Water-sector data and AI evaluation workflow

Prove it on real data, then scale it

WETEC works with your utility to choose a use case, pilot AI on real data against measurable criteria, and only then bring it into staff’s daily work.

Common situations

Obstacles before AI goes into use

Technical documents and equipment for AI deployment at a water utility

Data is not ready

Documents exist in several versions and figures from software and devices do not match. Data sources need reviewing before a pilot.

No definition of success

You know which work AI should support, but have no criteria to judge results or decide whether to expand.

Staff unsure how to use it

Staff do not yet know how to ask, how to check an AI answer, or when to pass a case to the person in charge.

See it in action

Pilot results decide the next step

AI is run on real business scenarios and scored against agreed criteria. Select a result group to see an example.

Pilot on 40 business scenariosIllustrative example
Pass rate85%
Agreed bar 80%

Above the bar, ready to integrate

Example · need fixingWhat is the current tier-2 domestic water tariff?

AI quoted an outdated tariff. Update the source document, then re-test.

Scope of work

From choosing a use case to handing over the application

The use case, permitted data and evaluation criteria are fixed first. AI is integrated into live systems only after the pilot meets the bar.

Assessment and pilot

Choose priority tasks, review which data may be used and run AI on real business scenarios to measure results.

Deliverables

  • Use-case scope and evaluation criteria
  • Pilot results on your real data

Integration and training

Based on the pilot, WETEC connects the AI application to the relevant software and data, sets user permissions and trains staff.

Deliverables

  • AI application integrated into your systems
  • User guides and training

Delivery process

Pilot and evaluate before expanding

Evaluation criteria are fixed at the start. Pilot results, not gut feeling, decide the integration scope.

  1. Choose the use case

    Pick the work AI should support, the data it may use and how results will be judged.

    Pilot scope and criteria
  2. Pilot

    Run AI on real business scenarios and score the results with your users.

    Pilot results report
  3. Integrate

    Connect the application to relevant software and data and set access rights.

    Integrated AI application
  4. Train and hand over

    Show staff how to use AI, check its answers and pass out-of-scope cases to the person in charge.

    User guides and training

How we deliver

AI grounded in real data and real users

In the BIWASE AI project, WETEC connected internal technical documents to a Q&A assistant and ran it entirely on the company’s private infrastructure.

Data used within agreed limits

Your utility decides which data sources may be used and which user groups may access them.

Judged on real scenarios

Answers that fall short are logged and fixed before the application goes into use.

People make the decisions

AI supports lookup and analysis. Staff check the information and stay responsible for business decisions.

Training and support

Staff are trained to use the application, with documentation and a support contact.

Which of your tasks should AI support?

WETEC will review the use case, available data and evaluation criteria with your team and propose a suitable pilot scope.

Get AI advice