the opportunity, updates the client record and proposes the next actions. You move from a data-entry role to a reading-over and approval role, which is far quicker.
Are my clients' data and my pipeline protected?
Yes. Your client database and your sales pipeline are at the heart of your value: our agents run locally within the company or are hosted in France, 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.
Does the agent connect to my CRM and my current tools?
Yes. The agent connects to your existing tools (CRM, email, calendar, document management) and complements them, without forcing a migration. Your teams call on it from Microsoft Teams, Slack or their email, and your clients reach it on WhatsApp Business or your website chat. 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 set out in your agent's quote; the catalogue grows without a surcharge. Only the fees charged by third-party platforms — WhatsApp Business bills per conversation — are passed on to you at cost, without margin, outside the subscription.
Are the follow-ups and emails sent automatically?
No, unless you explicitly decide so for simple cases. By default, the agent prepares the personalised follow-ups and messages, and it is the salesperson who reads them over and approves the send. The agent carries out what you have authorised, the decisions remain yours.
Do you have to be a large organisation to equip yourself?
No. The offer suits an independent consulting practice just as well as a multi-branch service firm.
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 an agent?
A few weeks as a rule, after a 30-minute meeting with our sales team that identifies the use case with the best return, then a phase of design, integration and testing before going live and training your teams.
Available agents
37 AI agents Blue Lemon Agent deploys for this scope
AI customer support agent for small businesses →
Answering your customers' frequent questions; Order, quote & case tracking; Level 1 resolution of everyday requests; Esc…
Tier-1 customer support agent (website) →
FAQ chatbot, human escalation, messaging channels included at no extra cost.
HR agent — employee support, onboarding and HR documents →
HR base, 1–2 integrations.
Purchasing / supplier agent (quotes, follow-ups) →
Supplier base, workflows.
Tier-2 customer support agent (diagnosis) →
Technical base, escalation.
Client accounting support agent (practice) →
Base, multi-client.
Technical documentation and after-sales support (industry) →
Technical and quality documentation; after-sales and customer support.
Document agent (FAQ, knowledge base) →
RAG on 20–50 docs, 1 language, dialogue in Microsoft Teams, Slack, WhatsApp Business or by email.
Customer review reply agent →
Brand tone, moderation.
Web research assistant →
Sourced summaries.
Simple quote generation agent →
Catalogue, PDF template.
Unpaid invoice chasing assistant (email) →
Templates, sequences.
Tier-1 technical support assistant (product) →
Product knowledge base.
Recurring appointment scheduling assistant →
Diary, reminders.
Sales agent (meeting notes, follow-ups, CRM) →
CRM integration, meeting reports.
Document processing agent (OCR, extraction) →
OCR + indexing pipeline.
Logistics / round planning agent →
Constraints, simple optimisation.
Multilingual support agent (2–3 languages) →
Multiple languages, channels.
HR document management agent →
Staff files, compliance.
Business document management agent (large RAG) →
100+ docs, several sources.
Qualification + scoring + multichannel follow-up agent →
Scoring, email/SMS, CRM.
Multilingual international support agent →
Several languages and channels.
Omnichannel customer relations agent →
Email/chat/phone/CRM.
Automated invoice processing agent (P2P) →
OCR + ERP + approval.
Legal document agent (litigation) →
RAG, deadlines.
Advanced technical support agent (diagnosis + fix) →
Technical base, actions.
Logistics planning & optimisation agent →
Constraints, optimisation.
Large-scale customer service multi-agent system →
Omnichannel, high volumes.
Sovereign AI document management platform →
Massive RAG, governance.
Sourced business document search →
Regulatory and legal bases, doctrine.
Forecasting and planning →
Footfall, demand for benefits, resource allocation.
Internal IT support (tier-1 helpdesk) →
Tickets, procedures; tier 1.
Technical support & documentation →
Tier-1/2 technical helpdesk and documentation kept up to date from the code.
Administrative and documentary support in customs →
Strictly in support (administrative). No operational or targeting use.
Level 1 client relations (banking, finance, insurance) →
Level 1 client relations; request qualification and routing; sensitive cases set aside and handed back to the adviser; p…
Firm memory and client reception (law firms) →
Firm memory; client reception and relations, under the lawyer's validation.
Firm memory and client relations (accounting firms) →
Firm memory; client relations, under the accountant's validation.
Related industries
Other service and consulting activities
Let's talk
Let's size up the potential in your service firm
30 minutes to identify the use case with the best return — hosted in France, supervised, with no commitment.