● Business offer — Communication & advertising
The AI visual creation agent: generate and adapt your visuals to your guidelines
Adapting a campaign across ten formats, resizing a banner, adjusting a social post, preparing A/B variants: this visual production work devours creatives' time without, on its own, carrying the idea the client pays for. Your AI agent absorbs those adaptations. Hosted in France — on local inference or an isolated resource — your content, brands and client data stay with you. The art direction keeps the lead.
Hosted in France
Content & brands protected
GDPR & AI Act: governed deployment
Human oversight
Updated on
Deployed in a few weeks
Visual creation assistant · hosted in France
For the “Summer Solidarity” campaign, adapt the master visual across all our social and display formats, respecting the client's guidelines.
Master visual taken up. 11 adaptations generated — Instagram (1:1, 4:5, 9:16 stories), Facebook, LinkedIn, display (300×250, 728×90, 160×600) and web banner.
Guidelines respected: logo in the protected zone, brand palette and typography, legal mention carried over. Crops and safe margins adjusted per format — for review.
⛓ Source · client guidelines + brand templates
Guidelines respected: logo in the protected zone, brand palette and typography, legal mention carried over. Crops and safe margins adjusted per format — for review.
⛓ Source · client guidelines + brand templates
The stories format is a bit crowded at the top, and give me two headline variants for the A/B test.
Stories rebalanced: the title moved below the interface zone, contrast reinforced for mobile legibility. Two headline variants prepared (“This summer, let's act together” / “One gesture, a big impact”).
I am preparing the exported set and the guidelines compliance sheet — the art direction approves before publication.
✎ Action · set ready to export — the creative approves
I am preparing the exported set and the guidelines compliance sheet — the art direction approves before publication.
✎ Action · set ready to export — the creative approves
Local inference · no data outside the EU
Content 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 art direction decidesThe agent assists, the human approves
✦ In brief
For a communication team, an agency or an advertiser, a Blue Lemon Agent agent automates repetitive visual production — generating visuals, multi-format adaptations, resizing, A/B variants — while respecting your brand guidelines. It runs on local inference or is hosted in France: your creations, your brands and your advertising clients' data are never exposed to a foreign service, architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity. Creatives' time is redirected towards the idea, the art direction and the strategy. Live within a few weeks, all inclusive.
100%
hosted in France in the target architecture
0
transfer outside the EU in the target architecture
9
uses ready to deploy on this scope
0
decision taken without human approval
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.
The context
Why visual AI appeals to communication — and why it worries people
Brands demand ever more content, across ever more channels and formats, at a pace manual production can no longer sustain. But entrusting your creations, your guidelines and your advertisers' data to a consumer tool raises questions of intellectual property and confidentiality that few teams can ignore.
! The issue
Creative teams are caught between an exploding demand for content (social, display, print, video, each format in its own dimensions) and production time that cannot keep up. Yet most consumer generative AI tools amount to entrusting your visuals, your guidelines, your briefs and your advertising clients' data to a third party, often hosted outside Europe and subject to the Cloud Act — with vague terms on the ownership of the creations and their reuse for training.
✓ Our answer
Visual AI is only of interest to a communication team if it is sovereign and confidential by design. Local inference or an isolated resource hosted in France, your models and templates staying with you, systematic human oversight, decisions reserved to the art direction: the time saved on production is never paid for in lost intellectual property or confidentiality. The aim is not to replace the creative, but to give them back time for the idea.
The decisive point
Ownership of the creations & advertisers' confidentiality: sovereignty & compliance
A communication team handles brands, confidential campaigns and its clients' data. Here is how the architecture of our agents protects them, campaign by campaign.
Local inference
The agent can run on a machine belonging to the team: no visual and no brief leaves the network, nothing passes through a foreign cloud.
Hosting in France
Otherwise, a dedicated and isolated resource, hosted in France under French law — your creations: processing and access within the European Union targeted by the architecture.
Reduced extraterritorial exposure
Architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity for your creations and briefs: our architecture stays out of the American hyperscalers' reach, even when hosted in Europe.
Your creations belong to you
Your visuals and briefs are never used to train a third-party model: what the agent produces stays your property.
One isolated resource per organisation
No pooling of campaigns: a strictly dedicated environment, with guidelines and advertiser data kept apart.
AI Act: governed deployment
An agent strictly in support; no automatic publication; 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 events logged, their content, their retention period and who may access them are defined for the deployment chosen.
For the most sensitive brands and campaigns, 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.
Demonstration
See the agent at work
5 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.
The company in this demonstration
Fictional companyAubevert — garden centre and outdoor living retailer
- Sector
- Specialist retail — plants, gardening and outdoor living; catalogues, in-store display, social media and display advertising
- Headcount
- 240 staff, including an in-house studio of 5 — 1 art director, 2 designers, 1 production coordinator, 1 project manager
- Market served
- 34 owned stores, 1.1 million visitors a year, a mix of consumers and professional landscapers
- Volume
- 18 campaigns a year, 6 seasonal catalogues, around 1,400 visuals produced annually, 11 standard social and display formats
- Tools in place
- Creative suite, image library of 12,400 files, social scheduler, ad accounts, editorial calendar and CRM — the agent plugs into them, nothing is replaced
- Who decides
- The art director approves everything that ships; the communication manager sets the campaign plan; the legal team rules on rights assignments
- Room for improvement
- Adapting a campaign across 10 formats takes 60 % of the studio's time; a poster requested by a store waits 3 weeks; 31 rights assignments expire within six months and no register tracks them
Aubevert isn't looking to produce worse, but to produce the same faster and give time back to art direction. The agent runs on local inference on a machine at head office and plugs into the image library, the creative suite, the social scheduler, the ad accounts and the CRM: it prepares, the studio approves. The exchanges below cover one quarter, from the reopening of the image library to the campaign review.
This company, its figures and the exchanges that follow were invented for the demonstration. They illustrate a common situation; they describe no real client.
Visual creation assistant · hosted in France
I read nine years of your image library last night — 12,400 files — and drew the working guidelines for each of your six brands.
Brand guidelines are the document that sets the logo, the colours, the typefaces and the layout rules of a brand. Working guidelines are what your published visuals actually apply: the same document, as it survived nine years of production.
The gap I measured, and it is the gap that decides the gain: across your last 340 published visuals, 4 in 10 break at least one point of the guidelines — and 8 times out of 10 it is the same point: the logo clear space. Clear space is the mandatory empty area around the logo that no other element may enter. This is not carelessness on your designers' part: it is written in millimetres in a 2019 document, and social formats did not exist at those sizes.
What each set of working guidelines contains, all of it checkable visual by visual: clear space expressed as a proportion of the logo's height rather than in millimetres, the palette with its exact values, the typefaces and their licence, the position of the legal notice, and the safe margins per support — the band at the edge of a format where nothing important may fall, on pain of being trimmed off in print.
