← All processes

Process 4 of 8

Performance Management

How much of performance management could AI take on, and where should you start? Label each task, add your hours, and watch the result.

Goals, check-ins, feedback and reviews through the year. Picks up after onboarding and runs alongside Learning and Development.

How to use this

  1. Label each taskOpen an activity. For each task, pick Human-led, Augment, Automate or Create. A bold square box marks a task where your choice differs from our starting view.
  2. Check the weightsEach task has a % share of its activity's time. We start with a guess. Change it to fit your team.
  3. Add the hoursPicture one full-time person in this area. We start at about 40 hours a week. Change any number and every total follows. Use Ignore it for activities your process does not have.
  4. Read the verdictThe verdict at the bottom combines your labels, weights and hours to show where to start.
Rules differ by country: examples from the United States and India
United States
Because employment is mostly at will, clear records of performance matter if a decision is challenged. Ratings and pay decisions must not discriminate. Some places regulate AI used in employment decisions.
India
Contracts and standing orders may set how performance and conduct issues must be handled, and fair process matters. Appraisals often link to annual pay revisions. Performance data is personal data under the DPDP Act, so limit who sees it.

These are starting pointers for comparing the two, not legal advice. Find the equivalent rules for your own country. Rules change and vary by state, company size and contract, so check the current text.

1Your activities

Ready to explore? Dive in. Open a stage below, choose who does each task, and watch the numbers move.

Performance Management activities

Open "Activity details" to pick who does each task.

Your changes are saved automatically.

How each activity gets its result

Each activity gets one result: the choice that covers most of its time. AI helps with every task that is not Human-led.

2Verdict

Our verdict on performance management

Your choices
Our starting view
  • Share of time:
  • Human-led
  • Augment
  • Automate
  • Create

What is our starting view? Our own view of each task, with the reasoning shown under it. It is a place to check, not measured data. Your choice may differ because of your organisation's own practice. Where it does, a bold square box on each task marks our starting view.

How this is calculated

AI helps with every task you do not pick as Human-led. An activity's hours are shared across its tasks by weights. We set the starting weights by hand, and you can change them. Tasks you add yourself all count the same. The big percentage is the share of time with AI involved. Each of the four boxes shows that label's share of the time, with the time itself and the number of tasks underneath. The same percentage is used on every process, so they can be compared. A tie goes to Human-led. Activities you ignore are not counted anywhere. It counts time, not job numbers. The hours are starting guesses for a general HR professional, not industry figures. Time spent checking AI's work is not taken off, so read the AI hours as a best case. A small change to a weight can move a result a little.

A caution. This is how it looks today. It comes from a first view of each activity and the linked sources, not from measuring your own process. Treat the task choices and hours as opinions to test. Before relying on any of it, try a small trial, check results for fairness across groups, tell the people affected when AI is used, and confirm local law. Ask a few of the people affected, and their managers, what they would change before you rely on any of this. This is not legal advice.

Before using AI with people
  1. Talk to your lawyer.
  2. Check local data protection rules. Many places ask for extra care with sensitive personal data.
  3. Check local rules on AI in employment. Some places treat AI used for hiring, pay, promotion or ending work as high-risk, or require bias audits and notices to people.
  4. Ask your tool providers. How do they test for bias, and what data do they keep?

These are pointers, not legal advice. Rules change, so check the current text.

3HR Tech

HR Tech reshaping these areas

Tap an area to see its products.

4Good to know

Common HR terms