A technique (observed in other teams, not personally used by the source) where a near-final presentation deck is given to an LLM prompted to act as a virtual replica of a specific stakeholder. The LLM is asked to flag what's missing or could be misunderstood from that stakeholder's perspective — it is not used to draft or write the content itself.
The source explicitly declines to use this herself, on the grounds that an LLM 'can't replace knowing and understanding the people and the personalities in the room' — treating real-time reading of stakeholders as irreducibly human. This puts the technique in tension with Presenting as Discovery, which relies on live body-language cues and direct in-room questions rather than a pre-simulated proxy. It's also a sharper form of AI delegation than LLM as Internal Reporting Agent: that concept offloads admin, this one offloads a piece of interpersonal judgment.