Theta Lake has been granted United States patent: US 2024/0364759 A1, Systems and Methods for Analyzing Content from Virtual Whiteboards, covering a framework that captures the full, evolving life of a virtual whiteboard — every object, every author, every change, at every moment — and turns it into structured, searchable data that Theta Lake’s AI can analyze for risk.
Connecting Innovation Across the Patent Portfolio
This patent intersects with and extends a broader portfolio of over 18 patents that together define Theta Lake’s approach to oversight of digital communications. From our foundational patent covering context-based policy detection across what is spoken, shown, and shared, to our patents for seeing what’s shared on screen, and hearing what was really said, this patent extends that same principle to one of the most challenging surfaces in modern collaboration: the virtual whiteboard.
Theta Lake’s AI risk detections are multimodal–analyzing text, images, and context across all spoken, typed, visual, and shared content. But AI can only analyze what it can see. A whiteboard captured as a static image, or as hours of video, provides either insufficient or overwhelming data. This patent solves that challenge at its core, providing our AI classifiers with a structured, comprehensive, and accurate picture of everything that happened on the canvas.
The Whiteboard Problem: Infinite, Living, and Easily Erased
Virtual whiteboards have become a default space for teams to brainstorm and collaborate, adding screenshots, leaving sticky notes and routinely drawing, sharing, and commenting on information. They are built into communications platforms like Zoom, Webex, and Microsoft Teams, and offered as dedicated platforms like Miro and Mural. But they create a blind spot for compliance.
The compliance challenge arises because digital whiteboards are unlike any other communication medium:
- They are infinite, scrolling endlessly in every direction.
- They are persistent — a single board can remain active indefinitely continuously updated by multiple contributors.
- They are highly visual — much of the content is drawings, icons, charts, and images rather than text.
- And they are positional — where an object sits, and what it sits next to, can completely change its meaning.
- Compounding these factors is their ephemeral nature; users can add, modify, or delete content seamlessly, often long before a static snapshot is ever captured.
When compliance, privacy, security, or HR risks materialize on a whiteboard—such as sensitive customer data embedded in a sticky note, a confidential roadmap sketched into a diagram, or inappropriate visual content dropped onto the canvas—the potential risk exposure for an enterprise is significant. Because this content is ephemeral, it frequently evades standard snapshot detection—and attempting to manually audit this high volume of dynamic content is operationally unfeasible.
What the Patent Covers
The patented framework introduces a highly effective, data-driven alternative to legacy compliance methods. Traditional approaches rely on taking static image snapshots or recording hours of continuous video—methods that result in either inadequate context or excessive data volumes. This framework avoids photographing or filming the whiteboard entirely; instead, it ingests the underlying data directly through an API and reconstructs the canvas from its individual component parts.
That data includes:
- The whiteboard objects themselves: The text boxes, images, videos, icons, sticky notes, comments, and sketches that make up the canvas.
- The coordinate metadata for each object: The precise x,y location that tells the system not just what is on the board, but where, preserving the spatial context that gives whiteboard content its meaning.
- The activity log: A structured, timestamped record of who did what and when—capturing every creation, edit, and deletion by every participant.
With these lightweight inputs, the system can recreate the whiteboard on an object-by-object basis at any point in its entire timeline, without a single screenshot. It can rebuild exactly how the board looked at a specific date and time, account for how it got there, and even surface content—like a comment that was written and then deleted—that no snapshot would ever have caught.
Capturing Time, Not Just a Moment
The result is the difference between a static photograph and a complete, replayable whiteboard history. Because every object carries its location and every action carries a timestamp, the framework reconstructs the board as a living document. A reviewer can navigate to any specific moment, see the canvas exactly as it stood, and review the accompanying stream of activity that explains how it came to be. It is also fully searchable. A compliance team can quickly locate every instance when two specific individuals were active on a board, or trace the history of every board a given user has interacted with.
And because the system understands the board as a sequence of changes rather than a fixed image, it can generate a dynamic, condensed replay of the relevant segments. This allows a reviewer to play back, in a seconds, the exact portion of a lengthy whiteboard session where a risk actually occurred.
Powering Smarter AI Risk Detection Downstream
This structured, object-level data serves as the critical foundation for Theta Lake’s downstream AI risk detection. A comprehensively reconstructed, fully searchable record of the whiteboard enables our targeted AI classifiers to analyze the content for risks
- Optical Character Recognition (OCR) extracts text from images and from freehand writing made with a finger or stylus, turning drawings and handwritten notes into analyzable text.
- Image detection examines the non-text visual elements that make whiteboards so hard to supervise.
- These signals feed the same multimodal classifiers that analyze the rest of an organization’s communications for regulatory, compliance, privacy, cybersecurity, and HR risk.
Crucially, because the framework preserves the full history rather than a single frame, the AI can analyze content that would otherwise have disappeared. When it flags a risk, the dynamic replay takes a reviewer straight to the moment it happened, in context, with a complete account of who did what.
Seeing the Whole Whiteboard — at Scale
Managing modern collaboration risk requires an understanding of content, context, authorship, and modification over time across text, images, and visual elements. Historically, virtual whiteboards have been one of the most difficult environments to supervise—proving too vast for standard video recordings, too dynamic for static screenshots, and too visual for basic text filters.
The technology covered in this patent bridges this gap. By reconstructing the whiteboard as a structured, searchable, and replayable asset, it provides Theta Lake’s AI with a complete view of collaborative workspaces. Ultimately, this ensures that the ideas brainstormed on a digital canvas can be governed with the same rigor applied to everything else users say, show, and share.