· Aubevert retail brand: logo top left, brand green on light backgrounds only, never a photograph behind the logo — a rule held in 96 % of cases and written down nowhere.
· Aubevert Pro (landscapers): darker palette, price shown excluding VAT, always — a constraint of your professional clientele.
· The two partner brands: their own guidelines, with what your co-branding contract allows, and what it leaves unsaid.
What that gives you tomorrow: an adaptation arrives ALREADY compliant. Your designers stop re-fixing a logo and start arbitrating an image — and that is where your studio shows.
What I suggest: the art director reads the six sets of working guidelines, and I apply them from the next campaign on. Each one can be corrected one word at a time: it is your material, I have only set it down clearly. working-guidelines_6-brands.pdfWhat your visuals actually apply
⛓ Sourced · studio image library, 12,400 files over 9 years, last 340 published visuals, 2019 and 2023 guidelines
Brand guidelines are the document that sets the logo, the colours, the typefaces and the layout rules of a brand. Working guidelines are what your published visuals actually apply: the same document, as it survived nine years of production.
The gap I measured, and it is the gap that decides the gain: across your last 340 published visuals, 4 in 10 break at least one point of the guidelines — and 8 times out of 10 it is the same point: the logo clear space. Clear space is the mandatory empty area around the logo that no other element may enter. This is not carelessness on your designers' part: it is written in millimetres in a 2019 document, and social formats did not exist at those sizes.
What each set of working guidelines contains, all of it checkable visual by visual: clear space expressed as a proportion of the logo's height rather than in millimetres, the palette with its exact values, the typefaces and their licence, the position of the legal notice, and the safe margins per support — the band at the edge of a format where nothing important may fall, on pain of being trimmed off in print.
· Aubevert retail brand: logo top left, brand green on light backgrounds only, never a photograph behind the logo — a rule held in 96 % of cases and written down nowhere.
· Aubevert Pro (landscapers): darker palette, price shown excluding VAT, always — a constraint of your professional clientele.
· The two partner brands: their own guidelines, with what your co-branding contract allows, and what it leaves unsaid.
What that gives you tomorrow: an adaptation arrives ALREADY compliant. Your designers stop re-fixing a logo and start arbitrating an image — and that is where your studio shows.
What I suggest: the art director reads the six sets of working guidelines, and I apply them from the next campaign on. Each one can be corrected one word at a time: it is your material, I have only set it down clearly. working-guidelines_6-brands.pdfWhat your visuals actually apply
⛓ Sourced · studio image library, 12,400 files over 9 years, last 340 published visuals, 2019 and 2023 guidelines
We're looking for the Spring 2024 catalogue template, the one before the format change. Nobody can find it.
Here it is, with the four documents that belong with it — and I'll tell you which one governs.
· Spring 2024 Catalogue template, filed on 08/01/2024, the last version before the move to the 21 × 27 format. That is the one you want.
· The campaign brief explaining the format change and its reason — useful if you have to justify it in committee.
· The 4 approved sample pages that came out of it, which serve as the layout reference.
· And 17 visuals derived from that template still on display in stores. Those deserve a decision: they carry the old logo, the one from before the 2023 refresh.
How I found it when its name says nothing — it is called AUB_cata_PRT_ok_final_v3.indd: I read the contents, not the name. The logo, the palette, the page format and the filing date are enough to identify it.
What that search was costing you: your own activity records show 28 hours of “looking for a file” over the quarter, which is more than three studio days for five people. The same document now comes back in a sentence.
What I suggest: a monthly pass over recent uploads to set the brand, the campaign, the season and the version. Of the 12,400 files, 3,900 have no usable name — and those are the ones you hunt for on a Friday evening before a print approval. Indexing renames nothing and moves nothing: your existing paths keep working. studio-memory_what-was-recovered.pdf3,900 files with no usable name, indexed
⛓ Sourced · image library, file contents rather than file names, studio activity records
· Spring 2024 Catalogue template, filed on 08/01/2024, the last version before the move to the 21 × 27 format. That is the one you want.
· The campaign brief explaining the format change and its reason — useful if you have to justify it in committee.
· The 4 approved sample pages that came out of it, which serve as the layout reference.
· And 17 visuals derived from that template still on display in stores. Those deserve a decision: they carry the old logo, the one from before the 2023 refresh.
How I found it when its name says nothing — it is called AUB_cata_PRT_ok_final_v3.indd: I read the contents, not the name. The logo, the palette, the page format and the filing date are enough to identify it.
What that search was costing you: your own activity records show 28 hours of “looking for a file” over the quarter, which is more than three studio days for five people. The same document now comes back in a sentence.
What I suggest: a monthly pass over recent uploads to set the brand, the campaign, the season and the version. Of the 12,400 files, 3,900 have no usable name — and those are the ones you hunt for on a Friday evening before a print approval. Indexing renames nothing and moves nothing: your existing paths keep working. studio-memory_what-was-recovered.pdf3,900 files with no usable name, indexed
⛓ Sourced · image library, file contents rather than file names, studio activity records
All of this means you read our briefs, our unpublished visuals and our partner brands' contracts. Where exactly do they go?
They do not leave your walls. I run on local inference on a machine at head office, and your image library is copied nowhere.
Local inference means the computation happens on your own machine: an unpublished visual has no outside network to cross in order to be processed. If you would rather not run a machine, the other route is an isolated resource hosted in France, under French law, dedicated to Aubevert — no pooling with another retailer, or with another brand.
What that changes, concretely:
· Your visuals and briefs train no public model. What I learn from your guidelines serves your guidelines. What the agent produces stays your property, and nothing you entrust to me surfaces anywhere else.
· Encryption in transit and at rest, and role-based access — rights follow the job: the studio's apprentice opens the social templates, not the co-branding contracts or the media plans.
· A complete log: who asked for what, when, what was approved and by whom. That log is what lets you answer a partner brand that asks.
· Architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity, including against an American host based in Europe.
And you change no tools: I plug into your creative suite, your image library, your social scheduler, your ad accounts, your approval space and your editorial calendar. No migration, no extra subscription for your teams.
The figure that makes this commercial rather than technical: both of your co-branding contracts already require that undisclosed visuals stay inside the European Union, and the last buying-group listing you applied for scored it as a 5-point criterion. You were able to tick the box without reservation, with evidence.
What I suggest: that I keep up to date the technical sheet your partners ask for at renewal — hosting, subcontractors, retention periods, who has access to what. It is asked for once a year and hunted for over three days. technical-framework_where-your-visuals-live.pdfLocal inference, role-based access, processing in the EU targeted
✎ Framework · deployment architecture, access log, co-branding contracts, buying-group scoring criterion
Local inference means the computation happens on your own machine: an unpublished visual has no outside network to cross in order to be processed. If you would rather not run a machine, the other route is an isolated resource hosted in France, under French law, dedicated to Aubevert — no pooling with another retailer, or with another brand.
