Home · Business · HR & People · Performance & training

PARTLY

As of 13 August 2026, AI can only partly identify employee training needs.

This still needs a person who signs their name to it.

Can you do it?

15 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0

Skill neededchat-fluent

Who has to check ita colleague

What the alternative costsThe supplied commercial information gives no price for a human alternative, so no pound figure is stated.

If this goes wrong: you direct training at the wrong people, miss a capability problem or treat a management issue as an employee deficit.

What to actually do

  1. Hand it to a person

    The route this page recommends

    A person who owns the outcome does this end to end, worth it when the failure is dear.

  2. Use a tool built for this

    Second choice

  3. Do it yourself

    The distant third

    A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

    How to actually do it

    1. Open the current role description, competency framework and team objectives, then copy the relevant requirements into the role-requirements section of the prompt.
    2. Gather recent review notes, quality checks, work samples, customer feedback and the employee's own assessment, remove unnecessary personal data, and paste the dated evidence into the prompt.
    3. Add completed training, coaching already offered, available resources and any known changes to the role so the model can distinguish a training gap from a resource or process problem.
    4. Paste the completed prompt into ChatGPT, Claude or Gemini and ask it to produce the evidence table and clarification questions.
    5. Compare every suggested gap with the original records and mark claims unsupported by a specific piece of evidence as unconfirmed.
    6. Discuss the unconfirmed points and the five questions with the employee, then ask a manager or HR colleague to approve the final training plan and record its objective evidence.

    Prompt

    Identify possible employee training needs from the information below. Use only the evidence supplied and do not invent performance facts, employee characteristics or skill gaps.

Role requirements:
[PASTE THE CURRENT ROLE DESCRIPTION, REQUIRED SKILLS AND EXPECTED STANDARDS]

Evidence from work, reviews or quality checks:
[PASTE ANONYMISED, DATED EVIDENCE]

Employee self-assessment or feedback:
[PASTE ANONYMISED FEEDBACK]

Existing training and support:
[PASTE COMPLETED TRAINING, COACHING, EQUIPMENT OR RESOURCE INFORMATION]

Business priorities:
[PASTE THE RELEVANT TEAM OR BUSINESS PRIORITIES]

Produce a table with these columns: possible gap, evidence, confidence level, whether training is a plausible response, other possible explanations, recommended training or support, priority, and evidence still needed. Separate observed facts from inferences. Do not rank employees against one another, diagnose anyone, make assumptions about protected characteristics, or recommend disciplinary action. Flag any conclusion that needs a manager or HR colleague to check. End with five specific questions I should ask the employee before deciding on training.

    Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

What it gets wrong

  • AI cannot see informal workplace context, such as unclear priorities, poor equipment or a team process that makes competent work difficult.
  • AI cannot reliably distinguish a genuine skill gap from motivation, workload, health, conduct or management issues.
  • AI cannot guarantee that its suggestions are fair when the underlying records are incomplete, inconsistent or biased.
  • AI cannot take responsibility for deciding whether an employee needs training or for the employment consequences of that decision.

What caps this at PARTLY: judgement under ambiguity, private data access and legal accountability.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

AxisScore (0–2)
Output2
Inputs1
Verification1
Liability1
Effort delta2
Total7 / 10

The methodology and its thresholds are published in full.

FAQ

Can AI assess my employees' training needs?
Partly. It can organise role requirements, performance evidence and feedback into possible gaps, but a manager must check the evidence and decide whether training is the right response.
What information does AI need to identify training needs?
Give it the current role requirements, dated work evidence, employee feedback, existing training and relevant team priorities. Remove unnecessary personal data and do not rely on the model to fill missing facts.
Can AI tell if an employee needs training or performance management?
No, not reliably. It can list possible explanations and questions to investigate, but that distinction depends on workplace context and must remain with the manager and HR process.
Is it safe to use AI with employee performance data?
Use only the minimum information needed, follow your organisation's data and employment policies, and anonymise material where possible. Do not paste sensitive employee information into a service unless your organisation has approved that use and the necessary privacy safeguards are in place.

Nearby answers

All Performance & training rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.