What is a practicum?

A practicum is a period of supervised practical work in real conditions, assessed against a standard, that a learner completes before being certified. It moves assessment from the classroom to the job. The learner does the actual work, on actual cases, while someone qualified watches the results and decides whether they meet the bar.

Why does a practicum matter for AI agent oversight?

Reviewing an agent's work is a judgement skill, and judgement only shows under real conditions: a full queue, ambiguous evidence, pressure to keep moving. A written test can show someone knows that a declared income figure needs verifying. A practicum shows whether they verify it at four in the afternoon on their fortieth item.

It also tests consistency. One good decision proves little. Two weeks of decisions show a pattern, and the rare serious error, the one that matters most, is more likely to surface.

How does a practicum work in practice?

A practicum for agent reviewers usually runs like this:

1. The learner is given review rights under supervision, on the queue they will work.
2. Each decision is logged: the item, the verdict, the decision, the reason.
3. An assessor samples or checks every decision against the right answer.
4. Red lines are defined in advance: errors that fail the practicum on their own.
5. At the end, the assessor reviews accuracy and red lines and signs off, or sets more practice.

During the practicum, a second person should check anything the learner approves that cannot be undone. The learner's decisions are real; the safety net is too. Logs make the process work, and ready formats exist for customer support and collections.

What is an example?

An internal audit team adopting an agent that drafts control test results runs a two-week practicum for each auditor. Each day, auditors review the agent's drafts, accept, edit or reject them, and log why. The assessor checks every decision. One auditor's accuracy is high, but she accepted a test result based on a screenshot rather than the system's own export. That is a red line, so she repeats a week. The approach is set out in how to run a two-week practicum for internal auditors, with a payments version in how to run one for payments operations staff.

How is a practicum different from a capstone?

A capstone is one built deliverable, judged as a whole: for example, an agent the learner designed that must pass certification. A practicum is many real decisions over a period, judged for accuracy and consistency. Capstones suit builders; practicums suit reviewers and operators, whose job is decisions, not designs. A practicum is also different from shadowing, where the learner watches but does not decide. A fuller treatment is in what a practicum is in AI agent training.

Which terms are related?

• Work-based assessment: the wider category a practicum belongs to.
• Red line: an error that fails the practicum regardless of accuracy.
• Assessor: the person who signs it off.

For the wider topic, read AI literacy and certification for people who work with agents.

Where the Academy fits

Certified Pryme Operator in Pryme Intelligence Academy includes a two-week practicum of real review decisions: at least 95 in every 100 correct, no missed red line, signed off by a Pryme assessor. It follows about 14 hours of learning across eight modules and a written exam.