What that changes, concretely:
· Your visuals and briefs train no public model. What I learn from your guidelines serves your guidelines. What the agent produces stays your property, and nothing you entrust to me surfaces anywhere else.
· Encryption in transit and at rest, and role-based access — rights follow the job: the studio's apprentice opens the social templates, not the co-branding contracts or the media plans.
· A complete log: who asked for what, when, what was approved and by whom. That log is what lets you answer a partner brand that asks.
· Architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity, including against an American host based in Europe.
And you change no tools: I plug into your creative suite, your image library, your social scheduler, your ad accounts, your approval space and your editorial calendar. No migration, no extra subscription for your teams.
The figure that makes this commercial rather than technical: both of your co-branding contracts already require that undisclosed visuals stay inside the European Union, and the last buying-group listing you applied for scored it as a 5-point criterion. You were able to tick the box without reservation, with evidence.
What I suggest: that I keep up to date the technical sheet your partners ask for at renewal — hosting, subcontractors, retention periods, who has access to what. It is asked for once a year and hunted for over three days. technical-framework_where-your-visuals-live.pdfLocal inference, role-based access, processing in the EU targeted
✎ Framework · deployment architecture, access log, co-branding contracts, buying-group scoring criterion
For the “Solidarity Summer” campaign, adapt the main visual across all our social and display formats, respecting the client's guidelines.
Master visual taken up. 11 adaptations generated, ready for review, all of them drawn from your templates — not invented by me.
The master visual is a campaign's reference image: the one every size derives from. An adaptation is the same creative work in another format, with the message unchanged.
The eleven: Instagram in 1:1, 4:5 and 9:16 stories · Facebook · LinkedIn · display in 300×250, 728×90 and 160×600 · homepage web banner · email header · share thumbnail.
Guidelines held, point by point: logo in its clear space, brand palette and typography, legal notice carried onto all eleven — that is the one people forget, and the only one visible from outside. Crops and safe margins adjusted format by format: a 160×600 is not treated like a 728×90, one reads top to bottom, the other left to right.
The time this shifts: adapting a campaign across 10 formats took 60 % of your studio's time; it takes 18 %. The rest is your review, and that cannot be delegated. Across 18 campaigns a year, your studio gets the days back, not your budget paying for them.
What I set aside rather than force: at 160×600, the full headline falls inside the safe margin. I did not shorten it on my own authority: I have prepared two possible crops and one short headline, and the art director settles it in thirty seconds.
What I prepare while you review: the export set in the right colour profiles and the guidelines compliance sheet, format by format — the one the partner brand will ask for. solidarity-summer_11-adaptations.pdfOne master, eleven formats, one compliance sheet
⛓ Sourced · campaign master visual, Aubevert working guidelines, brand templates, technical constraints of the 11 supports
The master visual is a campaign's reference image: the one every size derives from. An adaptation is the same creative work in another format, with the message unchanged.
The eleven: Instagram in 1:1, 4:5 and 9:16 stories · Facebook · LinkedIn · display in 300×250, 728×90 and 160×600 · homepage web banner · email header · share thumbnail.
Guidelines held, point by point: logo in its clear space, brand palette and typography, legal notice carried onto all eleven — that is the one people forget, and the only one visible from outside. Crops and safe margins adjusted format by format: a 160×600 is not treated like a 728×90, one reads top to bottom, the other left to right.
The time this shifts: adapting a campaign across 10 formats took 60 % of your studio's time; it takes 18 %. The rest is your review, and that cannot be delegated. Across 18 campaigns a year, your studio gets the days back, not your budget paying for them.
What I set aside rather than force: at 160×600, the full headline falls inside the safe margin. I did not shorten it on my own authority: I have prepared two possible crops and one short headline, and the art director settles it in thirty seconds.
What I prepare while you review: the export set in the right colour profiles and the guidelines compliance sheet, format by format — the one the partner brand will ask for. solidarity-summer_11-adaptations.pdfOne master, eleven formats, one compliance sheet
⛓ Sourced · campaign master visual, Aubevert working guidelines, brand templates, technical constraints of the 11 supports
The stories format feels a bit heavy at the top, and give me two headline variants for the A/B test.
Stories rebalanced, and the two variants are ready — built to be comparable, not merely different.
On the stories: title moved below the interface zone — the top and bottom bands where the social network places its own buttons, covering the image without warning —, contrast raised for reading on a phone in bright sun, logo lifted back into its clear space. The visual has not changed; its hierarchy has.
The two headlines: “This summer, let's act together” and “One small act, one big difference”. Same length, same call to action, same visual: if one wins, it is the headline that wins, not a better layout.
An A/B test runs two versions at the same time to two halves of the audience, everything else equal, so that the measured gap comes only from what changes.
The time this shifts, on your two most repetitive tasks: resizing a visual for the networks went from 32 % to 8 % of campaign time; preparing A/B variants of an ad, from 18 % to 6 %. Those are the two jobs nobody ever wanted to do by hand.
What I suggest next: run for five days, then move the losing headline's budget onto the winner — and I am already writing two new headlines in the winning angle, so the next test is ready the day this one ends. The export set and the compliance sheet are waiting for you; the art director approves before anything ships. ab-headlines_two-comparable-variants.pdfTest protocol and export set
⛓ Sourced · social templates, network interface zones, 12 months of Aubevert headline history
On the stories: title moved below the interface zone — the top and bottom bands where the social network places its own buttons, covering the image without warning —, contrast raised for reading on a phone in bright sun, logo lifted back into its clear space. The visual has not changed; its hierarchy has.
The two headlines: “This summer, let's act together” and “One small act, one big difference”. Same length, same call to action, same visual: if one wins, it is the headline that wins, not a better layout.
An A/B test runs two versions at the same time to two halves of the audience, everything else equal, so that the measured gap comes only from what changes.
The time this shifts, on your two most repetitive tasks: resizing a visual for the networks went from 32 % to 8 % of campaign time; preparing A/B variants of an ad, from 18 % to 6 %. Those are the two jobs nobody ever wanted to do by hand.
What I suggest next: run for five days, then move the losing headline's budget onto the winner — and I am already writing two new headlines in the winning angle, so the next test is ready the day this one ends. The export set and the compliance sheet are waiting for you; the art director approves before anything ships. ab-headlines_two-comparable-variants.pdfTest protocol and export set
⛓ Sourced · social templates, network interface zones, 12 months of Aubevert headline history
Our 34 stores each want their own local poster, with their opening hours and their offer. It takes three weeks and it clogs the studio.
The 34 posters are ready. Forty minutes, one template, and not a single point of the guidelines lost on the way.
What I did: I took the store poster template approved by your art director and fed it from your store reference data — store name, address, seasonal opening hours, local offer, campaign code. Every poster is the same creative work; only the variable fields change.
