An abstract iFortis Worldwide brand visual, converging filaments of light

Glossary

What is Model evaluation?

Definition. Model evaluation is the systematic testing of an AI model or agent against defined tasks and quality measures before and after release.

Model evaluation in practice

Evaluations check accuracy, safety, cost and consistency. They are rerun whenever the model, prompt or data changes.

Where iFortis Worldwide applies it

See AI Engineering and the Agentic SDLC for how this is used in client work.

Related terms

  • MLOps: MLOps is the set of practices for deploying, monitoring and maintaining machine learning models in production reliably.
  • Hallucination: A hallucination is an output from a generative AI model that is stated confidently but is false or unsupported by its sources.
  • Model risk management: Model risk management is the governance of risks that arise when decisions rely on models, including errors, bias and misuse.

All glossary terms

An abstract iFortis Worldwide brand visual, a lit data plane receding to a horizon

Want these ideas applied to your business?

Start with a focused opportunity assessment. We find where AI creates measurable advantage and what it takes to get there.

Contact us Submit an RFQ