Keeping it safe · for operations and project managers

Using AI safely in operations and project management

Operations and project managers handle confidential material every day: vendor contracts, employee schedules, headcount plans, and incident reports. Much of it belongs only in AI tools the company has approved for that data. A contract may forbid sharing its terms. Privacy law may cover employee data.

Keeping it safe also covers decisions. AI can help analyze a staffing schedule or a vendor proposal. It should not decide who works which shift or which vendor is dropped.

Where it comes up

  • Comparing vendor contracts

    Check whether the contracts include confidentiality clauses, and use only a tool approved for confidential commercial material. If none is approved, extract the terms you need into a table without vendor names.

  • Analyzing a staff schedule

    Replace names with role codes before analysis. Remove notes about leave reasons, health, or performance. Those details are not needed to find coverage gaps.

  • Writing up a workplace incident

    AI can help structure the report from your notes. The facts, the account of what happened, and any finding of fault must come from the people involved and the approved process.

What it looks like at each band

  1. E1 · Foundational

    Understands the rule about confidential material in principle. Misses it in practice when it sits in a rota column or a forwarded contract. Applies the policy after someone points it out.

  2. E2 · Proficient

    Strips names and personal details from employee data before every analysis. Puts contracts and internal plans only into approved tools. Recognizes decisions that should stay with a person, such as staffing and disciplinary matters.

  3. E3 · Distinguished

    Judges risk across a whole project, including data, contract terms, and the effect of decisions on people. Raises concerns at kickoff, not at the final review. Recognizes tasks that belong outside AI tools entirely.

  4. E4 · Exceptional

    Writes the team’s rules for AI use, including the approved tool for each kind of data and the decisions a person must make. When an error happens, reports it and changes the step that let it through.

Common mistakes

  • Pasting a full HR or rota export

    Exports from scheduling and HR systems carry names, contact details, and sometimes absence reasons. Remove the identifying and sensitive columns before any analysis.

  • Uploading contracts to a personal account

    Vendor contracts often prohibit disclosure of pricing and terms. A personal AI account is a disclosure to a third party. Use the approved tool or work from an anonymized extract.

  • Letting AI rank people

    Some requests ask a model who to cut from a rota or which team member underperforms. That puts a decision about people in a tool that cannot be held to account. Keep the decision with a manager.

A worked example

Finding coverage gaps in next quarter’s shift rota for 120 warehouse staff.

A weaker attempt
Upload the full rota export, with names, phone numbers, and leave reasons, to an unapproved personal chatbot. Then let it pick which staff cover the gaps.

Employee personal data, including leave reasons, now sits in a service outside the company’s control. The model is also choosing who takes extra shifts, which is a manager’s decision.

A stronger attempt
Replace names with employee codes and delete the phone and leave-reason columns. Open the AI tool your company has approved for staff data. Request a list of the shifts below minimum cover with the count of qualified staff free for each one. Check two of the gaps against the original rota. Then decide the cover assignments yourself under the scheduling policy.

The manager still gets the coverage analysis. Personal data is removed, the tool is approved, and the decision about people stays with the manager.

The full study guide for this skill, for every role, is Keeping it safe. The other five skills for operations and project managers are on the AI skills for operations and project managers page. A one-page summary is in the AI data safety checklist.

See where you stand

The free check grades keeping it safe and the other five skills in about fifteen minutes. Nobody sees the grade but you.

Then build the skills and prove it. The three steps for operations and project managers list the courses and templates that fit the work.