AttentionTag doesn't just measure engagement — it acts on it. Two complementary systems work together: Assistants intervene on the viewer's side, while Coaching Cards guide the speaker in real-time.
Automated interventions triggered on the viewer's device when engagement thresholds are breached. Configured by the speaker or admin.
AI-generated coaching suggestions delivered to the speaker in their meeting sidebar, helping them adapt their delivery in real-time.
Rule-based workflows that act on engagement signals
Assistants are workflows you configure in the speaker console. Each assistant has three components:
The engagement threshold that triggers the assistant — for example, "valence drops below 30%" or "3+ viewers sleepy for 2 minutes."
The type of intervention: Comprehension Support, Attention Reinforcement, Engagement Encouragement, or Stress Mitigation.
What happens when the condition is met — a visual prompt on the viewer's screen, an auditory cue, or a chat message.
Executes in real-time on the viewer's device during the session. When the condition is met, the action fires immediately — a visual prompt appears on their screen, or an auditory cue plays. No server round-trip needed.
Evaluated periodically on the server. Ideal for batch actions like sending summary emails, triggering webhook notifications, or SMS alerts when engagement patterns are detected over a session.
An on-screen message or overlay on the viewer's device
A sound alert to recapture attention
In-app chat message to the viewer
Condition
Viewer's focus drops below 40% for more than 3 minutes
Category
Attention Reinforcement
Action
Visual prompt: "Quick check — are you still with us?"
AI-generated, real-time guidance for the speaker
During a live session, AttentionTag continuously monitors aggregate engagement across all viewers. When it detects a meaningful dip — such as several students losing focus or showing signs of confusion — it automatically generates a coaching card and delivers it to the speaker's meeting sidebar.
Monitor
Engagement signals checked continuously
Detect
Threshold breach identified (individual or aggregate)
Generate
AI creates contextual coaching suggestion
Deliver
Card appears in speaker's meeting sidebar
4 students drifting for 3+ minutes
Suggested Actions
Interactive mockup — actual cards appear in the Zoom or Teams sidebar
Each card is tagged with an urgency level so you can prioritise:
The AI considers the full meeting context when generating suggestions — including engagement trends, which specific viewers are struggling, how far into the session you are, and even the meeting transcript. This means suggestions are specific and actionable, not generic advice.
The AI also remembers previous cards in the same session, so it won't repeat suggestions you've already seen or acted on.
Check off the suggestions you followed, then mark the card as Done. This helps the AI learn what works for your teaching style.
If the suggestion isn't relevant right now, dismiss it. The card is logged but won't distract you further.
Cards stay visible until you act — you're never forced to respond. Focus on teaching, address them when you have a natural pause.
| Assistants | Coaching Cards | |
|---|---|---|
| Target | Viewers / Students | Speaker / Teacher |
| Triggered by | Rules you configure | AI detects engagement dips automatically |
| When | Real-time (live) or scheduled (daily) | Real-time, during live sessions only |
| Content | Predefined actions (prompts, sounds, chat) | AI-generated teaching suggestions |
| Setup | Manual configuration in console | Fully automatic — no setup required |
| Delivered via | Viewer's app (visual, audio, chat) | Speaker's meeting sidebar (Zoom, Teams) |
Assistants and Coaching Cards are driven by the engagement signals AttentionTag captures in real-time. Learn about every signal we track and what it means.
