Frame AI independence like self-driving car levels: 'Human-in-the-loop' (AI as advisor), 'Human-on-the-loop' (AI acts with supervision), and 'Human-out-of-the-loop' (full autonomy). This tiered model allows organizations to match the level of AI independence to the specific risk of the task.

Enterprise buyers are drawn to the vision of full automation ("the sizzle"), but their immediate need is improving existing human workflows ("the steak"). A startup must offer both. The visionary product gets them in the door, while the practical agent-assist tool delivers immediate value and gathers necessary data for future automation.

AI model capabilities have outpaced their value delivery due to a fundamental design problem. Users are inherently scared and distrustful of autonomous agents. The key challenge is creating interaction patterns that build trust by providing the right level of oversight and feedback without being annoying—a problem of design, not technology.

To drive AI adoption, leaders must balance two opposing actions. They must 'do more' by setting a high bar for creating 'magical' customer experiences. At the same time, they must 'do less' by empowering teams with autonomy, reducing review overhead, and giving them freedom to experiment.

Individual employees want powerful, autonomous AI agents similar to consumer products. However, the enterprise prioritizes control, safety, and governance. This creates a fundamental tension that enterprise AI products must navigate, balancing user desire for freedom with the organization's need for security and oversight.

A common AI implementation failure is assuming users think like technologists. Trivial technical details can be huge adoption blockers. To succeed, focus on building user trust and actively partner with customers to operationalize the technology, rather than simply delivering it and expecting them to figure it out.

Instead of focusing on AI features, understand the two mental shifts it creates for customers. It either offers a superior method for an existing, tedious task ("a better way") or it makes a previously unattainable goal achievable ("now possible"). Your product must align with one of these two thoughts.

Unlike traditional software, AI products have unpredictable user inputs and LLM outputs (non-determinism). They also require balancing AI autonomy (agency) with user oversight (control). These two factors fundamentally change the product development process, requiring new approaches to design and risk management.

Design systems that can be operated by humans, AI agents, or a combination. This prevents projects from failing due to over-automation or requiring a complete refactor when human intervention is needed, ensuring flexibility and saving future development costs.