Performance review season comes around. The team writes self-assessments under duress. Managers spend a weekend writing reviews that say roughly what they would have said three months ago. The calibration meeting argues about ratings. Letters go out. Six weeks later, no one can name a thing that changed. This is the loop most companies are stuck in.
The teams that escape it run cadence, not a cycle: weekly evidence, monthly retro, quarterly calibration.
The three cadences
A working coaching and feedback system runs on three cadences at once:
- Weekly, in the 1-1. Evidence-led, lightweight, two or three concrete items each side. This is where most of the actual feedback lives.
- Monthly, at the team level. A short retrospective on team performance. What shipped, what stalled, what we are learning. The manager runs this with the team, not about it.
- Quarterly, at calibration. Cross-team conversation about levels, scope, and progression. Grounded in the evidence accumulated weekly, not in narratives written the night before.
The quarterly cycle most companies bet everything on is the easiest to get right and the least useful one in isolation. Without the weekly and monthly cadences underneath, calibration is theatre.
What "evidence-led" means
The phrase sounds heavy. The practice is light.
In a 1-1, both sides bring two or three pieces of recent evidence. A shipped artefact, a decision, a customer interaction, a moment of friction. The conversation starts from evidence, not from feeling. It produces specific feedback because it is grounded in specific work.
The manager's job is to do three things with each piece of evidence:
- Notice the pattern. This is the third decision like this in two weeks. Or, this is unusual for you.
- Interpret it. Here is what I see. Here is what I might be missing.
- Commit to a next move. What we will both do differently before the next 1-1.
Three sections. Same template every week. Eight weeks of this changes a team more than any review cycle.
Where AI fits without flattening the craft
There are three honest places for AI in a coaching system. None of them is delivering the feedback.
Capturing evidence. AI drafts, summaries, and transcripts mean the manager spends their time on judgment rather than note-taking. The 1-1 notes write themselves; the manager reads, edits, and signs. Time freed goes back into thinking about the person, not the paperwork.
Pattern detection. Across a team or a quarter, AI surfaces themes a single manager would miss: the same kind of slippage in three different reports, a theme in stakeholder feedback that no one person noticed. The output is a prompt for the human to investigate, not a verdict.
Practice scenarios. Managers learning the craft of feedback need reps. AI generates realistic scenarios from the team's actual work patterns: the report who is over-committing, the senior IC drifting from the role, the pair stuck in a disagreement. Practising the conversation before having it raises the quality of the real one.
What AI must not do: generate the feedback itself. The moment the manager is rubber-stamping a model's verdict, the relationship has changed and the person on the other side knows.
The smallest version that works
If your team has nothing today, do not build the whole system. Build the 1-1 template:
- Evidence: what happened this week.
- Interpretation: what it means.
- Commitment: what we agree to do differently.
Run it for eight weeks. Add nothing. Most systems fail because they start with the calibration meeting, not the 1-1. Get the 1-1 right and the rest of the system has somewhere to stand.
How this connects to AI-native People work
Coaching and feedback are the two systems that decide whether AI capability actually sticks across a function. The team that has weekly evidence-led 1-1s will surface AI experiments, share patterns, and standardise practice without needing a separate enablement programme. The team that does not will have isolated power users and no compounding.
This is why the AI-native People team work and the enablement operating model both depend on a working coaching system underneath. AI craft becomes part of how managers develop their reports, alongside every other craft. No separate track.
What this connects to
Auto-recommended next reads in the People Ops cluster, ranked by shared concepts and headings:
- Designing values that stick
- An AI policy blueprint for People teams
- Choosing AI models for HR work
- Leading the AI transformation in People
Common questions
- Why do most performance systems fail to compound?
- Wrong cadence, not wrong intent. By the time a quarterly review comes round, the specific moments that mattered are gone, replaced by a fuzzy impression assembled the night before. Feedback has to land within days of the moment it describes, or it stops being feedback and becomes an opinion. Weekly cadence is what keeps the evidence fresh enough to be useful.
- What is an evidence-led 1-1?
- A 1-1 where both sides show up with two or three concrete items, not vague impressions: a shipped artefact, a decision, a customer interaction, a moment of friction. Weak evidence sounds like 'things feel slower lately.' Strong evidence names the artefact and the date. The stronger the evidence, the shorter and more useful the conversation.
- Where does AI fit, without flattening the craft?
- Three places, all of them upstream of the actual conversation: capturing evidence so the manager isn't taking notes, surfacing patterns a single manager would miss across a team, and running practice scenarios for managers still building the craft. None of them touch the moment itself. A simple test: if a report could tell the feedback came from a model rather than their manager, AI has gone too far into the loop.
- What is the simplest version of this system to start with?
- Three sections, same order every week: evidence, interpretation, commitment. Run it for eight weeks with nothing else added. You'll know it's working when the interpretation section gets shorter over time, both sides are already reading the evidence the same way before the conversation starts. That's the signal to add the monthly retro on top.
- How does this connect to the AI-native People team work?
- Coaching and feedback decide whether AI capability sticks, or stays with the two or three people who taught themselves. The tell is in the 1-1: if AI use never shows up as evidence (a draft it sped up, a pattern it caught), it isn't part of the system yet, it's still a side project. Once it shows up as evidence like anything else, standardisation has already started.
When reading turns into doing
The Grain Audit maps one People Ops process end to end, ranks the highest-return automations, and hands you a 90-day plan you keep whether or not we work together.
Two weeks. GBP 2,000, credited in full against a programme. Three slots a month.
Book a Grain AuditIf this resonated, there's more.
Subscribe to receive new Intelligence pieces as they're published. No noise, just the work.
By subscribing you agree to our Privacy Policy. Unsubscribe any time.
