Why we moved the definition of a good candidate out of the recruiter’s head and into the role, and what changed in our own scoring.
On 2 September, two of us scored the same backend engineer for the same role. Kavita gave 4.0; I gave 2.5. We were both right — we just had different roles in our heads.
That gap is the whole reason Role DNA exists. A role now carries a written definition of a good candidate: a short summary, then Must, Nice and Exclude. Your agent drafts it from the JD and your call notes; you edit a line or two and approve it. Every score after that is against the same text.
Since we switched, our recruiters change 14% of agent scores, and most changes are half a star. Before, two recruiters disagreed by a full star or more on one in four candidates.
The first version let the agent approve its own draft. Nobody read it. Now a draft waits in the Inbox until a person approves it, and the attribution shows both names.