This isn't something that's been documented publicly in detail yet, so rather than guess at an answer, here's what the underlying design tells us, along with what's still worth confirming directly with your Hyring contact.
What is known for certain is the underlying design intent: the AI report and the human scorecard are kept as two separate documents rather than merged into one final number, specifically so that your team debates the same evidence instead of relying on memory. That suggests the human scorecard is meant to function as your team's actual point of view, sitting alongside the AI's output rather than being silently overridden by it or replaced by it. That's a reasonable read of the intent behind the design, but it isn't the same thing as a documented, published answer to exactly what happens mechanically when the two disagree in practice.
A few specific questions are worth raising with your Hyring contact directly if this matters for your process. Does the human scorecard function as the de facto final word in practice, since it's a separate, human-authored document that naturally takes precedence when it conflicts with the AI's read? Is there a structured way to flag or annotate a specific disagreement, such as a comment tied to one particular AI-assigned score, versus writing free-text notes in the scorecard that happen to contradict the AI report elsewhere? And is the AI's score ever purely informational in your records, or does it factor into any downstream calculation, ranking, or ATS sync in a way you'd need to correct or flag if you believed it was wrong?
The honest, current answer is that the AI report and human scorecard live side by side by design, but the specific mechanics for resolving a disagreement between them haven't been published yet. If this is important for how your team plans to use Hyring Meet, ask your Hyring contact for a direct, written answer before building a process around an assumption either way.
