What we are learning, while we build it.
Notes on the decisions inside the product — how a review score is composed, what anonymous has to mean, and why an AI-written lesson waits for a person. Written by the team, not by an agency.
Everything we have written.
How to practise difficult employee conversations with an AI avatar
Difficult conversations go wrong in the minute nobody planned for. How to prepare them in four steps, and how to rehearse them out loud with an AI avatar that scores every turn.
Read article
What belongs in an employee database: a checklist
Personal details, contract, organisation, payroll and tax, development: what an employee database should hold, what to leave out, and how to get from a spreadsheet to a working database.
Read article
One yardstick at a time, across the whole team
Judging a whole person in one pass invites the halo effect, where a strong impression on one measure bleeds into every other rating. So the scoring board is built the other way round.
Read article
In, out, or unresolved — the third answer that matters
Eligibility used to have two outcomes, and a blank cell quietly meant "out". Somebody could be dropped from their own performance review with no warning anywhere. Now a round refuses to launch instead.
Read article
What anonymous has to mean to actually be anonymous
A checkbox that hides a name is not anonymity. It is a promise with several ways to break it — the audit log, the job role, an error message, a reorganisation. Here is each one, and what we do about it.
Read article
Why the interview opens with a line from the candidate's own CV
The first thirty seconds decide whether a candidate treats the conversation as real. So the avatar starts with something only a reader of their CV could know.
Read article
Every AI-written lesson lands as a draft
The model can write a lesson in seconds. It cannot know that your company reorganised last quarter, or that the policy it just described was replaced. So a person publishes.
Read article
The Human Component: Why AI Needs Real Emotion to Train Real Skills
Most corporate training ignores the one thing that makes difficult conversations hard: the emotional charge. We break down the neuroscience of learning under social stress — and why only realistic practice changes behaviour.
Read article
Simulation 2.0: Our New AI Feedback Engine
How we rebuilt our response model so the AI stays in character under pressure — and why that changes everything for sales training.
Read article
Why We Founded Persohap
A letter from our founder on human connection in an automated world — and the gap no corporate tool had tried to fill.
Read article
Privacy first, and why the boring answers are the right ones
Everything sits in Frankfurt, one company's data can never surface in another's, and coaching data never reaches a performance review. The principles behind those choices, without the parts we do not publish.
Read article
What eight dimensions of a conversation actually tell you
Completion rates say someone turned up. We score every turn of a conversation across eight dimensions instead — and this is what those numbers can and cannot tell you about how a person is improving.
Read articleSee it working, not just described.
Every decision in these posts is visible in the product. We will walk you through whichever one you care about.