For managers, not for HR
Your performance reviews are already written.
Define what you expect from each report. They write down what they did. Impact Plan links every win to the behaviours it proves and the skills it used — and keeps a growth plan and a performance review current all year, so review season is reading a page instead of writing one.
Free forever for up to 5 reports. No credit card. Your reports never pay.
Nobody trained you for the hardest part of the job.
You were promoted for being good at the work. Then you were handed six careers to steer, two review cycles a year, and no system at all.
You cannot remember March.
By November you are rating a year on the strength of the last six weeks. That is not unfairness, it is a memory problem — and memory problems have a fix.
Review season eats a weekend.
Eight to sixteen hours, twice a year, spent reconstructing from Slack scrollback what you already watched happen.
"Where is my career going?"
The question every good report eventually asks, and the one you cannot answer with anything better than encouragement.
Nobody wrote down what good is.
You know it when you see it. Your reports cannot read your mind, and "be more senior" is not feedback anyone can act on.
How it works
Three moving parts, and two of them are not yours.
You set the expectations. Your reports supply the evidence. Impact Plan does the filing.
Say what you expect
Write the handful of behaviours you actually want from each person, or let AI draft them from the role and edit what it gets wrong. Specific, observable, and visible to the person they apply to.
They log what they did
A win takes two minutes to write down in the moment and is impossible to reconstruct nine months later. You approve it in one click — or write it yourself, which is how most managers start.
Impact Plan maps it
AI reads each win and suggests the behaviours it proves and the skills it used; you pick the goals it moved. The growth plan and the review update themselves as the evidence lands.
Say what good looks like, per person
Expectations are yours, not a competency framework somebody in HR wrote in 2019. Write four or five for each report, or generate a draft from the role in one click and cut the ones that do not apply.
- Personal expectations for one report, alongside the ones the role carries
- AI drafts six to eight observable behaviours from the role in seconds
- Three honest states: not demonstrated, developing, consistently demonstrated
Wins get filed the moment they happen
Your report writes two lines in the IMPACT + HOW format — what changed, and how they did it. AI reads it and suggests which of your expectations it proves and which skills it used. You approve, and it is on the record with a date on it, permanently.
- AI-suggested skill and behaviour links, which you accept or drop
- One-click approval from your inbox; anything you write is approved on save
- Link a win to the goal it moved, so contribution is checkable rather than claimed
A growth plan that updates itself
Pick where someone is heading — up, sideways, or into a stretch — and you get a readiness percentage and the list behind it: which expectations are still missing, and which skills would enable them. It moves every time a win lands, which means you never build one by hand.
- Readiness as one number and a colour, not a twelve-page scorecard
- Split into what this level asks for and what the next one will
- The evidence that moved the number is one click away from the number
The review is written before you need it
Because the evidence went in as it happened, the review already exists: what they did, when, which expectations it met, which skills it proved, and which goals it moved. You read it, add your judgment, and you are finished. No weekend, and no reconstructing a year from memory.
- Every claim traces back to a dated, approved piece of evidence
- Walk into calibration with a case instead of an adjective
- Nothing in it is a surprise, because your report watched it accumulate
One-to-ones that remember
A monthly check-in that opens with everything since the last one already on it: the wins, the expectations that moved, and the things you both agreed to last time. Agreements belong to the person rather than to one meeting, so an open one is still there next month whether or not anyone mentioned it.
- Talking points raised by either side, carried forward until they are dealt with
- Agreements with a deadline, closed as done, not delivered, or cancelled
- A moved deadline keeps the one it moved from, so "we said the 15th" stays checkable
Agents
Run it without opening it
Impact Plan speaks Model Context Protocol: 74 tools, 31 of them writes. Connect Claude or any MCP client and capture a win on the walk back from a meeting, ask what a report has done since June, or have your assistant draft the check-in before you sit down.
- Same permissions and same audit log as the screens — the agent is you, not a back door
- OAuth with per-token scopes, so you can hand out a read-only connection
- Every feature here ships its tools in the same release as its screens
It has to be worth their time too
A tool your reports resent is a tool that stops being filled in by March. So the deal runs both ways: they get their work on the record instead of in your memory, a growth plan specific enough to act on, and one-to-ones that pick up where the last one stopped.
Nothing here is scored in secret. Everything you can see about a report, that report can see about themselves.
What your reports get
Pricing
Free for five. €29 for the rest.
Up to five reports is free forever, with every feature switched on. Above that it is €29 per manager per month, however many people you look after. Your reports never pay, whatever the plan.
Frequently asked questions
The things managers ask before they sign up.
No. There is no payroll, no headcount reporting, no compliance module, and nothing a People Ops team has to configure before you can start. Impact Plan is a tool you set up for your own team in an afternoon. It keeps records good enough to hand to HR at review time — but HR is not in the way of you using it.
One person who reports to you and has an active account. Up to five of them is free forever. People you have deactivated do not count.
No. Pricing is per manager. Your reports get full accounts at no cost, however many of them there are — which is deliberate, because a tool that charges by the head is a tool you end up rationing.
Then you write the wins yourself, which is how most managers start and is still far less work than reconstructing them in November. Anything you write is approved the moment you save it. Most teams flip round within a month or two, once people see their own record actually being read in a check-in.
Yes, and most people do. Impact Plan is where the year's evidence accumulates. When the company's tool opens its review window, you copy out of a page that is already written instead of starting from a blank form.
Yes. Impact Plan exposes 74 Model Context Protocol tools, 31 of them writes, so Claude or any MCP client can capture achievements, run check-ins, assess expectations and read a growth plan on your behalf — under the same permissions and the same audit log as the screens.
The person it is about. Expectations, assessments, growth plans and evidence are all visible to the report they concern. There is no private manager scratchpad, on purpose — a record somebody cannot read is a record they cannot correct, and it would surface for the first time in a review.
Stop writing reviews from memory.
Set up your first report in about ten minutes. Free forever for teams of five.
No credit card. Your reports never pay.