AI
The strategist is the author. The AI is a drafter.
Every AI function in Steerpoint receives only the engagement objects relevant to its specific job. Structured context in, specific output out. The engagement data is the constraint.
Bounded context, not open prompting
AI in Steerpoint does not operate on free-form input. Each function receives a specific package of engagement objects: the relevant domains, their maturity states, linked value cases, active decisions, and connected deliverables. The precision of the output depends on the structure of the input.
Named roles with visible inputs
Each AI function operates through a named role with a defined job. The strategist can see exactly what the role was given and what it produced. There is no black box. The context assembly is visible, the output is reviewable, and the decision to accept or dismiss stays with the strategist.
Review model, not auto-apply
AI output enters the engagement through a review step. Nothing is applied automatically. The strategist reviews what was generated, sees the evidence it was based on, and decides what to accept, modify, or dismiss. The workflow preserves human judgment at every point.
Strategic briefing with staleness awareness
The workspace-level strategic briefing synthesises situation, key insights, next steps, and risk signals from live engagement data. It tracks whether the workspace has changed since the last generation and signals when a refresh is needed. The briefing stays current because the engagement data stays current.
Object chain
AI functions operate against the full engagement object graph. Domains, Domain States, Gaps, Value Cases, Decisions, Deliverables, and Artifacts all serve as structured context for AI generation. The quality of AI output is determined by the quality of the engagement evidence.