IT and AI strategy
An assessment, a costed roadmap, and the order in which to tackle things.
- Assessment of your tools, workflows and data
- A twelve- to eighteen-month IT master plan
- The AI use cases worth doing — and those that are useless
// services
Advice, support, delivery, training
Four ways to work together, on their own or in a row: knowing where to start, steering a project, having it built, then handing it over to your teams. For SMEs, mid-size and large enterprises and local authorities, from the Gers and the Toulouse area.
Small business, shop, craftsperson or freelancer: the offer for you has its own page.Small businesses
// 01 · advice
Advice
Knowing where you stand, what is worth doing, and in which order.
An outside view of your tools, workflows and data before committing to anything: what works, what gets stuck, the AI use cases worth doing and those that are useless.
Formats
Assessment
Half a day on site: tools, workflows, data, AI use cases. You leave with a written report.
IT master plan
A twelve- to eighteen-month roadmap: what, in which order, and what each step costs.
Examples
Taking over legacy systems
Understanding software nobody can explain anymore, documenting it, deciding what to keep.
- We start from the code and the real data, not from outdated documentation
- A single business vocabulary: yours
- Three costed scenarios: keep, evolve, rebuild
// 02 · support
Project ownership support
You steer your project; it gets scoped, compared and followed up until go-live.
For an SME, a mid-size enterprise or a local authority that stays in charge: the need written down, solutions compared with no vendor commission, the chosen provider followed up until go-live. The decision stays yours; method, technical arbitration and follow-up are provided.
Formats
Scoping workshop
Your project becomes a written file you can take to a bank or a board.
Project ownership and management
You stay in charge: method, technical arbitration and follow-up, alongside you, through acceptance.
Examples
Choosing your tools
ERP, CRM, DMS, point of sale, servers: choosing, connecting to what exists, migrating the data.
- A reasoned choice, with no vendor commission
- Data you can always get back out
- No technical debt: a tool you can update, and leave
// 03 · delivery
Delivery
A tool built, put into service and delivered finished, on your accounts.
When something must be built: connecting tools, automating workflows, putting AI into service or writing business software. Delivered in stages, with its backups, monitoring and documentation.
Formats
Fixed price per deliverable
Written scope, fixed price, dated milestones: the default format.
Examples
Automating workflows
Connecting tools that ignore each other, removing re-keying and manual reminders.
- We follow a real case end to end to see where information gets lost
- Data entered once flows from one tool to the next
- Reminders, exports and alerts go out on their own — and every automation is monitored
AI in production
Controlled use cases: capped costs, controlled data, hosting you choose.
- Incoming documents, tenders, reports, an assistant over your own documents
- On your premises, on your server or in a controlled cloud, decided use case by use case
- A spending cap, and data not reused for training
Business software and intranets
When off-the-shelf isn't enough: a custom, documented web and mobile tool.
- Built from your spreadsheets and your real practices
- Delivered in stages, with a quick first go-live
- Role-based access, audit trail, export always possible
// 04 · training
Training
Hands-on, on your tools and documents, so that your teams become self-sufficient.
Not a lecture on AI: workshops on your teams' own tools and documents, during the engagement and at its end, so that they use it, evolve it and take it over by themselves. No project is sold through job cuts.
Formats
Hands-on workshops
On your tools and documents, with your teams' real cases.
Handover at the end of the engagement
What was built is explained to those who take it over: documentation, accounts, handover.
01
Listen first
“Is my job going to disappear?” The question comes up at the first workshop, out loud. A tool installed against the team ends up unused.
02
Then show
Ten minutes on their own documents. Seeing the machine make mistakes, and keeping the last word, defuses most of it.
03
Then build
Those who will use it scope, test and decide. That's where adoption happens, not at delivery.
Not Qualiopi-certified, so the training cannot be funded by an OPCO (French training fund).
// how it runs
How an engagement runs
01
Scope
The need gets written down: what's in, and above all what's out. You approve before anything is quoted.
02
Quote
A quote, not a range: every deliverable priced, dated milestones. Nothing starts without your written agreement.
03
Build
You see the tool running regularly, not just at the end. What's wrong gets fixed while it's still cheap.
04
Deliver and hand over
Delivered finished, with backups, monitoring and alerts in place, then explained to those who take it over.
// terms
Written down before we start
A fixed price per deliverable
Every engagement is quoted: no meter running, no extension halfway through.
Two options, both quoted
The classic way, and the AI-assisted way, cheaper because faster. You see the difference, and it goes to you.
Turnkey, no subscription
No recurring fees. Hosting and supplier accounts are opened in your name: you pay them directly. Monitoring is delivered with its alerts; having them watched for you is an option.
Reversible
Code, documentation and data are handed over. No proprietary tool: another provider can take over.
// questions
Frequently asked questions
- How much does an engagement cost?
- It's quoted, with a fixed price per deliverable; the first conversation is free.
- Why not just do it myself with AI?
- For anything non-critical, do: a landing page, a first spreadsheet, a draft. With AI, a demo takes a weekend. What separates it from a tool in production doesn't show: who can see what, verified backups, an alert before the outage, security updates, documentation that lets someone else take over. As soon as personal data, payments, several users or a legal obligation are involved, that is where most of the work lies.
- If AI is faster, why this price?
- AI has lowered the price of code, not the price of mistakes. What you pay for is a finished tool — access rights, backups, monitoring and alerts, documentation, reversibility. What it saves goes to you: it is the difference between the two options priced on the quote.
- Do we need to change software to use AI?
- Rarely. Most useful use cases — reading incoming documents, drafting a report, querying your own documents — plug into the tools you already have.
- Where is our data hosted?
- Wherever you decide: on your machines, on a server rented in Europe in your name, or with a cloud provider under a spending cap. The choice is made use case by use case, and sensitive data stays put.
- Will we depend on you?
- No. Everything is in your name, and your teams are trained to take over — or to hand the next steps to another provider.
- Who are these services for?
- SMEs, mid-size and large enterprises and local authorities, from the Gers and the Toulouse area. Small businesses and freelancers have their own page, with a dedicated offer.
// limits
What I don't do
- Visual identity and printing
- Reselling hardware and licences
- 24/7 on-call and managed services
- Paid advertising
- High-volume e-commerce
For these, you will be pointed to someone else.
// contact
A free one-hour conversation
You describe your situation and priorities, and you leave knowing where to start — advice, support, delivery or training.