I’ve been evaluating Sembly for a potential rollout to our engineering teams, specifically for sprint retrospectives and post-incident reviews. The sales page highlights "industry-specific vocabulary," but I need to know how that translates for dev/ops conversations.
My main concerns are:
* **Jargon handling:** Does it accurately identify and log terms like "downtime," "rollback," "hotfix," "tech debt," "SLO breach," or specific cloud services (e.g., "Lambda timeout," "K8s pod eviction")? Or does it get confused and produce generic summaries?
* **Action item extraction:** In retros, the gold is in the "action items" and "owner assignments." How well does it pull out tasks like "Migrate DB schema by next sprint" or "Add circuit breaker to payment service" from the natural discussion?
* **Cost implication:** We're looking at the "Teams" plan. If it fails on technical context, that's a significant spend for just a meeting logger.
Has anyone stress-tested this in a real dev environment? I'm particularly interested in:
* The accuracy of meeting minutes when discussions get deep into code or infrastructure.
* Whether the "AI Meeting Assistant" asks for clarifications on ambiguous terms or just makes its best guess.
* Any configuration or custom vocabulary options that are available.
I’m about to run a trial, but real-world feedback would help me set the right expectations before I push for budget approval.