Home · Business · HR & People · Performance & training

PARTLY

As of 13 August 2026, AI can only partly create a coaching plan for an employee.

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

Can you do it?

5 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 costsNo comparable human-service price is provided in the available tool data.

If this goes wrong: the plan misreads the cause of the problem or applies inconsistent expectations, damaging trust and potentially contributing to an unfair employment process.

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 your current role description, objectives, relevant workplace policy and any existing performance notes, and remove unrelated personal information.
    2. Gather dated, factual examples of the work or behaviour that needs development, the expected standard, the employee's strengths and any support already offered.
    3. Ask the employee what they think is affecting performance and record agreed constraints or training needs without presenting assumptions as facts.
    4. Paste the role, evidence, expectations, support, timescales and policy details into the prompt, then ask the chatbot to produce the coaching plan.
    5. Compare every objective and proposed action with the role requirements, your evidence and the current workplace process, and delete anything the source material does not support.
    6. Ask a colleague or HR contact to check the plan for consistency, confidentiality, reasonable support and any risk that it could be treated as an undisclosed disciplinary process.
    7. Discuss the checked draft with the employee, amend agreed actions and dates, then send the final plan and schedule the review meetings.

    Prompt

    Create a practical coaching plan for an employee in the UK. Use only the information supplied below and mark any missing information as a question rather than guessing. This is a development plan, not a disciplinary decision or legal opinion.
    
    Employee role: [role and level]
    Relevant responsibilities: [responsibilities]
    Development goal: [specific outcome]
    Observed behaviour or performance gap: [dated, factual examples]
    Expected standard: [role requirement, target or agreed objective]
    Strengths to build on: [strengths]
    Possible causes or constraints already discussed: [causes or constraints]
    Support available: [training, shadowing, mentoring, resources]
    Time available: [time]
    Review period: [period]
    Relevant workplace policies or process: [policy or process]
    
    Produce:
    1. A short, neutral summary of the development need, separating facts from assumptions.
    2. Two to four measurable coaching objectives linked to the expected standard.
    3. For each objective, specific actions for the manager and employee, support needed, evidence of progress, owner and target date.
    4. A simple review schedule with questions for each meeting.
    5. Risks, reasonable alternatives and information that must be checked with HR before using the plan.
    6. A short opening script for discussing the plan respectfully with the employee.
    
    Do not diagnose the employee, invent performance evidence, recommend unlawful treatment, label the plan as disciplinary, or make a final judgement about capability. Keep the tone direct, supportive and suitable for sharing with the employee.

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

What it gets wrong

  • AI cannot know whether the recorded examples show a genuine capability issue, unclear expectations, inadequate support or a wider workplace problem.
  • AI cannot decide what is fair and proportionate for this employee or how the plan fits your organisation's employment process.
  • AI cannot obtain the employee's trust, hear tone and body language, or repair a difficult coaching conversation.
  • AI cannot take responsibility for confidentiality, equality considerations, reasonable adjustments or the consequences of using the plan.

What caps this at PARTLY: judgement under ambiguity, legal accountability and context depth.

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

FAQ

Can ChatGPT create a coaching plan for an employee?
Yes, it can draft a useful plan from factual examples, role expectations, objectives and available support. You still need to check the plan with the employee and, where appropriate, HR because AI cannot judge fairness or take responsibility for the employment process.
What information does AI need to create a coaching plan?
Give it the role requirements, the specific development goal, dated observations, the expected standard, strengths, constraints, available support and review period. Do not include unnecessary personal or sensitive information, and tell it to mark gaps rather than guess.
Is an AI coaching plan legally compliant in the UK?
AI cannot confirm that a plan is legally compliant or suitable for your organisation's process. Check it against your workplace policies and ask HR or an employment solicitor about serious cases, discrimination risks, reasonable adjustments or any plan that may become disciplinary.
Should I let an employee know AI helped write their coaching plan?
The important point is that you own and understand the plan, and that it is based on accurate evidence and discussed openly. Follow your organisation's rules on AI and confidentiality, and do not present unverified AI wording as a decision made by an impartial system.

Nearby answers

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.