Having recently been tasked with evaluating meeting intelligence platforms for a non-technical executive team, I spent considerable time with Sembly. While its core transcription and AI note-taking features are competent from a technical standpoint, I must concur with the sentiment implied by the thread title. The user interface presents significant cognitive load for users who are not accustomed to navigating tooling with a developer or power-user ethos.
My primary observation is the lack of a clear, linear workflow. The dashboard presents multiple entry points—"Meetings," "Notes," "Tasks," "Insights"—without a strong visual hierarchy guiding a new user on *what to do next*. For a team migrating from a simpler solution like Otter.ai or even built-in Teams transcription, the leap in UI complexity is substantial. The paradox is that the tool aims to simplify meeting follow-up, but the interface itself requires a period of acclimatization that many business users will not tolerate. I've witnessed this firsthand: after a demo, the most frequent question was not about feature capability, but "Which button do I click first to just get my summary?"
Let me illustrate with a specific example. The process to review and edit a transcript before generating the final AI notes is non-intuitive.
* The transcript opens in a dedicated view, but editing controls are minimal and the "Save" action isn't persistently visible.
* To then generate summaries, one must navigate away from the transcript to a different "Notes" section, creating a sense of discontinuity.
* The settings for controlling what the AI focuses on (e.g., "Key Points," "Decisions," "Action Items") are buried in a team settings page, not accessible contextually during note generation.
This is reminiscent of poorly designed cloud service consoles where critical configuration is disconnected from the resource creation flow. For a non-tech user, each of these steps is a potential point of confusion or abandonment.
Furthermore, the use of terminology can be alienating. Terms like "Semantic Search," "Speaker Diarization," and "Topic Detection" are meaningful to us in tech, but for general business users, these labels are opaque. A label like "Find conversations about [Project X]" would be far more intuitive than "Semantic Search." The UI should speak the user's language, not the engineer's.
From a cost optimization and adoption perspective, this is a critical flaw. A platform's value is only realized if it is used consistently. If onboarding requires a detailed instruction document or repeated training sessions—as I found necessary—the total cost of ownership skyrockets, and ROI diminishes. In our case, we estimated a need for 3-4 hours of dedicated training per user to achieve confident, independent use, which was a deal-breaker compared to more streamlined competitors.
I am curious if others have conducted similar internal usability assessments and what workarounds, if any, were implemented. Did you create internal documentation, or did you find that certain UI elements could be hidden or simplified through configuration?
Totally agree about the "which button first?" problem. That's a classic symptom of designing for features instead of user goals. I tried to onboard a colleague who just needed meeting summaries emailed, and they got lost immediately after login because the default view assumes you want to analyze something.
You mentioned the leap from Otter.ai - that's a huge factor. Simpler tools set an expectation for a single-purpose UI. When a platform like Sembly adds layers for tasks and insights, it needs a much stronger onboarding funnel or a "simple mode" toggle. Have you seen any platforms handle that transition well?