The three traps I handled, and they are what costs the three weeks:
· Long store names. “Aubevert Saint-Genis-Laval Centre” does not fit on one line at the intended size. I have a rule that wraps to two lines rather than shrinking the type — shrinking the type breaks the hierarchy on all 34 at once.
· Summer hours that differ for 11 stores. They come from the reference data, dated as at the export. A wrong opening time on a poster is a customer standing at a closed door.
· Three stores had no local offer entered. I invented none and left no blank: their posters came out in the no-panel version, which already exists in your template. The three managers received the request, with the deadline for the next print run.
The delay this changes: a requested local poster waited 3 weeks. The full set comes out in 40 minutes, art director review included — and a store that corrects its opening hours at 5 p.m. has an updated poster at 5:05.
What I suggest next: open the same template to the 6 seasonal campaigns and pre-fill each poster 15 days before the print run, so stores correct beforehand rather than afterwards. You set the cut-off date; I start no print run. local-posters_34-stores.pdfA three-week wait brought down to 40 minutes
⛓ Sourced · reference data for the 34 stores, approved poster template, calendar of the 6 seasonal campaigns
What I did: I took the store poster template approved by your art director and fed it from your store reference data — store name, address, seasonal opening hours, local offer, campaign code. Every poster is the same creative work; only the variable fields change.
The three traps I handled, and they are what costs the three weeks:
· Long store names. “Aubevert Saint-Genis-Laval Centre” does not fit on one line at the intended size. I have a rule that wraps to two lines rather than shrinking the type — shrinking the type breaks the hierarchy on all 34 at once.
· Summer hours that differ for 11 stores. They come from the reference data, dated as at the export. A wrong opening time on a poster is a customer standing at a closed door.
· Three stores had no local offer entered. I invented none and left no blank: their posters came out in the no-panel version, which already exists in your template. The three managers received the request, with the deadline for the next print run.
The delay this changes: a requested local poster waited 3 weeks. The full set comes out in 40 minutes, art director review included — and a store that corrects its opening hours at 5 p.m. has an updated poster at 5:05.
What I suggest next: open the same template to the 6 seasonal campaigns and pre-fill each poster 15 days before the print run, so stores correct beforehand rather than afterwards. You set the cut-off date; I start no print run. local-posters_34-stores.pdfA three-week wait brought down to 40 minutes
⛓ Sourced · reference data for the 34 stores, approved poster template, calendar of the 6 seasonal campaigns
We've already been caught out by a photo that was lying around. That's our real fear before every campaign.
Every visual in the campaign now carries its origin, its licence, its territory and its end date — and three expire before the campaign ends. I have already drafted the three letters.
A rights assignment is the photographer's or illustrator's written authorisation: it is bounded by supports, a territory and a duration. Past that duration the image must come off the supports, even if the file is still in your library.
What I check before adapting anything, every time: the file's origin, the type of licence, the supports assigned, the territory, and the end date compared with the planned run.
The three cases in this campaign:
· Two photographs from the 2023 “nursery” shoot: three-year assignment, ending 30/09/2026, and the campaign runs into November. An extension has to be requested from the photographer. The letter is written, with the exact supports, territory and requested duration — it goes out the moment you sign it, and I have looked up what his last extension cost so you know what to expect.
· One stock image whose licence covers web but excludes large-format display. The campaign plans 34 posters, one per store. Two ways out, and both are open: upgrade the licence, for which I have the price and lead time, or take one of the four alternatives from your own library that I have pulled — same crop, same green cast, unlimited rights and zero euros.
What this changes in your planning: these checks were done by hand, campaign by campaign, and only when somebody thought of it. Across your last six campaigns, two ran with at least one out-of-term image — with no consequence, but with no knowledge either.
What I suggest: that I keep the assignment register across the whole library. There are 31 expiries in the next six months, 9 of them on supports still running. You set the warning period — I recommend 90 days, the average reply time of your photographers — and I bring them to you with the letter already written. An expiry does not stop a campaign: it gets prepared for. rights-register_31-expiries.pdf3 extension letters ready to sign
⛓ Sourced · image library, assignment contracts, stock image licences, last 6 campaigns
A rights assignment is the photographer's or illustrator's written authorisation: it is bounded by supports, a territory and a duration. Past that duration the image must come off the supports, even if the file is still in your library.
What I check before adapting anything, every time: the file's origin, the type of licence, the supports assigned, the territory, and the end date compared with the planned run.
The three cases in this campaign:
· Two photographs from the 2023 “nursery” shoot: three-year assignment, ending 30/09/2026, and the campaign runs into November. An extension has to be requested from the photographer. The letter is written, with the exact supports, territory and requested duration — it goes out the moment you sign it, and I have looked up what his last extension cost so you know what to expect.
· One stock image whose licence covers web but excludes large-format display. The campaign plans 34 posters, one per store. Two ways out, and both are open: upgrade the licence, for which I have the price and lead time, or take one of the four alternatives from your own library that I have pulled — same crop, same green cast, unlimited rights and zero euros.
What this changes in your planning: these checks were done by hand, campaign by campaign, and only when somebody thought of it. Across your last six campaigns, two ran with at least one out-of-term image — with no consequence, but with no knowledge either.
What I suggest: that I keep the assignment register across the whole library. There are 31 expiries in the next six months, 9 of them on supports still running. You set the warning period — I recommend 90 days, the average reply time of your photographers — and I bring them to you with the letter already written. An expiry does not stop a campaign: it gets prepared for. rights-register_31-expiries.pdf3 extension letters ready to sign
⛓ Sourced · image library, assignment contracts, stock image licences, last 6 campaigns
And if we ask you for an image that exists nowhere? A shot we've never taken.
I produce it, and your art director decides whether it stays. Three routes, from cheapest to longest.
· Re-cropping and retouching your own library — often the shot exists, badly framed. Of the seven missing shots in this campaign, three were already with you, in a 2022 shoot nobody had reopened.
· Compositing from your own elements: product cut out, setting taken from your shoot, lighting matched. Your rights, your material, no new licence.
· The synthetic image — an image built by a model, coming from no camera — when neither of the first two routes yields the shot. I did this for the remaining four shots: 40 minutes instead of a day of shooting. The art director kept two, rejected one, and commissioned a photograph for the last — which is exactly the split you want.
What every synthetic image carries with it: its provenance note in the file's metadata, the prompt that produced it, the date and the tool — enough to reproduce it identically in six months, when a winter variant is needed. An image you cannot reproduce is an image you will pay for twice.
Why that note serves you beyond principle: both of your co-branding contracts already require it, and the European AI regulation provides in time for the marking of machine-produced content. You are ahead of a requirement your partners will put to you, and you can show it in writing.
What it opens commercially: testing a visual angle no longer costs a production. You can present six routes in the campaign committee where you showed two, and the shoot is commissioned on the one that was chosen, not on the one you had to guess. seven-missing-shots_three-routes.pdf40 minutes against a full day of shooting
⛓ Sourced · image library, 2022 shoot, generation log, co-branding contracts
· Re-cropping and retouching your own library — often the shot exists, badly framed. Of the seven missing shots in this campaign, three were already with you, in a 2022 shoot nobody had reopened.
