Frequently Asked Questions
Q1: What pattern in my own speech did AI uncover that I hadn’t noticed for years?
The AI flagged that I use hedge phrases like ‘kind of’ and ‘sort of’ an average of 12 times per 30-minute meeting — way more than I thought. My one-on-ones with my manager had 47 instances of self-diminishing language over two weeks. When I mentioned this discovery, my manager said she’d noticed it in performance reviews but didn’t want to bring it up directly. Cutting those phrases improved how my ideas landed in meetings.
Q2: How did AI analysis change my actual meeting behavior?
I started tracking talk time ratios after the AI showed I spoke only 31% of meeting time in my own team syncs. Within a week I consciously spoke up earlier in discussions instead of waiting to see consensus first. My contribution percentage went from 31% to 44% by week two. My teammates noticed I was driving decisions more instead of just附和.
Q3: What specific AI tool worked best for reviewing meeting transcripts?
I used a combination — Otter.ai for the raw transcription, then Claude for the pattern analysis. Otter gave me the timestamps and speaker separation, which I fed into Claude with a custom prompt asking it to count filler words, measure talk time, and surface repeated topics. The total cost was about $20 for Otter Pro plus free Claude access. A dedicated tool like Fireflies.ai has some of this built in, but the custom combo gave me deeper insights.
