Skip to content
FenixIAm by AIworks 2028 FenixIAm

Tailor-made

Your process, turned into an agent that is accountable.

When the way you work does not fit a packaged product, we build the agent on the same engine other businesses already use. Models, voice, documents, measurement and branding are already solved; what gets built is whatever makes your process unique.

Paper scene created with AI.

The method

From process to agent, in four stages.

Every stage ends with something you can see and try out. If a stage does not convince you, it stops there, before anything is invested in the next ones.

  1. Process discovery

    We understand the work as it is done today, not as the manual says.

    What we do
    We walk through the process with the people who do it every day: the steps, the documents that come in and go out, the decisions, the exceptions and the points where time is lost or mistakes are made.
    What Fenix delivers
    A map of the process that separates what the AI will do from what a person will always decide, along with the criteria the result will be measured against.
    What you bring
    A few sessions with the person in charge and real examples of the documents they work with.
  2. Prototype on the engine

    You see it working with your sample data before committing further.

    What we do
    We configure the agent on the Fenix engine and build only what is specific to your process: screens, rules that block progress and document templates.
    What Fenix delivers
    A clickable prototype with your branding that already reads your sample documents and produces the first drafts.
    What you bring
    Sample data, anonymised or fictitious, and short reviews: what is superfluous, what is missing and what is worded wrongly.
  3. Measured pilot

    A team uses it in their real work, measured from day one.

    What we do
    We put the agent in the hands of a small team, with the agreed criteria. We fix whatever real use uncovers and measure quality, time and cost.
    What Fenix delivers
    The agent in use, results reports by criterion, the cost measured message by message and an honest list of what is still missing for production.
    What you bring
    A small group of users, time from the person in charge to validate results, and the decision to continue, adjust or stop.
  4. Production and scale

    What worked in the pilot is opened up to the rest of the organisation.

    What we do
    We extend the agent to more teams, sites or organisations, each with its own branding and isolated data, without rebuilding what already exists.
    What Fenix delivers
    The agent in production, with per-person access, a spending cap per agent and the engine kept up to date with the best available models.
    What you bring
    Someone to manage usage in your organisation and the roll-out schedule by team.

Built on the same engine

Two processes that have nothing in common. One engine underneath.

A municipal procurement file and the management of a farm share neither vocabulary nor users. Both were built on the same Fenix engine; what changed was the process.

Case 1 Built on the same engine · Municipal procurement Pilot

A procurement file that will not let you skip a requirement.

The municipal officer works through the file across 16 screens, from the start to the final package. The AI extracts the data from the sources, calculates and drafts; if a mandatory document is missing, the file does not move forward. Each town council sees its own name and coat of arms on the screens and in the documents.

  • 16 screens
  • Blocks that stop progress
  • Word and Excel documents
  • Exportable package
  • Each town council’s branding
Paper scene created with AI.

Services contract file · screen 9 of 16

The report on the lack of in-house resources is missing. The file cannot move forward without it.

Justification report ready in Word · cost calculation in Excel

Case 2 Built on the same engine · Farm Demo

Running a farm, with every month’s close sealed.

Plots, seasons, work reports and costs in one place. The monthly close flags what is missing with a traffic light, gets reviewed and is sealed. Then the 17-section report comes out ready to share, and in Excel.

  • Monthly close with traffic light
  • Sealed on closing
  • 17-section report
  • Excel export
Paper scene created with AI.

September close

In review · 5 days left August closed and sealed 2 work reports without an assigned cost
17-section report · Excel

What the engine provides and what gets built

We do not start from scratch. That is why you see results early.

Already solved in the engine

  • Models from seven providers, interchangeable without touching the agent
  • Real-time voice when the process needs it
  • Document reading and search across your own files
  • Word and Excel documents generated from templates
  • Data isolated per company, your own branding and per-person access
  • Cost measured on every message and a monthly cap per agent

What gets built for you

  • The screens of your process, in your own vocabulary
  • The rules that block progress when something is missing
  • The calculations specific to your business
  • The templates for your documents and reports
  • The criteria the result is measured against

Getting started

What you need and what you do not.

You need

  1. One specific process

    Just one, clearly scoped and with a result that can be checked. Better the one that eats the most time than the most ambitious one.

  2. One person in charge

    Someone who knows the process from the inside, can validate results and has the authority to decide whether to continue.

  3. Sample data

    Real or anonymised documents and cases to build the prototype. You do not need to open up your systems to get started.

You do not need

  • A long project before seeing results: the prototype comes before the big investment.
  • To change your systems: the agent starts by working with documents and is integrated once the pilot justifies it.
  • An in-house technical team: we maintain the engine.
  • A closed requirements document: the process is discovered with the people who do it.

Tell us which process takes up most of your time.

In a first conversation we check whether it fits the engine, what would need to be built and what the pilot would look like.