· Compositing from your own elements: product cut out, setting taken from your shoot, lighting matched. Your rights, your material, no new licence.
· The synthetic image — an image built by a model, coming from no camera — when neither of the first two routes yields the shot. I did this for the remaining four shots: 40 minutes instead of a day of shooting. The art director kept two, rejected one, and commissioned a photograph for the last — which is exactly the split you want.
What every synthetic image carries with it: its provenance note in the file's metadata, the prompt that produced it, the date and the tool — enough to reproduce it identically in six months, when a winter variant is needed. An image you cannot reproduce is an image you will pay for twice.
Why that note serves you beyond principle: both of your co-branding contracts already require it, and the European AI regulation provides in time for the marking of machine-produced content. You are ahead of a requirement your partners will put to you, and you can show it in writing.
What it opens commercially: testing a visual angle no longer costs a production. You can present six routes in the campaign committee where you showed two, and the shoot is commissioned on the one that was chosen, not on the one you had to guess. seven-missing-shots_three-routes.pdf40 minutes against a full day of shooting
⛓ Sourced · image library, 2022 shoot, generation log, co-branding contracts
The models from the 2023 campaign — can we use them again this summer? And our typefaces: we don't even know what we're allowed to do with them.
Two more registers, kept in the same place as the photographs — and both answers are yes, on conditions I give you with their route.
The people photographed, first.
An image release is the agreement signed by the person appearing in the photograph: it too is bounded — by supports, a territory and a duration, often shorter than the photographer's assignment.
· Of the 6 people in the 2023 campaign, 4 can be used straight away: five-year release, all supports, France. Nothing to do.
· One release covers web and excludes print. And you want the store poster. I have drafted the amendment — same wording, extension to printed supports, duration aligned on the photographer's assignment so that both expire on the same day. Two signatures, and the strongest shot in the campaign is usable again.
· One release expires in March. I am telling you now, not in March: the person is a customer of one of your stores, and it is better to ask before than after.
Typefaces next, and this is the subject nobody raises.
A font licence is the contract permitting the use of a typeface: it is almost always limited — number of workstations, monthly page views for web, sometimes advertising use excluded.
· Your retail brand typeface: desktop licence for 5 seats. There are 5 of you in the studio, plus 2 partner agencies using it. That is the only genuine gap in your estate, and it is settled by an extension to 12 seats for which I have the price.
· Your web display typeface: licence for 250,000 monthly page views. You have been at 310,000 since the spring campaign. The next tier is 500,000: I suggest taking it before peak season, not during.
What this gives you back: each of your last six campaigns took half a day of rights checking, done under pressure and never complete. The register is now kept continuously, and an alert arrives with its solution, not with a problem. You decide to extend, replace or drop; I remove no visual on my own. releases-and-fonts_what-is-covered.pdfOne amendment drafted, two licence tiers to anticipate
⛓ Sourced · 2023 image releases, font licence contracts, website traffic measurement
The people photographed, first.
An image release is the agreement signed by the person appearing in the photograph: it too is bounded — by supports, a territory and a duration, often shorter than the photographer's assignment.
· Of the 6 people in the 2023 campaign, 4 can be used straight away: five-year release, all supports, France. Nothing to do.
· One release covers web and excludes print. And you want the store poster. I have drafted the amendment — same wording, extension to printed supports, duration aligned on the photographer's assignment so that both expire on the same day. Two signatures, and the strongest shot in the campaign is usable again.
· One release expires in March. I am telling you now, not in March: the person is a customer of one of your stores, and it is better to ask before than after.
Typefaces next, and this is the subject nobody raises.
A font licence is the contract permitting the use of a typeface: it is almost always limited — number of workstations, monthly page views for web, sometimes advertising use excluded.
· Your retail brand typeface: desktop licence for 5 seats. There are 5 of you in the studio, plus 2 partner agencies using it. That is the only genuine gap in your estate, and it is settled by an extension to 12 seats for which I have the price.
· Your web display typeface: licence for 250,000 monthly page views. You have been at 310,000 since the spring campaign. The next tier is 500,000: I suggest taking it before peak season, not during.
What this gives you back: each of your last six campaigns took half a day of rights checking, done under pressure and never complete. The register is now kept continuously, and an alert arrives with its solution, not with a problem. You decide to extend, replace or drop; I remove no visual on my own. releases-and-fonts_what-is-covered.pdfOne amendment drafted, two licence tiers to anticipate
⛓ Sourced · 2023 image releases, font licence contracts, website traffic measurement
The visuals are ready, but we still write every caption and headline by hand, the night before.
Forty-eight pieces ready for review, written for the visuals you have just approved — and I'll give you the order to take them in.
The core: a 600-word campaign text carrying the “Solidarity Summer” argument. Everything else follows from it: if you change the argument, I carry it through the other forty-seven without you having to ask again.
The derivatives, each in the brand's voice:
· 18 social captions, one per adapted visual, written from the image rather than beside it — the caption says what the image does not, it does not repeat it.
· 12 ad headlines, three angles × four formats, built comparable for the test.
· 2 emails, one announcement, one follow-up to non-openers after fifteen days.
· 34 store poster panels, one per store, with the local offer.
· The legal notices carried over and checked on all 48.
What I take from your own history: your posts opening with a question earn 1.7 times the engagement of the rest over your last twelve months. So six captions out of eighteen open with a question — not eighteen: a device applied systematically stops being a device.
What I did not write, and left highlighted: the percentage donated to the partner charity. The brief states it, no signed document confirms it. Have it confirmed before publication: that is the kind of figure that ends up in the local press.
Approval order I suggest: the campaign text, then the 12 headlines — they are what commits media budget. The captions can wait until tomorrow morning. solidarity-summer_48-pieces.pdfOne core, five derivatives, one highlight
⛓ Sourced · campaign brief, working guidelines, engagement by post format over 12 months
The core: a 600-word campaign text carrying the “Solidarity Summer” argument. Everything else follows from it: if you change the argument, I carry it through the other forty-seven without you having to ask again.
The derivatives, each in the brand's voice:
· 18 social captions, one per adapted visual, written from the image rather than beside it — the caption says what the image does not, it does not repeat it.
· 12 ad headlines, three angles × four formats, built comparable for the test.
· 2 emails, one announcement, one follow-up to non-openers after fifteen days.
· 34 store poster panels, one per store, with the local offer.
· The legal notices carried over and checked on all 48.
What I take from your own history: your posts opening with a question earn 1.7 times the engagement of the rest over your last twelve months. So six captions out of eighteen open with a question — not eighteen: a device applied systematically stops being a device.
