Skip to content

Artificial intelligence

Put AI to work in your business.AI agents, assistants on your documents and models connected to your software

GNR brings artificial intelligence to the tasks where your teams lose time: reading and sorting emails, finding information in your documents, drafting a quote, answering common questions. The AI is connected to your software and data with controlled access, and every action that matters is still approved by a person.

(01)What we deliver

What AI can do for you

  • AI agents that act

    An agent reads a request that arrived by email, finds the customer in your CRM, drafts the reply or the quote and submits it to you before sending.

  • An assistant on your documents

    An internal assistant answers from your procedures, contracts and product sheets, citing its sources so anyone can check.

  • ChatGPT or Claude connected to your tools

    With the MCP protocol and your APIs, your teams query their business data in plain language, with the same permissions as in your software.

  • Document reading

    Invoices, purchase orders, tender files, receipts: AI extracts the useful information and files it in the right place.

  • AI hosted on your side

    When your data must not leave, we deploy a model on your own servers or with the hosting provider of your choice.

  • A guided rollout

    Your teams learn the tool on their own tasks, with clear rules on what the AI does alone and what must be approved.

(02)When to call us

When AI becomes useful

  • Your teams copy information from an email or a PDF into a piece of software.
  • The same questions keep coming back, and the answer sits in a document nobody can find.
  • You already use ChatGPT, but with no link to your data and no shared rules.
  • Answering a tender or a quote request takes hours of reading.
  • You want to test AI on a real case before investing more widely.

(03)Example

Start small, on a real case.

We start from one precise, measurable task, such as sorting incoming requests or searching your documents. It is delivered in a few weeks, used by a pilot team, then extended if it delivers. In our construction demo, the AI reads a 214-page tender file and drafts the technical proposal section by section, which the company reviews and approves.

(04)How we work

One lead, a demo every week.

Fixed scopes, weekly demos and one accountable lead. You always know what ships next and why.

  1. 01

    Discover

    1 to 2 weeks · Workshops, user interviews and a technical audit. You leave with a clear scope, a plan and a fixed estimate.

  2. 02

    Design

    2 to 4 weeks · Flows, prototypes and a design system, tested with real users before a line of production code is written.

  3. 03

    Build

    8 to 16 weeks · Weekly releases to a live staging environment, so you can see and steer progress every Friday.

  4. 04

    Evolve

    Ongoing · Monitoring, support and a roadmap for what comes next. Most of our clients stay with us for years.

(05)Pricing

What it costs

An assistant on your documentsUsual lead time3 to 5 weeks
€3,000 to €5,000 excl. tax
An agent that acts in your softwareUsual lead time4 to 8 weeks
€4,500 to €7,000 excl. tax

Indicative estimate, excluding tax. You get a firm price after a first conversation.

See our pricing

(06)Tools

The tools we connect

  • ChatGPT
  • Claude
  • Mistral AI
  • Gemini
  • MCP
  • n8n
  • Make
  • Google Drive
  • Notion

And more than 3,000 other applications, connected through their APIs or ready-made connectors.

(09)FAQ

Your questions

Do we need to change software to use AI?

No. We connect AI to the tools you already use (email, CRM, management software, shared folders) through their APIs or connectors such as n8n or Make. Your teams keep their habits.

Is our data sent to OpenAI or anyone else?

Only if you decide so. Depending on how sensitive the data is, we use business plans that do not train models on your data, a European model such as Mistral AI, or a model hosted on your own servers.

What is an AI agent, in practice?

A program that does more than answer: it chains actions in your software, such as reading a request, looking up information, drafting a document and passing it on. Together we set what it may do alone and what it must have approved.

How long does a first AI project take?

A first, focused use case is up and running in a few weeks: one to two weeks to scope it, then a delivery every week on a test environment. You see the AI working on your real documents before rolling it out.

What if the AI gets it wrong?

It can. That is why our assistants cite their sources, and actions that commit the company (sending, paying, signing) go through human approval. Answer quality is also measured on real examples before going live.

Got a project in mind?

One of the founders replies within one working day with a few questions about your context, and you get a written proposal within a week.