Back to home AI consulting firm — Poitiers · Montréal

AI consulting, by engineers who ship it to production.

An AI project is won or lost before the first line of code: which use case to address, which architecture to choose, under which legal framework. We inform these decisions with the experience of a team that designs and operates its own systems in production. You leave with a clear roadmap, independent of any vendor.

We work out of Poitiers and Montréal, remotely or on your premises.

A first video call, no commitment. Reply within 24 business hours.

ADEME Our advice is grounded in practice: ADEME asked us to take its funding-application screening agent into production.
  • Since 2019

    we design, ship and operate our own Data & AI products in production

  • 500+

    organisations use our GDPR compliance software

  • ADEME

    asked us to take its screening agent into production, after its challenge

  • 0 licences

    zero reselling, zero vendor commissions — advice without a conflict of interest

the essentials

What to know in 30 seconds

AI consulting means scoping an artificial intelligence project, choosing its architecture and steering its delivery. Three decisions shape that work: which use cases to prioritise, which models and hosting to pick, how to comply with GDPR or Quebec's Law 25. Good scoping gets the prototype into production, instead of leaving it in the lab.

DPLIANCE is an independent AI consulting firm, founded in Poitiers, France, in 2019 and present in Montréal. It offers three formats: AI project audit and scoping, architecture and technical choices, and a fractional CTO for AI and data. The first two are fixed-fee engagements; the third is a subscription of a few days per month to steer your roadmap alongside your teams.

The difference: our consultants have designed and operated AI solutions in production since 2019, including a case-analysis agent that won the Dis-ADEME challenge. Our advice stays independent from any sale: zero licences to place, zero developer days to fill.

The first call is 30 minutes, with an engineer, no commitment. Reply within 24 business hours.

Request a quote
the problem

Advice on AI without an agenda is rare

You are already convinced you need to "do AI" — everyone keeps telling you so. Your problem is elsewhere: every vendor recommends their software, every consultancy recommends an engagement, every cloud recommends its infrastructure. When the advice and the sale come from the same pocket, the recommendation is a pitch.

The result: technology choices you endure rather than make, impressive demos that stay stuck at prototype stage. Budgets get spent before anyone examines compliance. An IT director is then left alone to decide on topics that few hires cover.

The classic alternative is to recruit a full-time AI-specialised technical director. The profile is scarce and expensive, and most project volumes are rarely enough to justify one. That is exactly the space our consulting covers.

Sound familiar? Describe your situation to an engineer. Talk to an engineer

the three offers

Three ways to secure your AI decisions

Two fixed-fee engagements to decide, one subscription to steer. Every engagement is quoted for your context: scope, data readiness, integration depth. The price stays firm before we start.

Fixed fee

AI project audit & scoping

You have use cases in mind — or an "AI" mandate and a blank page.

  • Assessment of your data and information system
  • Use cases prioritised by business value and feasibility
  • Roadmap with realistic timelines and workloads
  • GDPR / Law 25 compliance built into the scoping
  • Staged budget estimate, from prototype to production

Duration: 2 to 3 weeks

Fixed fee

Architecture & technical choices

Your project is decided — what remains is how to build it, with what, and where.

  • Documented target architecture, sized for production
  • Model choices: French, European, or deployed on your servers
  • Hosting aligned with your jurisdiction (GDPR, Law 25)
  • Privacy by design: records, DPIA or PIA, decision traceability
  • A technical blueprint your teams or vendor can execute

Duration: 3 to 4 weeks

Subscription

Fractional CTO for AI & data

You want an AI technical director at your side, without hiring one full-time.

  • Steering of your AI and data roadmap, a few days per month
  • Technical arbitration and review of vendor proposals
  • Support for hiring and coaching your data and dev teams
  • A seat at your committees: leadership, IT, vendors
  • Contextual watch: what matters to you, not the hype

Duration: adjustable monthly commitment

Torn between two formats? That is exactly what the first call is for. Talk to an engineer

the signature

Like a fractional CTO — specialised in AI, data and privacy

Most SMEs, mid-caps and public bodies rarely have the need or the budget for a full-time AI technical director. Yet they face the same decisions: which project to launch, which architecture to choose, which vendor to trust, how to stay compliant. Today those decisions get made with nobody to instruct them.

That is the role we play: a part-time technical director who knows your systems and sits on your committees. We combine a specialisation few CTOs share: artificial intelligence, data and privacy, practised together since 2019.

We operate our own products in production. Our opinions come from that ground, not from an analyst report. When we advise against an architecture, it is often because we have already paid to learn why.

Discuss a fractional CTO engagement

independence

Advice without a conflict of interest

A delivery consultancy lives off its developers' billable days: its advice naturally leads to a delivery engagement. A vendor lives off licences: its audit rarely concludes its product is useless to you. Our consulting aims at something other than delivery: our deliverables stay actionable by any competent team, yours as much as a vendor's. Delivery through us remains one option among others, not an obligation.

That independence lets us make the recommendations that sting. An off-the-shelf tool can be enough, a project can need more maturity, AI can be the wrong answer to your problem. Advice that always says yes is worthless.

Zero licences to place

Tool choices follow your criteria alone: zero software reselling, zero vendor or cloud commissions.

Zero days to fill

Fixed-fee engagements have a deliverable and an end. The CTO subscription scales up, down or off, always under your control.

Zero engineered dependency

Deliverables are documented and actionable without us. If we vanished tomorrow, your roadmap would still work.

two modes

We build for you, or we steer with you

DPLIANCE works in two complementary ways. The right mode comes down to one question: who builds?

