
Omnifact is now built around agents for autonomous, multi-step workflows.
Published on August 28th, 2026
Today, Omnifact is receiving its biggest update in over two years. We have rebuilt the core Omnifact Chat experience from the ground up around agents to provide a faster, more useful foundation for your daily work with AI.
With today's release, Chat in Omnifact is no longer just a simple conversation with an AI assistant for occasional searches or quick lookups. We have fully rebuilt the experience around agents capable of taking action on your behalf, thinking through complex problems, and executing multi-step workflows that can run over extended periods. Whether you need to conduct deep research, synthesize information from different sources into a grounded, fact-based report, or automate repetitive tasks, these agents represent a true step change in the productivity of work.
The first change you will notice is live execution feedback instead of a static loading spinner, giving you full visibility into how Omnifact is thinking and planning before it acts. The new Chat interface displays each step as it happens, whether the agent is searching the web, analyzing a document from your knowledge base, or running a complex multi-step task.

You can expand any step to inspect the exact queries run or documents read. Once execution completes, each step collapses into a concise summary so your conversation history remains clean and scannable.
Because agents can now run in the background and in parallel, you don't need to watch them, allowing you to be more productive, waste less time, and get results while focusing on other important work.
We have refreshed our interface to place the Chat Input at the center of your workflow. Model selection has moved out of the top navigation bar and directly into the Chat Input, where you can switch models without looking away from your prompt.

We also cleared visual clutter from the Chat sidebar, making it much easier to browse and organize previous conversations. The interface is cleaner and more focused, while keeping all the core workflows you rely on.
In our previous interface, you chose a Response Mode such as Auto, Quick Response, or Simple Lookup to control how much the assistant analyzed a prompt before answering.
Omnifact now determines on its own whether a prompt requires a rapid direct answer or deeper multi-step reasoning and tool execution. For most teams, this removes an extra setting to manage.
Giving agents the ability to plan steps and use tools does not change our commitment to data protection. Everything your team relies on, from Space Knowledge and Web Search to Image Generation, operates under the exact same strict privacy and security standards.
Every prompt, tool input, and document retrieved during an agent workflow continues to run through our Privacy Filter, ensuring your organization gets the full power of autonomous agents without compromising data control.
Underneath the refreshed interface sits an entirely new architecture. We engineered this foundation to support advanced agent workflows, persistent execution environments, and complex multi-tool integrations that the previous chat system could not accommodate.
This lays the groundwork for what is coming next: connecting assistants directly to external systems and business tools via MCP, creating and editing complete business documents directly in the platform, and running background automations for repeatable knowledge work. Today's release is just the beginning, and we cannot wait to put these next capabilities into your hands.
The new agent-driven Chat experience is live today across all organizations. When you log in, a short onboarding walkthrough introduces the key interface updates, with detailed documentation available in the coming days.
We would love to hear how your team puts the new agents to work. Please share your thoughts through your dedicated account contact or directly via the in-app feedback tool as you explore the update.