self-paced

Evaluating AI Systems

Evaluating AI Systems — practical automation and agent patterns with governance.

This page is the course itself. Commercial AI training for UK companies is owned by the AI training hub.

Delivered by AI Build

Learning platform: SAVI.education

Certificate issued by SAVI.education

Duration
5 hours
Price
£99
Level
Intermediate

Who this course is for

Who it's for
People whose work sits in Automation and Agents and who want a practical, workplace-focused course rather than a generic overview.
Prerequisites
No specialist background required. A laptop and an approved workplace AI tool will help you practise the exercises.
What you'll be able to do
  • Complete “From One-Off Prompt to Repeatable Workflow” and apply it to a real task at work.
  • Complete “Human Review and Accountability” and apply it to a real task at work.
  • Complete “Deployment Monitoring And Evaluation” and apply it to a real task at work.
  • Complete “Prompt Evaluation” and apply it to a real task at work.
Duration & level
5 hours · Intermediate
  • Self-paced lessons
  • Credential

Course syllabus

Lessons are grouped into clear steps with visual cues, so each course is easier to scan before you start.

Lessons unlock after purchase. Create your account with the same email you paid with — no course code is required — or choose “Already purchased?” above.

  1. Step 1 · Lesson

    From One-Off Prompt to Repeatable Workflow

    Turn useful prompts into reusable, governed workplace workflows.

  2. Step 2 · Lesson

    Human Review and Accountability

    Keep humans accountable for decisions supported by AI outputs.

  3. Step 3 · Lesson

    Deployment Monitoring And Evaluation

    Learn how to take an AI application from prototype to production with clear deployment patterns, monitoring, evaluation and rollback plans. You will be able to define practical quality, cost, safety and reliability controls for RAG, agentic and API-based AI systems.

  4. Step 4 · Lesson

    Prompt Evaluation

    Learn how to evaluate prompts systematically rather than relying on intuition or one-off tests. You will be able to define evaluation criteria, build small test sets, compare prompt versions, and create a repeatable evaluation workflow suitable for enterprise use.