Visual Analysis
Visual Analysis
Section titled “Visual Analysis”When a guest shares their screen during a meeting, your AI avatar can see and analyze the content being presented. This allows the AI to respond to visual context — not just what the guest says, but what they’re showing.
How It Works
Section titled “How It Works”Visual analysis activates automatically when a guest starts a screen share. The AI periodically captures and analyzes the shared content, using it to inform the conversation. No setup or configuration is required.
What the AI Sees
Section titled “What the AI Sees”During a screen share, the AI can analyze:
- Presentation slides — key points, data, charts, and diagrams
- Documents — text content, tables, and formatting
- Dashboards — metrics, graphs, and KPIs being displayed
- Websites — pages the guest is browsing or demonstrating
- Spreadsheets — data tables and figures
The AI processes visual content in context with the ongoing conversation, combining what it sees with what the guest is saying.
How the AI Uses Visual Context
Section titled “How the AI Uses Visual Context”The AI references visual content naturally in conversation:
- Acknowledges what’s shown — “I can see the dashboard you’re sharing” or “Looking at those metrics…”
- Asks contextual questions — based on specific content visible on screen
- Connects visual and verbal — relates what’s being shown to topics already discussed
- Identifies relevant details — picks up on numbers, labels, and key information from shared content
Keeping Up With the Screen
Section titled “Keeping Up With the Screen”Taking a proper look at a screen takes the AI a few seconds, so it will often say something like “give me just a second to see your screen” first.
- The guest can see when it’s looking. While the AI is reviewing the screen, a “Looking at your screen…” indicator appears next to the avatar, so a normal pause is never mistaken for a frozen or broken meeting.
- If it’s taking a while, it says so. Should the review run longer than usual, the AI tells the guest it’s still looking rather than leaving them in silence. It won’t interrupt to do this — if the guest is mid-sentence, it waits.
- It speaks as soon as it has looked. Once the AI has taken in the screen, it describes what it sees straight away — the guest doesn’t have to prompt it a second time or fill the silence to get an answer.
- It won’t describe a page the guest has left. If the guest navigates somewhere new while the AI is still working out what it saw, the AI discards that answer and takes a fresh look instead of talking about the previous page. On a screen that hasn’t changed, it reuses what it already saw rather than re-reading it, which keeps the conversation quick.
There is still a short gap between the guest changing the screen and the AI having taken in the new one — a few seconds, much like a person glancing over. What it won’t do is describe the page they’ve already moved on from.
Common Use Cases
Section titled “Common Use Cases”Prospect Showing Their Current Setup
Section titled “Prospect Showing Their Current Setup”When a prospect shares their screen to show their existing tools or workflow, the AI understands the visual context and can ask informed follow-up questions about what it sees.
Requirements Documents
Section titled “Requirements Documents”If a guest shares a requirements document or RFP, the AI can reference specific sections and requirements during discussion.
Data and Metrics
Section titled “Data and Metrics”When guests share dashboards or reports, the AI can discuss the specific numbers and trends being displayed.
Privacy
Section titled “Privacy”- Visual analysis only processes content that the guest actively shares via screen share
- The AI does not access the guest’s camera feed for analysis — only screen-shared content
- Screen share content is processed in real-time and used for the current conversation
For more on the meeting experience, see Video Interface and Audio & Video Controls.