In late-August of 2025, we introduced AI Canvas, a native, no-code data-exploration tool that lets users visualize complex use cases that fall outside of standard reports. Admins can simply prompt in natural language and generate a custom widget in seconds, eliminating that 'data lag' where getting an answer takes hours of manual labor in third-party BI tools.
Today, AI Canvas gets smarter with the release of our Canvas Intent Engine, a new foundational capability that helps Canvas align on your intent before any work begins. This is an important step towards creating an AI experience designed for enterprise decision-making, where users can confidently leverage reliable, explainable insights to act quickly and effectively.
Our work reflects months of user feedback from network admins, security teams, and IT leaders across industries who have used AI Canvas in real environments.
Enterprise cybersecurity environments are inherently complex. AI Canvas spans a broad ecosystem of products, from network security and connectivity to identity and application performance. Each surface has its own terminology, data model, and domain-specific concepts.
As a result, even seemingly straightforward questions can have multiple valid interpretations. A request like "What are my riskiest sites?" could reasonably rank locations by malware activity, authentication anomalies, or degraded user experience. None of those interpretations are incorrect. They're simply different questions. Without understanding your intent, AI has no way of knowing which one you actually meant.
Most AI systems optimize for immediacy. Faced with ambiguity, they make an assumption and generate an answer. But in enterprise security, the cost of a wrong assumption isn't just an inaccurate chart. It can lead users down the wrong path, forcing them to retrace their steps or, worse, make decisions based on the wrong data.
The Canvas Intent Engine takes a more thoughtful approach.
Rather than immediately generating a visualization, Canvas first determines whether there's enough information to answer confidently. When multiple interpretations are possible, it asks targeted follow-up questions to clarify your intent before any work begins. Once your objective is clear, Canvas presents a transparent execution plan that outlines the logic and the data sources it plans to use.
Instead of asking users to trust an opaque process, the Intent Engine makes Canvas's reasoning visible from the very beginning. The result isn't more friction. It's fewer misunderstandings, fewer iterations, and a faster path to answers you can trust.
The Canvas Intent Engine is the first step toward a broader vision for AI Canvas. Understanding the user is only one side of the equation. To consistently produce accurate, trustworthy results, Canvas also needs a deep understanding of the enterprise data it's reasoning over.
That's why we're continuing to invest in what we call the Contextual Layer. This growing source of truth captures the relationships, semantics, and domain knowledge that exist across our product ecosystem, giving Canvas a richer understanding of the data behind every request.
That same foundation will also power explainability across Canvas. Rather than simply generating a visualization, Canvas will be able to clearly communicate how it arrived at its conclusions. For admins making operational decisions, understanding why an answer was generated is just as important as the answer itself.
Together, the Canvas Intent Engine and the Contextual Layer address both sides of AI reasoning. These investments continue to improve the accuracy and transparency of every Canvas output, enabling Canvas to integrate more deeply into everyday administrative workflows.
Canvas was built to help admins move quickly and confidently from data to decision, and every investment we make is aimed at bringing Canvas closer to that goal.
Stay updated on all things AI Canvas here: https://sasesensai.paloaltonetworks.com/new-ai-canvas/
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
| Subject | Likes |
|---|---|
| 5 Likes | |
| 2 Likes | |
| 1 Like | |
| 1 Like | |
| 1 Like |
