Asana's AI agents share memory across your company — but not your secrets
Enterprise teams building AI agents keep hitting the same wall: a chatbot that can answer a prompt but can't remember what the last five people asked it, and can't tell you whether last month's version actually worked. In a fireside chat with VentureBeat's Sam Witteveen at VB Transform 2026 , Asana's chief product officer, Arnab Bose, unpacked how his team tackled this problem to build a new…
Enterprise teams building AI agents face a critical challenge: AI agents can answer prompts but lack memory of previous interactions, making it difficult to track progress and maintain continuity. At VB Transform 2026, Asana's chief product officer Arnab Bose discussed how his team solved this issue by developing a new operating system called Agentic Work Management (AWM).
AWM treats AI agents as coachable teammates that work alongside humans, rather than as one-to-one assistants. Bose explained that AWM engineers the system on top of Asana's 18-year-old architecture: the Work Graph. This graph-based database organizes information through the Pyramid of Clarity, a structure that breaks down work into tasks, projects, portfolios, and company-wide goals.
This architecture allows AWM to provide a multiplayer teammate that views overarching company goals, updates project statuses, and shares memory with human colleagues.
AWM already has production deployments with several customers, including FedEx. To build AWM, Asana had to tackle several technical hurdles. First, they addressed data governance concerns by implementing access controls to prevent unauthorized access to confidential information. Second, they developed dynamic model routing to abstract prompt engineering from the user, allowing the system to automatically route prompts to appropriate models based on task complexity.
Lastly, they designed a billing architecture that charges a static cost per task completion to ensure predictable enterprise pricing.
This new operating system addresses a specific problem with current enterprise AI deployments: statelessness. Basic chat-based agents lack persistence, making it difficult to create reusable workflows for future tasks. AWM solves this by creating a permanent state, recording metadata on task completions and registering how they impact project statuses and company goals. CoreWeave is an early adopter of AWM, using it to streamline complex new product launches.
Written by urgent.news from VentureBeat's reporting — not their text. Machine-written — it may contain errors, so check the original before relying on it.
