Current LLMs are intelligent enough for many tasks but fail because they lack access to complete context—emails, Slack messages, past data. The next step is building products that ingest this real-world context, making it available for the model to act upon.

Don't treat evals as a mere checklist. Instead, use them as a creative tool to discover opportunities. A well-designed eval can reveal that a product is underperforming for a specific user segment, pointing directly to areas for high-impact improvement that a simple "vibe check" would miss.

An AI model can meet all technical criteria (correctness, relevance) yet produce outputs that are tonally inappropriate or off-brand. Ex-Alexa PM Polly Allen shared how a factually correct answer about COVID was insensitive, proving product leaders must inject human judgment into AI evaluation.

The review of Gemini highlights a critical lesson: a powerful AI model can be completely undermined by a poor user experience. Despite Gemini 3's speed and intelligence, the app's bugs, poor voice transcription, and disconnection issues create significant friction. In consumer AI, flawless product execution is just as important as the underlying technology.

Don't ask an LLM to perform initial error analysis; it lacks the product context to spot subtle failures. Instead, have a human expert write detailed, freeform notes ("open codes"). Then, leverage an LLM's strength in synthesis to automatically categorize those hundreds of human-written notes into actionable failure themes ("axial codes").

AI tools can handle administrative and analytical tasks for product managers, like summarizing notes or drafting stories. However, they lack the essential human elements of empathy, nuanced judgment, and creativity required to truly understand user problems and make difficult trade-off decisions.

When asked to describe a user process, an LLM provides the textbook version. It misses the real-world chaos—forgotten tasks, interruptions, and workarounds. These messy details, which only emerge from talking to real people, are where the most valuable product opportunities are found.

The common mistake in building AI evals is jumping straight to writing automated tests. The correct first step is a manual process called "error analysis" or "open coding," where a product expert reviews real user interaction logs to understand what's actually going wrong. This grounds your entire evaluation process in reality.

Assigning error analysis to engineers or external teams is a huge pitfall. The process of reviewing traces and identifying failures is where product taste, domain expertise, and unique user understanding are embedded into the AI. It is a core product management function, not a technical task to be delegated.