What I did not write, and left highlighted: the percentage donated to the partner charity. The brief states it, no signed document confirms it. Have it confirmed before publication: that is the kind of figure that ends up in the local press.
Approval order I suggest: the campaign text, then the 12 headlines — they are what commits media budget. The captions can wait until tomorrow morning. solidarity-summer_48-pieces.pdfOne core, five derivatives, one highlight
⛓ Sourced · campaign brief, working guidelines, engagement by post format over 12 months
And the social calendar? We post when we think of it, which is to say not often enough.
Six weeks of editorial calendar are built, network by network, ready to schedule. Nothing is scheduled until you say so.
The editorial calendar is the plan of what gets published, on which network and on which day.
What I paced, and on what evidence:
· Instagram — 14 posts, 6 of them stories: Tuesday and Thursday 6:30 p.m., the two slots carrying 2.4 times your account's average engagement over twelve months.
· Facebook — 9 posts, offer- and store-led, Saturday morning: that is when your audience plans its day.
· LinkedIn — 5 posts, Aubevert Pro angle, expert tone, never a promotional offer. Your landscapers do not come there for a price.
· Two peaks: the launch, and the announcement of the amount donated at the end of the campaign. The second is the one that gets forgotten, and it is the one that brings people back.
Every slot in the calendar carries its visual, already adapted to the right format, and its caption: you open the scheduler, you review, you schedule. There is no longer a production step between the decision and the post.
What that changes, measured at your end: your 34 stores each posted on their own, with 6 off-brand visuals last quarter. The calendar gives them ready material: they post more, and they post on brand.
What I suggest: that I prepare the following month's calendar on the 20th, visuals already adapted, leaving you only the arbitration. I schedule nothing: it is your scheduler, your credentials, your decision. editorial-calendar_6-weeks.pdf28 posts paced, ready to schedule
⛓ Sourced · engagement by time slot over 12 months, posts from the 34 stores, campaign calendar
The editorial calendar is the plan of what gets published, on which network and on which day.
What I paced, and on what evidence:
· Instagram — 14 posts, 6 of them stories: Tuesday and Thursday 6:30 p.m., the two slots carrying 2.4 times your account's average engagement over twelve months.
· Facebook — 9 posts, offer- and store-led, Saturday morning: that is when your audience plans its day.
· LinkedIn — 5 posts, Aubevert Pro angle, expert tone, never a promotional offer. Your landscapers do not come there for a price.
· Two peaks: the launch, and the announcement of the amount donated at the end of the campaign. The second is the one that gets forgotten, and it is the one that brings people back.
Every slot in the calendar carries its visual, already adapted to the right format, and its caption: you open the scheduler, you review, you schedule. There is no longer a production step between the decision and the post.
What that changes, measured at your end: your 34 stores each posted on their own, with 6 off-brand visuals last quarter. The calendar gives them ready material: they post more, and they post on brand.
What I suggest: that I prepare the following month's calendar on the 20th, visuals already adapted, leaving you only the arbitration. I schedule nothing: it is your scheduler, your credentials, your decision. editorial-calendar_6-weeks.pdf28 posts paced, ready to schedule
⛓ Sourced · engagement by time slot over 12 months, posts from the 34 stores, campaign calendar
Our search ads show visuals that have nothing to do with what we're promoting. Nobody looks after them.
The twelve ads are realigned on the campaign, visuals included, and ready for testing. Nothing is switched on.
Organic search (SEO) is what lifts a page in a search engine's free results; paid search (SEA) is the ads on the same engine. Both are worked on, and not in the same place.
What I found, and it explains the gap: 7 of your 12 ads still show the autumn campaign visual, four months after it ended. Someone clicking a summer promise lands on an autumn image — and leaves. Your bounce rate on those ads is 71 %, against 44 % on the other five.
What I have prepared:
· 12 ad visuals in the formats the platform requires, drawn from the approved master visual, clear space and legal notice held.
· 3 angles × 4 formats so the test decides the angle, not the layout.
· Titles and descriptions aligned with the visual: the promise in the ad and the image on the landing page say the same thing, which is the only optimisation that holds over time.
· Nine negative keywords — the searches your ad must on no account appear on: “free”, “second-hand”, “artificial plant”, “jobs”… They cost you €410 in wasted clicks last month.
On organic search: the campaign page is written for “eco-friendly garden centre”, 1,900 searches a month, where you do not appear in the first two pages. Three secondary queries are carried by the same page. What I advise against, and this is where six months get lost: targeting “garden centre”, where the top three results are national chains with twenty times your authority. Writing for that query is writing for page four.
The time this shifts: preparing A/B variants of an ad took 18 % of campaign time; it takes 6 %.
What I suggest: you approve the twelve, I switch on at your signal, and I bring you the first reading at day 3 with the first ad to cut. search-ads_12-realigned.pdf7 stale visuals replaced, 9 negative keywords
⛓ Sourced · ad accounts, bounce rate by ad, triggering queries, search volumes
Organic search (SEO) is what lifts a page in a search engine's free results; paid search (SEA) is the ads on the same engine. Both are worked on, and not in the same place.
What I found, and it explains the gap: 7 of your 12 ads still show the autumn campaign visual, four months after it ended. Someone clicking a summer promise lands on an autumn image — and leaves. Your bounce rate on those ads is 71 %, against 44 % on the other five.
What I have prepared:
· 12 ad visuals in the formats the platform requires, drawn from the approved master visual, clear space and legal notice held.
· 3 angles × 4 formats so the test decides the angle, not the layout.
· Titles and descriptions aligned with the visual: the promise in the ad and the image on the landing page say the same thing, which is the only optimisation that holds over time.
· Nine negative keywords — the searches your ad must on no account appear on: “free”, “second-hand”, “artificial plant”, “jobs”… They cost you €410 in wasted clicks last month.
On organic search: the campaign page is written for “eco-friendly garden centre”, 1,900 searches a month, where you do not appear in the first two pages. Three secondary queries are carried by the same page. What I advise against, and this is where six months get lost: targeting “garden centre”, where the top three results are national chains with twenty times your authority. Writing for that query is writing for page four.
The time this shifts: preparing A/B variants of an ad took 18 % of campaign time; it takes 6 %.
What I suggest: you approve the twelve, I switch on at your signal, and I bring you the first reading at day 3 with the first ad to cut. search-ads_12-realigned.pdf7 stale visuals replaced, 9 negative keywords
⛓ Sourced · ad accounts, bounce rate by ad, triggering queries, search volumes
After a quarter, what has it actually given us? And did the campaign work?
Seven person-days a month given back to the studio — seven working days of one person, roughly a third of your creative headcount — and the campaign review has been ready since Sunday evening.
On the three measures you tracked:
· Adapting a campaign across 10 formats: 60 % → 18 % of the time. The biggest gain, and the most mechanical.
· Resizing a visual for the networks: 32 % → 8 %.
· Preparing A/B variants of an ad: 18 % → 6 %.
