Interview scorecard template

Personal AI Trainer interview scorecard

Evaluate Personal AI Trainer candidates across 4 weighted areas: subject and technical command, how they actually teach, group management and safeguarding, and progress and communication. Subject and technical command leads at 30%, so check fluency across ChatGPT, Claude, Gemini and Copilot: prompt patterns, custom GPTs, retrieval over personal files, API. Use the rubric to compare role-specific evidence consistently.

See AI scoring
education coachingai literacyllm toolingone to one coachingprompt engineering
TL;DR
For subject and technical command, look for evidence the candidate names specific models and versions, shows working custom GPTs or automations built, and explains where each tool fails rather than overselling. For how they actually teach, look for evidence the candidate describes a repeatable session structure, starts from the learner's real tasks, and adapts pace when someone freezes at the keyboard. Apply the written 1–5 anchors to every answer, record the evidence behind each rating, and use the factor weights to reach a consistent overall assessment.
Complete evaluation framework

What to assess and how to score it

Review the evidence signals before interviewing. Then use the anchored descriptions—not instinct alone—to choose the score that best matches each answer.

01
Evaluation factor

Subject and technical command

30% weight

Check fluency across ChatGPT, Claude, Gemini and Copilot: prompt patterns, custom GPTs, retrieval over personal files, API basics, and honest framing of hallucination limits.

Evidence to listen for

  • Knows the material or discipline well beyond the level they teach
  • Holds the qualifications, licences, or certifications required, and they are current
  • Can answer an unexpected question honestly rather than bluffing
  • Keeps learning in their own field

Five-point scoring guide

1
Poor

Knowledge barely ahead of the learners; bluffs when asked something unexpected.

2
Needs Improvement

Adequate on the syllabus only; gaps show under questioning.

3
Satisfactory

Solid command of the material for the level taught.

4
Very Good

Depth well beyond the taught level; comfortable being asked anything.

5
Excellent

Names specific models and versions, shows working custom GPTs or automations built, and explains where each tool fails rather than overselling.

02
Evaluation factor

How they actually teach

30% weight

Probe their session format for a non-technical learner: diagnostic intake, live screen-share practice, worked prompts on the client's own documents, homework between sessions.

Evidence to listen for

  • Describes how they teach a specific thing, not their philosophy of teaching
  • Adapts when the first explanation does not land
  • Checks understanding rather than assuming it
  • Differentiates for varying ability within one group

Five-point scoring guide

1
Poor

Only philosophy, no method; one explanation and no plan B.

2
Needs Improvement

Delivers content but cannot adapt when learners are lost.

3
Satisfactory

Sound instruction; differentiation is limited.

4
Very Good

Multiple routes to the same concept, with real checks for understanding.

5
Excellent

Describes a repeatable session structure, starts from the learner's real tasks, and adapts pace when someone freezes at the keyboard.

03
Evaluation factor

Group management and safeguarding

25% weight

Assess how they handle client confidentiality: what data goes into a prompt, enterprise versus consumer accounts, chat history settings, and coaching minors or vulnerable adults.

Evidence to listen for

  • Has a concrete approach to a disruptive individual and a whole group losing focus
  • Knows the safeguarding or duty-of-care obligations for this age group and setting
  • Escalates a concern through the right channel
  • Sets boundaries without becoming punitive

Five-point scoring guide

1
Poor

No behaviour strategy; unaware of safeguarding obligations.

2
Needs Improvement

Struggles with disruption; vague on duty of care.

3
Satisfactory

Manages a normal group; less confident with serious disruption.

4
Very Good

Confident group management with clear safeguarding awareness.

5
Excellent

Sets clear data rules before any live demo, spots regulated or personal data in a client's file, and holds the line politely.

04
Evaluation factor

Progress and communication

15% weight

Look for how they evidence progress: baseline task timings, before and after prompt quality, tools adopted after 30 days, renewal or referral rates.

Evidence to listen for

  • Measures whether learners actually improved, not just attended
  • Gives feedback that changes what someone does next
  • Communicates with parents, clients, or managers about difficult progress honestly
  • Records progress usefully

Five-point scoring guide

1
Poor

No sense of whether anyone improved.

2
Needs Improvement

Tracks attendance rather than progress; avoids hard conversations.

3
Satisfactory

Monitors progress; feedback is general.

4
Very Good

Measures progress concretely and handles difficult conversations well.

5
Excellent

Tracks concrete learner metrics such as hours saved per week, and reports them to the client with examples from their own workflow.

Evidence-led prompts

Interview questions for a Personal AI Trainer

Use these prompts to surface evidence for the weighted factors above and compare candidates against the same role-specific criteria.

  1. 01

    Do you have experience implementing AI in real-world applications?

  2. 02

    What is your level of expertise in Python and other programming languages?

  3. 03

    How comfortable are you with machine learning and deep learning?

  4. 04

    What is your approach to teaching AI concepts?

  5. 05

    How would you explain artificial intelligence to a complete beginner?

See the complete Personal AI Trainer question set
Put this rubric to work

Score every candidate against the same standard

Add these weighted factors to Hirevire and let AI evaluate recorded answers against your rubric.

Explore AI Scorecards