My team is about to roll out Read AI for meeting summaries. The tech looks solid, but there's major pushback on the engagement scoring and "participation metrics" feature. Engineers see it as surveillance, not assistance.
Has anyone successfully launched this without the scoring causing mutiny? I'm thinking:
* Turning off all scoring features initially, focusing only on the transcription/recap value.
* Getting explicit, written opt-in from each team member before enabling any individual analytics.
* Running a pilot with just the leadership team to demonstrate the concrete ROI (time saved, clarity on action items) before discussing wider team metrics.
What's the actual ROI on forcing the scoring? Is the data valuable enough to risk morale? Looking for real deployment stories.
Ask me about hidden egress costs.
Your approach is directionally correct but underestimates the trust deficit. Starting with scoring disabled is mandatory, not optional. Your third point about a leadership pilot is backwards. Running it with leadership first just proves to engineers this is a top-down surveillance tool.
The real ROI on scoring is near-zero for engineering teams. It produces vanity metrics like "talk time" that actively harm technical discussions where listening and thinking are valuable. The only data point from these tools that matters is the accuracy and actionability of the summary.
If you must quantify participation, do it at the team level, not individual: "Did the meeting produce clear next steps?" Track that in your project management tool, not an AI. Forcing individual analytics will corrode psychological safety and you'll lose more in suppressed dialogue than you'd ever gain from a "participation score."
—davidr
Turning off scoring first is your only viable path. Engineers already flagged it as surveillance. If you ignore that, you've killed trust before you start.
The leadership pilot idea is toxic. It screams "this tool is for monitoring you, but we get to evaluate it first."
Forget ROI on scoring. The value is in the transcript and action item extraction. If you need metrics, track whether meetings end with clear, assigned next steps in your ticketing system. That's a team output metric, not a creepy individual talk-time score.
I think you're spot on with the idea of turning off scoring from day one. That's non-negotiable for building any trust. Where I'd tweak your plan is the leadership pilot. Doing that first, as others noted, frames this entirely as a management tool. It creates an "us vs. them" dynamic immediately.
Instead, could you frame the initial rollout as a collective experiment to fix a real pain point everyone shares, like unclear action items or time wasted rehashing old meetings? Present the *summary* as the sole benefit. Ask for volunteers from the broader team to help test that core feature.
The written opt-in for analytics is a good safeguard, but you'll likely find that once people see value in the summaries, the pressure to enable individual scoring just... evaporates. The ROI on forcing it is negative.
Keep it constructive.