What the campaign did, format by format — this is the reading you did not have: Instagram 4:5 carries 41 % of interactions for 9 % of impressions; the 728×90 carries 3 % for 22 % of display budget. The creative is not at fault, the format is: I suggest dropping the 728×90 from the plan and moving its 22 % onto the 300×250, which performs five times better at the same cost.
On the A/B test: “One small act, one big difference” earned 1.6 times the clicks of the other headline over 180,000 impressions. The gap is clear and the sample sufficient: the test has decided. The next two headlines are written in that angle.
What those hours became, from your own records: +38 % on the “art direction and concept” line, and two seasonal catalogues produced in house instead of being outsourced. That is the only figure on this list that shows up in your budget, and it comes from time given back, not from one more tool.
The figure that does not flatter me: of 980 visuals prepared this quarter, 138 were reworked by more than half by the art director — 14 %. Two brands account for 96 of those 138: the two partner brands, for which I have only 40 visuals of history each, against 4,100 for the retail brand. My working guidelines are thin where your material is thin.
What I do about it: give me ten approved visuals per partner brand and I rebuild both sets of guidelines. On Aubevert Pro, where you have already done it, rework fell from 29 % to 8 % in five weeks — it is the same correction, applied to the last two.
And the measure you did not have: zero visuals published without human approval over the quarter, across 980 logged approvals. That is not a setting, it is how I am wired in. quarter-review_what-the-studio-won-back.pdf7 person-days a month, 14 % rework explained
⛓ Sourced · studio activity records, approval log, ad accounts, successive versions of the visuals
On the three measures you tracked:
· Adapting a campaign across 10 formats: 60 % → 18 % of the time. The biggest gain, and the most mechanical.
· Resizing a visual for the networks: 32 % → 8 %.
· Preparing A/B variants of an ad: 18 % → 6 %.
What the campaign did, format by format — this is the reading you did not have: Instagram 4:5 carries 41 % of interactions for 9 % of impressions; the 728×90 carries 3 % for 22 % of display budget. The creative is not at fault, the format is: I suggest dropping the 728×90 from the plan and moving its 22 % onto the 300×250, which performs five times better at the same cost.
On the A/B test: “One small act, one big difference” earned 1.6 times the clicks of the other headline over 180,000 impressions. The gap is clear and the sample sufficient: the test has decided. The next two headlines are written in that angle.
What those hours became, from your own records: +38 % on the “art direction and concept” line, and two seasonal catalogues produced in house instead of being outsourced. That is the only figure on this list that shows up in your budget, and it comes from time given back, not from one more tool.
The figure that does not flatter me: of 980 visuals prepared this quarter, 138 were reworked by more than half by the art director — 14 %. Two brands account for 96 of those 138: the two partner brands, for which I have only 40 visuals of history each, against 4,100 for the retail brand. My working guidelines are thin where your material is thin.
What I do about it: give me ten approved visuals per partner brand and I rebuild both sets of guidelines. On Aubevert Pro, where you have already done it, rework fell from 29 % to 8 % in five weeks — it is the same correction, applied to the last two.
And the measure you did not have: zero visuals published without human approval over the quarter, across 980 logged approvals. That is not a setting, it is how I am wired in. quarter-review_what-the-studio-won-back.pdf7 person-days a month, 14 % rework explained
⛓ Sourced · studio activity records, approval log, ad accounts, successive versions of the visuals
Our 34 store managers chase us constantly to know where their visual is. Our production coordinator takes all of it.
Ninety-two “where is my artwork?” requests a month. I answer 78 within ten minutes with the real state of the schedule; 14 go up to your project manager.
What I answer, and where it comes from: the current step, who holds it, the expected return date, and the direct link to the proof awaiting approval. I read the production schedule, not a promise: if a job has slipped, the store hears it from me before discovering it.
The 14 I pass on: four requests outside the campaign plan, three priority calls between stores, seven where the honest answer is “we need to talk”. Committing to a date is not information, it is a project manager's decision.
What needs attention this week: three proofs have been awaiting store approval for 9 days, and the print run is in six. I have chased twice, with the direct link. I flag it because it is the studio that will be held responsible for the delay.
On account tracking, what I keep current in your CRM with nobody entering it: the state of every creative job, the budget consumed by campaign and by partner brand, the deliverables owed under the co-branding contracts and their deadlines. Two partner brands have €11,400 of unspent budget expiring at year end — the kind of sum that gets lost for want of being looked at in time.
The time this gives back: those chases took 5 hours a week from your production coordinator, split across forty interruptions. She has one left — the 14 real subjects. This is not time saved on work, it is time saved on interruption, and that is the most expensive kind in a studio.
What I suggest: a weekly status sent on Friday to each store, with what is waiting on them. Across the six stores where we tried it, chasing fell by 68 % — people chase when they do not know. artwork-tracking_92-requests-a-month.pdf78 answers in 10 minutes, 14 for a human
⛓ Sourced · studio inbox, production schedule, CRM, co-branding contracts
What I answer, and where it comes from: the current step, who holds it, the expected return date, and the direct link to the proof awaiting approval. I read the production schedule, not a promise: if a job has slipped, the store hears it from me before discovering it.
The 14 I pass on: four requests outside the campaign plan, three priority calls between stores, seven where the honest answer is “we need to talk”. Committing to a date is not information, it is a project manager's decision.
What needs attention this week: three proofs have been awaiting store approval for 9 days, and the print run is in six. I have chased twice, with the direct link. I flag it because it is the studio that will be held responsible for the delay.
On account tracking, what I keep current in your CRM with nobody entering it: the state of every creative job, the budget consumed by campaign and by partner brand, the deliverables owed under the co-branding contracts and their deadlines. Two partner brands have €11,400 of unspent budget expiring at year end — the kind of sum that gets lost for want of being looked at in time.
The time this gives back: those chases took 5 hours a week from your production coordinator, split across forty interruptions. She has one left — the 14 real subjects. This is not time saved on work, it is time saved on interruption, and that is the most expensive kind in a studio.
What I suggest: a weekly status sent on Friday to each store, with what is waiting on them. Across the six stores where we tried it, chasing fell by 68 % — people chase when they do not know. artwork-tracking_92-requests-a-month.pdf78 answers in 10 minutes, 14 for a human
⛓ Sourced · studio inbox, production schedule, CRM, co-branding contracts
We'd like to go after partner brand budgets, not just spend our own.
Five brands to approach, chosen from what has worked for you — and the visual recommendation for each one is built.
Where the five come from: I compared your 2 profitable co-branding operations with the 3 that were not. Three traits recur: a brand already listed in your 34 stores (2 of 2 against 0 of 3), an annual media budget above €60,000, and no in-house studio on their side — that last trait is the decisive one, because it is what gives yours its value.
What I prepare for each: a mood board — the plate of images that sets the visual direction of a proposal before any production —, three campaign routes adapted across your 11 formats, and a costing of what the brand would pay elsewhere for the same set. You walk into the meeting with visuals, not with an intention.
