Home · Business · Operations & Logistics · Scheduling & rotas

PARTLY

As of 13 August 2026, AI can only partly balance working hours between staff.

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 supplied tool data.

If this goes wrong: staff receive an unfair or impractical rota, coverage fails, and you have to repair the schedule and handle the resulting dispute.

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 rota, staff availability, leave record and demand or coverage plan for the relevant period.
    2. Create one table for required shifts with the date, start time, end time, location, role and number of people needed.
    3. Create a second table for each staff member showing role, skills, available times, leave, contracted or target hours, maximum hours and any fixed commitments.
    4. Add the fairness information that matters, including previous weekend or late shifts, undesirable shift definitions and preferences, while using staff IDs instead of unnecessary personal details.
    5. Paste both tables and the workplace rules into the prompt, then ask the chatbot to produce the rota, hour totals, conflicts, assumptions and fairness rule.
    6. Compare every proposed assignment with the original availability, skills, leave, coverage and hours tables, and check the totals using a spreadsheet.
    7. Ask a manager or colleague who knows the workplace to check rest, safety, equality, contractual and operational consequences before publishing the rota.
    8. Send the approved rota through your normal staff scheduling channel and record any manual changes and the reason for each change.

    Prompt

    Create a balanced staff rota from the data below. Treat the stated shift requirements, staff skills, availability, leave, contracted hours, maximum hours, minimum rest rules, fixed commitments and stated preferences as constraints. Do not invent availability, qualifications, demand or employment rules. Aim to distribute total hours and less desirable shifts fairly, but show the exact fairness rule you used. Produce: (1) a rota grouped by date, shift and staff member, (2) each person's total hours and shift count, (3) a comparison of each person's assigned hours with their contracted or target hours, (4) a list of uncovered shifts, conflicts and constraint breaches, and (5) any assumptions or information still needed. Do not silently resolve conflicts. If the data cannot support a fair rota, say why and give the smallest changes needed. Use only the information supplied here.
    
    Dates and planning period: [insert dates]
    Required coverage by date and shift: [paste table]
    Staff names or IDs, roles and skills: [paste table]
    Availability and unavailable times: [paste table]
    Leave and fixed commitments: [paste table]
    Contracted or target hours: [paste table]
    Maximum hours and rest requirements supplied by the employer: [paste rules]
    Shift lengths and undesirable-shift definitions: [paste details]
    Preferences and previous undesirable shifts: [paste details]
    Other workplace rules: [paste rules]

    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 your demand forecast is realistic or whether a thinly covered shift is operationally safe.
  • AI cannot decide how to trade equal hours against preferences, continuity, skill mix and unpopular shifts without a rule you provide.
  • AI cannot reliably establish that the proposed rota complies with every employment, contractual or workplace rule affecting your staff.
  • AI cannot take responsibility for the effect of a disputed allocation, missed cover or staff member working beyond a safe limit.
  • AI cannot replace a manager's knowledge of relationships, reliability and circumstances that are not recorded in the rota data.

What caps this at PARTLY: legal accountability, judgement under ambiguity 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 make a fair staff rota?
Partly. It can produce a rota and calculate hours from structured data, but you must define fairness and check availability, skills, rest, contracts and coverage. A manager still decides whether the result is workable and fair.
Can AI make sure everyone gets the same hours?
It can compare assigned hours with contracted or target hours and highlight differences. It cannot decide whether equal hours are appropriate when staff have different contracts, availability, skills, leave or previous unpopular shifts.
Can AI check if a staff rota is legal in the UK?
It can flag apparent clashes against rules that you provide, but it cannot reliably confirm full UK employment compliance. Not professional advice. Ask your HR adviser or an employment solicitor about a serious or disputed case.
What information does AI need to balance staff hours?
Give it required coverage, shift times, staff skills, availability, leave, contracted or target hours, maximum hours, rest requirements, preferences and previous undesirable shifts. Remove unnecessary personal information and make the model list any missing data rather than guessing.

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.