We build for you

The agency builds your solution and keeps it running afterwards. That is the right mode when you want a result in production without tying up an internal technical team.

Discover the agency offer

We steer with you

Consulting instructs your decisions; the build stays with your teams or your vendors. You keep your hands on the work, we secure the choices that commit it.

See the three consulting offers

In both cases, the same engineers apply the same Prod First method. The same sovereignty doctrine applies too: your data stays in your jurisdiction, under your legal framework.

how it works

How a consulting engagement starts

  1. First call

    30–45 min

    An engineer reviews your situation over video, no commitment. We talk projects, teams and constraints. We tell you frankly whether a consulting engagement makes sense.

  2. Proposal

    within 1 week

    You receive a written proposal: scope, deliverables, schedule and a firm quote. The rate card stays transparent — you know exactly what you are buying before you sign.

  3. Engagement

    2 weeks to several months

    Scoping or architecture as a short fixed-fee engagement, fractional CTO over the months: either way, you work directly with the consultant, never through a junior.

  4. Handover & next steps

    at every stage

    Every engagement ends with actionable deliverables and a debrief to your leadership. You continue alone, with your vendor, or with us — your call.

The first step takes only 30 minutes of your time.

frequently asked questions

Frequently asked questions

What does an AI consultant do?

An AI consultant helps an organisation decide on and succeed with its artificial intelligence projects. Their role: identifying profitable use cases, scoping the project, choosing the architecture and models, selecting vendors and steering delivery. At DPLIANCE this advice rests on something specific: our consultants have designed and operated AI solutions in production themselves since 2019. GDPR and Law 25 compliance is a daily discipline here.

Why external consulting rather than hiring?

An AI-specialised technical director is a scarce, expensive profile that is hard to keep busy full-time outside large organisations. External consulting gives you that level of expertise on demand. Budget a few weeks for a scoping, a few days per month for ongoing steering. And the day your project volume justifies a hire, we help you get it right — it is part of the fractional CTO offer.

How do you choose an independent AI consulting firm?

Three criteria are enough to sort the field. Independence first: does the firm also sell the build, licences or hosting? If so, its recommendation is a sales pitch. Practice next: have its consultants actually shipped an AI solution to production, or only produced reports? A scoping written by someone who runs models in production themselves is worth more than a theoretical one, which usually underestimates integration and compliance. The legal framework last: hosting, transfers outside the EU, lawful basis, AI Act documentation. A firm discovering those topics mid-engagement will make you pay for them twice. DPLIANCE ticks all three criteria: zero reselling, zero commissions, AI solutions designed and operated in production since 2019. Both the GDPR and Quebec's Law 25 are practised daily.

How is this different from a consultancy or strategy firm?

A delivery consultancy sells developer days: its advice prepares its own engagement. A strategy firm delivers recommendations that are rarely production-tested. Our consulting combines the two missing qualities. Independence first: we stay away from delivery. Practice next: our practitioners operate their own AI solutions in production. Our deliverables are written to be executed, not archived.

Are your recommendations genuinely independent?

Yes, structurally so: zero licences resold, zero vendor or hosting commissions, and our consulting engagements exclude any delivery. If the best answer to your need is an off-the-shelf tool at €30 a month, that is what the deliverable will say. The freedom to conclude "walk away" is precisely what you are paying for.

How much does an AI consulting engagement cost?

Every engagement is quoted individually, based on scope, data readiness and the depth of analysis required. Fixed-fee engagements, scoping and architecture, come with a firm price announced before we start. The fractional CTO subscription is sized in days per month and adjusts over time. Describe your need to an engineer: you will receive a priced proposal within days, with a reply within 24 business hours.

Can we start small?

It is even recommended. A two-to-three-week scoping is enough to prioritise your use cases and avoid bad investments. You then decide, with full knowledge, whether to go further. The fractional CTO subscription starts at a few days per month and adjusts. And the first 30-minute call remains free of charge.

Do you work with our IT department and dev teams?

That is the whole point of the offer: we support your teams, we don't replace them. In practice, we instruct decisions for your IT director and review architecture with your developers. We also evaluate vendor proposals and upskill your data teams. Consulting succeeds when your teams are more autonomous after the engagement than before.

What if we then want the solution built?

You are free to choose. The deliverables, roadmap and architecture blueprint, are written to stay actionable by any competent team: yours or a third-party vendor's. If you choose to entrust the build to our agency, we apply the Prod First method described on the agency page. That build remains one option among others, not a condition of the consulting.

Do you work in Quebec and Canada?

Yes. Florian Gadal, our CTO, is based in Montréal, and we practise Law 25 daily — records, PIAs, data governance. Quebec and Canadian organisations get the same consulting as in Europe. Hosting and architecture recommendations align with their own legal framework, rather than a simple transposition of the European one.

Your question isn't here? Ask an engineer directly

Let's talk about your next AI decision.

In practice: you write to us, and we reply within 24 business hours to set a slot. You then spend 30 minutes on a video call with an engineer. That reply comes straight from the engineer, outside any sales process and without an automated email sequence.

  • 30 minutes with a consultant who has already shipped AI solutions to production — they understand your context technically.
  • A reasoned opinion: the right approach for your case, realistic timelines and the data prerequisites to gather.
  • A firm quote within days if an engagement makes sense — zero commitment otherwise.

Reply within 24 business hours, from a human.