You pay for AI where it pays off. Nowhere else.

For whoever answers for a process in a French SME and can decide to change it: the owner or managing director, the operations or plant director, a branch director, a technical director. Most often in a firm of 50 to 249 people that sells expert judgment — engineering and design offices, architecture practices, consultancies, specialist trades. You do not need to be the AI expert: that is the starting assumption, not a gap.

The rung you want depends on what is blocking you. A decision you have to make and defend — that is Counsel. A process that still waits on the same two or three people — that is the redesign of that process. If ten files out of ten are handled by applying the procedure to the letter, or if the method belongs to a software vendor or an accreditation body whose sign-off any change needs, neither is for you, and we will say so.

You describe a process, or your business if you do not know which one to pick yet. We ask five questions: how often it comes back and how long it takes, who exercises judgment in it, what a mistake costs when one gets through, where your past jobs live — bids, quotes, post-mortems — and in what shape, and who can say yes.

The call ends with a straight answer: the process we would start with, what it costs you today per year, estimated from your own figures, and what we would do in your place. No deck, no diagnosis to buy, no surprise quote: a quote only comes afterwards, and only if both sides want to carry on. The person who would do the work is the person on the call.

The twenty minutes cost nothing. The three reference files cost €4,500 excl. VAT; if you go on, that amount is deducted from the price of the software. The process rebuilt is quoted at the end of those three files. We cannot price it earlier, because we do not know the process until we have looked at it with you. The follow-up is included.

One signature covers everything, and you can stop at the end of the three files, with no penalty. You keep your annotated files, your three numbers and the plan.

Before we start, we also fix what this process costs you today (real lead times, rework, rejected files) and what will count as a success. Those are the numbers you will follow the software against afterwards. You are buying a process that runs: the software is on your premises or on sovereign infrastructure, it can evolve without us, and there is no subscription to our presence.

Two answers, because these are two different things.

The system we build for you runs on your own infrastructure or on sovereign hosting. Your files never have to leave your walls.

We choose each AI model according to what your files contain. Whatever is confidential goes through models installed on your premises or on sovereign infrastructure: you keep control of them, and their cost does not move at a vendor's whim. A commercial model can serve on a step where it does better, provided it sees nothing confidential. It then works on documents that are not, or on files from which we have first removed the names and whatever you want protected. You decide what leaves, and the record shows which model served at each step.

Counsel, the online product, is more exposed and we would rather say so: the text of the decisions you submit is analysed in the United States by Google LLC (Gemini API) and Tavily Inc., processors under Article 28 of GDPR, who do not use it to train their models. Everything else — account, reading, billing — is hosted in France with OVH and does not leave the European Economic Area. The full list of processors is in our privacy policy.

The Agenda (free) is the monthly edition: what AI puts on your table, by business function, chosen for what you will have to decide — thousands of articles read, a dozen kept, and each issue names what it refused. One-to-many.

Counsel (Solo) is decision-grounded: you describe an open decision, and each generation surfaces the field's published evidence ranked by how directly it bears on your situation. Every claim links to a verbatim passage from the source article, and each item is honestly tagged as direct, adjacent, or tangential to what you are deciding. It reads like a consultant's synthesis, not a feed.

Because it is not content delivered to you: it is a position argued against your situation, quoting a real source for every claim, and re-argued when the evidence moves. If it does not pay for itself on the first decision it changes, it is not for you, and you have fourteen days to find out.

As many as you actively need. There's no slot meter, no per-decision charge. We size capacity for the typical Solo user (1–5 active decisions); if you're running more in parallel, the Team plan gives you shared context across seats and is built for that load.

You archive it in your dossier and free up the mental space. The Counsel history stays — you can re-open any past decision, export it as Markdown for your team, or re-argue it if the context changes. No billing change either way; Solo is one flat price.

No. Refresh is unlimited — you're not metered on LLM usage. We chose this so you never face surprise bills; one price, all decisions, all refreshes.

A one-shot Q&A surface that complements your decision-grounded Counsel. You type a question — "what does the evidence say about putting client files through an AI tool?" — and Counsel answers from the public corpus over a 180-day window, with verbatim citations and an honest confidence band (high / medium / low). Different from your daily Counsel: no stored context, no decision required, single answer. Included on Solo and Team.

A chatbot writes first, then attaches links that seem to back what it said. Counsel reads first, then writes only what it can quote literally.

On approach. Perplexity and ChatGPT pull snippets from the open web at query time, generate a response from them, then attach citations post hoc without verifying — often beside what the source actually says. Test in 30 seconds: click the links in your last Perplexity reply. Half don't say what they're credited with. Counsel does the opposite: selection from a curated, quality-rated AI corpus, structured extraction at ingestion, scoring by an LLM judge against your specific decision, synthesis built only from the extracted passages, passage-by-passage grounding verification, public marking of remaining uncertainty. It doesn't invent a tool, a metric, or a citation — and when it can't anchor a phrase, it says so in the document itself.

On usage. Your decisions, questions, and monitored subjects are stored and automatically re-checked against every new article. The system learns from your explicit signals (useful / off-target / pinned) and your clicks, tracks what changed between two Counsels (new, carried, promoted, dropped), and names the decisions with no recent signal — "agent framework: no signal in 41 days" — exactly where the six-month-too-late reversals come from.

It's not a better prompt. It's a different architecture.

Team plans give every seat unlimited Counsel decisions plus shared context across the workspace. €499/mo flat, or €399/mo billed annually (€4,788/yr, 20% off) — up to 5 seats either way. A workspace admin manages members and billing. Shared decision context and team-wide feedback loops are available once two or more seats are active.

Yes. Cancel from your account settings. Annual plans keep access until the billing period ends. Monthly plans stop at the next cycle. You never lose your Counsel history or your bookmarks.

All major credit cards (Visa, Mastercard, Amex) via Stripe. Annual plans are charged once upfront. Monthly plans renew on the same day each month. EU-issued invoices include reverse-charge VAT where applicable.

We offer a 40% discount on Solo for verified students, researchers at academic institutions, and registered nonprofits. Email us at contact@aissential.tech with your institutional email.