Case studies

For them, it's already running.

Five projects in production, with the figure measured at the client's site and what it took to get there. Not pilots gathering dust in a corner.

50+ SME and mid-sized companies supported

  • Accessite logo, a StratIA client
  • Hello Prépa logo, a StratIA client
  • Bastide Manutention logo, a StratIA client
  • Paymetrust logo, a StratIA client
  • Comité National Olympique et Sportif Français logo, a StratIA client
  • Union des Marchands de Biens logo, a StratIA client
  • Ordre des Géomètres-Experts logo, a StratIA client
  • BTP Ingénierie logo, a StratIA client
  • EM Lyon Business School logo, a StratIA client
  • Logo of a StratIA client
  • Fandi logo, a StratIA client
  • Caratelli logo, a StratIA client
  • Maison Colombo logo, a StratIA client
  • Colibrimm logo, a StratIA client
  • Ligue Occitanie de Tennis logo, a StratIA client
  • Disney logo, a StratIA client
  • Quoting

    Toulouse-based manufacturer, 300 employees · Custom quotes · The AI quoting agent

    From several days to one or two hours to produce a quote.

    The problem.
    Every request, even highly specific ones, started from scratch. Quoting took several days, the volume of responses was capped by the availability of a few people, and margins on custom parts varied depending on who did the estimate.
    What we did.
    We built a quoting AI based on about ten years of quotes and previously produced parts. Each new request is compared with the history to produce a consistent quote, which the estimator approves or corrects.
    The result.
    Quoting time cut from several days to one or two hours. Two to three times more quotes handled over the same period. Consistent margins on custom parts.
    A large, empty industrial hall
  • Quality

    Metal parts manufacturer · In-line inspection · Custom

    100% of parts inspected, without slowing down the line.

    The problem.
    Inspection was done by sampling. A defect caught late means an entire batch to rework or scrap, and a non-conformity that reaches the customer.
    What we did.
    We added a computer vision layer to lines already equipped with cameras. Every part is analyzed in real time, defects and non-conformities are detected automatically, and the full history is logged.
    The result.
    35% less scrap. Non-conformities cut sixfold. 100% of parts inspected.
    A close-up of a metal mechanical part
  • R&D

    Manufacturer running many tests · Test knowledge capture · The Referent

    Finding a comparable test: from several hours to a few minutes.

    The problem.
    Ten years of tests and reports, and every new question meant reopening files or asking a colleague. Tests were rerun when the answer already existed somewhere.
    What we did.
    We built a knowledge base covering all their tests and reports. Teams ask their question in plain language and find the comparable test with its results, with sources.
    The result.
    Search time cut tenfold. Over ten years of tests indexed. Dozens of redundant tests avoided every year.
    A close-up of an automated test bench
  • Compliance

    Aerospace manufacturer · Certification files · The Writer

    Writing time cut by three to five times, submission lead time shortened by several weeks.

    The problem.
    Building a certification file means digging by hand, through years of internal documents, for the evidence matching each requirement of the standard: material certificates, test reports, process qualifications. The engineering office spent a considerable amount of time on it, one file after another.
    What we did.
    We built an agent that identifies the requirements to cover, finds the evidence among existing documents and assembles a structured, traceable file, with every statement linked to its source. Requirements with no evidence found are flagged rather than filled in. The quality engineer approves before submission.
    The result.
    The engineering office moved from manual compilation to verification. Writing time cut by three to five times, files handled in parallel rather than one by one, submission lead time shortened by several weeks.
    A fighter jet in an assembly hangar
  • Funding

    French National Olympic and Sports Committee · Grant engineering · Custom

    Grant application writing time cut fivefold.

    The problem.
    Building a grant application means tracking calls for projects, understanding what each one expects, and producing an institutional-quality file. Most organizations go through an outside firm or give up.
    What we did.
    We designed Novia, a platform that cross-references public data from the French Ministry of Sports with the applicant's project, identifies relevant calls for projects and structures the application, from the opportunity memo to submission.
    The result.
    Writing time cut fivefold, and the entire process on a single platform. The same mechanism applies to the BPI, regional or European funding applications a manufacturer puts together to finance its investments or its transformation.
    A person looking at charts on a tablet
A calculator lying on financial charts, in front of screens showing graphs

What these figures mean

They are all measured at the client's site, within the project scope. None is a market average or a projection. They don't tell you what you'll get, because your volume, your data and your processes aren't theirs. They tell you what happened for someone who had the same problem.

When we estimate a gain for you, during an audit or scoping, the calculation method comes with it. You can challenge it.

What every project has in common

  • A narrow scope to start.

    One topic, one department, one document type. We expand once it's running.

  • Business experts in the loop from day one.

    They are the ones who define what a good answer is.

  • A supervised phase before production.

    The agent proposes, a human approves, and we measure the gap until it reaches the agreed threshold.

  • What already existed, used as raw material.

    Ten years of quotes, ten years of tests, a messy DMS. We start from what you have, not from a cleanup project first.

  • The decision stays with you.

    None of these agents signs, approves or sends anything on its own.

AI is redefining industry standards. It might as well be you who sets them.

Our job is to keep you one step ahead.

Thirty minutes to understand your situation. At the end, we tell you whether we can help, and if so, how.

Let's talk about your case

We support manufacturers in France, Europe, the Middle East and Africa.

  • An aircraft under maintenance in a hangar
    Aerospace
  • Orange robots on a production line
    Manufacturing
  • An automated production line
    Plastics processing
  • A forklift in a warehouse
    Material handling

Frequently asked questions

Can we talk to one of your clients before getting started?

Yes. Ask us for the reference closest to your situation, and we'll put you in touch with the client, with their consent. It's often the best way to find out how a project really goes.

Why aren't all your clients' names displayed?

Many manufacturers don't want their competitors to know what they've put in place, and we respect that. We describe the industry, the size and the problem. We share the details in a meeting, with the client's consent.

Will I get the same results?

We won't promise you that. These figures are measured at our clients' sites, within their scope. For you, we estimate the gain during the audit or scoping, based on your volumes and your data, and the calculation method comes with it.

How do you measure the results of an AI project?

Against the client's metrics, not ours: time spent per quote, scrap rate, time to find a piece of information. We measure before starting, then during the supervised phase, until we reach the threshold agreed together.

My industry isn't in your case studies. Can you help?

Probably. What manufacturers have in common is the problem: quotes that take too long, incomplete quality control, know-how locked in a few people's heads. Thirty minutes is enough to see whether we've already solved a problem like yours.