Assess candidates with written tasks.
Even when they use AI.

Yes. You can design a written assessment in which AI use is expressly permitted and the candidate remains responsible for the work they submit.

Set the rules before candidates start: what the exercise is testing, which tools are available, what they must produce and how their work will be assessed. Use criteria relevant to the role and brief your assessors consistently.

In Meritia, that can mean reviewing a recommendation together with the AI conversation behind it. What context did the candidate provide? Did they check the answer against the evidence? What did they challenge? How did their final response reflect the task?

Assessors can review the prompts and responses in Meritia’s built-in AI conversation alongside the candidate’s submitted written work. This gives your panel specific evidence to discuss when assessing a polished final answer.

For example, a candidate may ask AI for a launch recommendation, spot that it has overlooked unresolved issues, and ask it to reconsider. The panel can review that exchange and the final recommendation together.

This record covers AI used inside Meritia. It does not show activity in a candidate’s personal ChatGPT or Claude account, another browser or another device.

A simulated workplace message can bring a deadline forward. A priority email can introduce a new constraint. The candidate needs to decide what matters now, update their approach and respond.

Your assessors can consider the AI conversation and submitted response alongside the new information: did the candidate account for it, challenge an unsupported suggestion and explain what should happen next?

These messages and emails are part of the assessment simulation. They do not require a live Microsoft Teams or Outlook integration.

Start with a role where interpreting evidence, explaining a recommendation and responding to changing information matter. Possible scenarios include operations and project delivery, commercial and customer teams, product or business analysis, finance and legal work.

The task, evidence and assessment criteria are agreed for the role. These examples describe possible applications; the fit for your role is part of scoping the pilot.

AI does not mark any part of Meritia’s assessment. Human assessors review the work against agreed criteria and award every mark. Your hiring team makes the hiring decision.

The AI assistant is a tool available to the candidate within the exercise. Its answer is something the candidate may need to question, check or improve.