Blog · Product notes
How the evaluation draft shows its evidence
2 min read
Every dimension of an AI-drafted evaluation in Nextdot shows how much evidence stands behind it. Here is what that means, and what the leader does with it.
Drafted per dimension of your framework
The draft follows the organization's own performance framework: its dimensions and weights, its key skills, and the technical skills of each job role. It is built from the evidence on record: the dotsDot: Work worth talking about — time-bound, owned by whoever's doing it, open for conversation while it happens. a person kept, the feedback on them, and goal reviews, which are read like feedback.
Evidence strength, in four factors
Next to each dimension, the draft shows how much evidence stands behind it, from four factors:
- Volume: how much evidence there is.
- Variety of sources: whether it comes from different tools and people.
- Consistency: whether it points the same way.
- Relevance: how closely it relates to the dimension.
A dimension with thin evidence says so before anyone reads the text. That is where a leader should look hardest, and where a conversation is more useful than a rating.
The leader writes the evaluation
AI drafts; it does not decide. The leader reads the draft and writes their own evaluation, and nothing is final until they sign it. AI-generated content is labelled as such in the product.
The person sees the same evidence
The person being evaluated sees the same evidence their leader sees, and what appears in it is what they chose to keep. When the evaluation is final, it can be exported to Excel and PDF.