What that costs to produce, and this is where it becomes a commercial weapon: building those three routes took you two studio days per prospect, which made the whole approach impossible with a team of five. The five recommendations took half a day in total, art director review included. This is not a time saving, it is an action you could not carry out.
What I advise against, with figures: approaching the three powered-garden-equipment brands on your initial list. They all have an in-house studio, and across your three past operations with that profile, production was imposed by them 100 % of the time: you would have carried the distribution without carrying the creative. The budget would have come, the value would not.
And the argument to keep for the end of your meetings: all of this lives in France, on your own machine, with the deployment objective of processing and access operated within the European Union. A partner brand's unreleased visuals, its launches, its prices: none of it leaves your walls, and you can prove it in writing. For a brand entrusting you with a campaign before its launch, that is not a legal argument — it is the one that separates you from a cheaper supplier. co-branding_5-brands-and-their-routes.pdf3 routes per brand, half a day in total
⛓ Sourced · profitability of the 5 past co-branding operations, store listings, brands' published media budgets
Where the five come from: I compared your 2 profitable co-branding operations with the 3 that were not. Three traits recur: a brand already listed in your 34 stores (2 of 2 against 0 of 3), an annual media budget above €60,000, and no in-house studio on their side — that last trait is the decisive one, because it is what gives yours its value.
What I prepare for each: a mood board — the plate of images that sets the visual direction of a proposal before any production —, three campaign routes adapted across your 11 formats, and a costing of what the brand would pay elsewhere for the same set. You walk into the meeting with visuals, not with an intention.
What that costs to produce, and this is where it becomes a commercial weapon: building those three routes took you two studio days per prospect, which made the whole approach impossible with a team of five. The five recommendations took half a day in total, art director review included. This is not a time saving, it is an action you could not carry out.
What I advise against, with figures: approaching the three powered-garden-equipment brands on your initial list. They all have an in-house studio, and across your three past operations with that profile, production was imposed by them 100 % of the time: you would have carried the distribution without carrying the creative. The budget would have come, the value would not.
And the argument to keep for the end of your meetings: all of this lives in France, on your own machine, with the deployment objective of processing and access operated within the European Union. A partner brand's unreleased visuals, its launches, its prices: none of it leaves your walls, and you can prove it in writing. For a brand entrusting you with a campaign before its launch, that is not a legal argument — it is the one that separates you from a cheaper supplier. co-branding_5-brands-and-their-routes.pdf3 routes per brand, half a day in total
⛓ Sourced · profitability of the 5 past co-branding operations, store listings, brands' published media budgets
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
The uses of AI in communication and advertising
Each use corresponds to an agent we deploy. All of them 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 432 € excl. VAT / month
Multi-format adaptations
Adapt a master visual to every social, display and web format, with crops and safe margins adjusted.
Copywriting & content creation
Headlines, captions, ad copy and key messages consistent with the visual and the brand voice.
Knowledge base & briefs
Instantly find guidelines, briefs, templates and brand assets to frame every production.
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 validation, exceptions and escalation
Status, safe closure and audit trail
Human validation before publication, with versions retained
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.
Link content to performance, costs and campaign objectives
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.
Community management
Prepare the posts, calendars and adaptations per network, ready to schedule and approve.
Community management from 610 € excl. VAT / month Community management →Monitoring & campaign reporting
Track creative performance, gather the indicators and prepare the campaign reviews.
Campaign watch & reporting from 539 € excl. VAT / month Monitoring & reporting →CRM & account follow-up
Centralise the tracking of budgets, advertiser accounts and creative projects in progress.
Sales agent (meeting notes, follow-ups, CRM) from 592 € excl. VAT / month CRM & follow-up →AI search & campaigns agent
SEO optimisation, keyword research, content briefs, preparation of paid search and social campaigns: a sovereign AI agent, hosted in France.
Search & campaigns (SEO/SEA) from 534 € 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 creative team win back?
By automating resizing, adaptations and variants, a team can aim for a reduction by half in visual production time on repetitive tasks — reinvested in the idea and the art direction.
Adapting a campaign across 10 formats
Resizing a visual for the networks
Preparing A/B variants of an ad
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.
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
Three options, one agent
A visual creation agent (generation, adaptations, variants), installed and operated for you, to your guidelines. Choose according to how you work. Prices exclude VAT — annual subscription, the time it takes for the gains to settle in.
Agility
Setup + controlled subscription
3,665 € excl. VAT setup
then 432 € 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
637 € 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
6,760 € excl. VAT setup
then 629 € 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 to a communication team
Your creations and briefs never leaveLocal inference or an isolated resource hosted in France; no visual entrusted to a foreign third party.
Content in France, under French lawFor your creations and briefs: hosting under French law, native minimisation, architecture designed to reduce exposure to extraterritorial legislation, location alone not being enough to guarantee immunity.
The art direction keeps the decisionThe agent produces verifiable visuals and variants; no publication is automated.
Human oversight & traceabilityMonitoring and logging frame the multi-format adaptations; compliant with the requirements of the AI Act.
Frequently asked questions
Your questions, our answers
Can AI really adapt my visuals to my guidelines?
Yes, to a large extent. The agent takes a master visual, your templates and your guidelines (logo, palette, typography, protected zones) to generate the adaptations across every social, display and web format, with crops and safe margins adjusted. The creative moves from production to review and approval, much faster and of higher value.
Who owns the visuals the agent produces?
You do. Our agents run locally or are hosted in France: your creations, your briefs and your guidelines are never used to train a third-party model and are never reused elsewhere. What the agent produces stays your property, unlike many consumer tools whose terms of use are vague on this point.
Is my advertising clients' data protected?
Yes. Campaigns, briefs and your clients' data are sensitive: our agents run locally 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. Each organisation gets an isolated resource, with no pooling of data.
Is AI going to replace my creatives?
No. The agent takes on the repetitive production (adaptations, resizing, variants), not the idea or the art direction. It gives creatives back thinking time for design and strategy. The agent assists, the human decides: nothing is published without approval.
Does the agent integrate with my existing creative tools?
Yes. The agent connects to your environment (creative tools, digital asset management, approval space, editorial calendar) and complements it, without imposing a migration. We adapt the integration to your workflow and your templates.
Do you have to be a large agency to equip yourself?
No. The offer suits the communication team of a small business, an in-house studio and a multi-account agency alike. All inclusive, with no setup fee, and with a free audit to identify the use case with the best return.
How long does it take to deploy an agent?
A few weeks as a rule, after a free 15-minute audit that identifies the use case with the best return, then a phase of design, integration of your guidelines and testing before going live and training your teams.
Related use cases
Other agents for your communication
Let's talk
Let us estimate the potential for your communication team
15 minutes to identify the use case with the best return — hosted in France, supervised, with no commitment.