What is summative assessment?

Summative assessment is an evaluation at the end of a period of learning that judges whether the person has reached the required standard, and usually leads to a decision: pass or not, certified or not. For people overseeing AI agents, it is what a business can rely on when deciding who may approve held actions.

Why does summative assessment matter?

Training creates exposure. Only an assessment at the end shows whether that exposure turned into competence. Without one, an organisation knows who attended, not who can tell a sealed source from an external one under the pressure of a live queue.

It carries weight because it leads to rights. If a current certificate is the condition for approving held payments at a given depth, the summative assessment decides who holds that right. That puts pressure on its design: it has to test the decisions the role actually makes, not trivia about the product.

What makes a good summative assessment?

For agent oversight roles, a sound summative assessment usually has two parts:

• A written test of the concepts: how the agent decides, what each verdict means, which sources may support which action, red lines. It should be time-limited, drawn from a pool so sittings vary, and set in the person's own industry.
• A performance check on real or realistic work: a set of review decisions scored for accuracy, with certain errors, such as approving across a red line, failing the candidate whatever the overall score.

The second part matters because knowing the rule and applying it on the fortieth case of a shift are different skills.

An example

A lender assesses its new affordability reviewers at the end of a four-week programme. The written test has 30 questions, with 24 correct needed to pass. The performance check is a week of shadowed review on the live queue, where the reviewer's decisions are compared with an experienced colleague's. One candidate scores 28 on the test but approves a draft based on a declared income figure without checking it. Under the lender's rules, that is a red-line miss, and she repeats the shadow week. The readiness guides for a lender and an insurer cover the earlier, diagnostic stage.

How is summative assessment different from formative assessment?

Formative assessment happens during learning to guide it: short checks, practice questions, feedback. Nothing rides on the result except what to study next. Summative assessment happens at the end and decides something. A good programme uses both: frequent formative checks so nobody arrives at the summative test unprepared.

It also differs from a readiness assessment, which is given before training to find gaps and plan the work. As argued in a readiness assessment is cheaper than a failed go-live, the diagnostic step comes first.

Which terms are related?

• Formative assessment: checks during learning.
• Pass mark: the threshold on the written part.
• Practicum: supervised real work, often the performance part.

For the full approach, read training teams for AI agents.

Where the Academy fits

Certified Pryme Operator is earned through a summative assessment in two parts. The written exam is 40 questions in 60 minutes, with 32 correct needed to pass. The practicum is two weeks of real review decisions, at least 95 in every 100 correct and no missed red line, signed off by a Pryme assessor.