Conference Presentation, Fireside Chat, Interview
The Missing Layer: Why Agentic AI Fails Without a Work Operating System | Asana | RAISE 2026
- Current productivity gains are largely isolated to individual levels rather than translating to broad business productivity, as platforms for real business process AI are only beginning to emerge and face challenges in connecting to systems while ensuring governance and multiplayer capabilities.
- With an existing base of 185,000 customers and 80% of the Fortune 100, the company plans to provide a harness to identify business processes, extending its collaboration substrate to enable humans and agents to work together.
- Future enterprise architecture will not rely on a single platform but will feature agents operating across multiple elements, necessitating systems that aggregate identity, audit trails, delegation, attribution, debugging, discovery, and cost control.
- Companies must treat intellectual property as evolving from data to traces, evaluations, and prompts, while managing a diverse model strategy that balances open source, frontier, and other options based on geopolitical factors and company structure.
- Output quality is determined by the harness and data cleanliness rather than the model alone, requiring firms to develop "muscle memory" to assign tasks to appropriate models instead of using expensive frontier models for every task.
- To ensure quality in business-critical processes, agents must break work into stages and intersperse execution with human checkpoints or deterministic verification, a standard that enterprise buyers will demand for identity verification and approval workflows.
- Engineering strategies will focus on maximizing AI usage without causing burnout or compromising craft quality, ensuring teams are not burdened with low-quality content while maintaining rigorous test coverage and migration of legacy systems.
- The company intends to build its own vector between models or utilize automatic systems to avoid passing all margins to LLM providers, ensuring that even as token prices decrease, firms retain control over their infrastructure.
- Internal operations are accelerating regarding feature development and code migration, supported by a stance that legacy systems can be advantageous if utilized effectively to serve customers in the market.