Worked Example: Score Two AI Summaries
Score two AI summaries using factual coverage, unsupported additions, and decision usefulness.
You have two AI summaries of the same meeting. The source says conversions fell after a pricing page change, but paid traffic also increased and the team has not segmented results yet. Summary scorecard: coverage, faithfulness, usefulness The polished summary blames pricing as the cause because it makes a cleaner story. Coverage Check whether each summary includes both facts: pricing page changed and traffic mix changed. A summary that omits traffic mix hides an important alternative explanation. Faithfulness Mark any added causal claim that the source did not prove. Blaming pricing may be plausible, but the source does not establish…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in