Four lanes. One question: where is the system carrying more than it can show?

When the dashboard says green but the work still feels stuck, the problem is usually not effort. It is drift. Roadmaps look like prioritization documents, but the prioritization is doing structural work - each strategic decision creates load that gets absorbed somewhere, by a team, a partner, a customer segment, a downstream system. When the absorption is invisible, the strategy looks healthy until it does not.

The Product Strategy practice covers two kinds of engagement, scoped to the actual need:

• Senior PM advisory and fractional product leadership. Roadmap design, prioritization frameworks, OKR setup, market positioning, board-readiness preparation, and product-org coaching. Recurring monthly engagements where the work runs alongside your in-house team, or fixed-scope projects that ship a specific deliverable.
• The Direction (Strategy Read & Direction Map). A two-to-four-week structural read on a planned strategic move or current plan-of-record. Maps where prioritization is quietly exporting fragility across team, partner, or downstream boundaries; names the decisions that should change before the next move loads something the current pace was not built to carry.

Engagement shape locks at the start of the conversation. Discovery is a 50-minute structured call; if a fit is not there, we say so directly.

Many organizations no longer have an AI capability problem. They have an operating-model problem. The tools are in the building, but the operating model around them - review, escalation, handoff, evidence, trust rituals - has to keep up by design, not by reaction. AI implementations begin as a tool decision and end as a load decision: the model and the integration get the attention; the operating model around them ends up being whatever the team improvised. The practice treats AI as disciplined operational sensemaking, not automation theater.

The Applied Intelligence practice covers three named engagement shapes - The Read scoped to a single initiative, an Implementation Stand-up before AI is stood up, and a Governance Install for AI that is already running:

• The Read (deployment scope). The same Read shape used at full operational scope in Risk Management, applied here to one named AI or automation initiative - an internal system being built, a third-party AI-powered tool being installed, or an applied-AI program where the operating measurement is unclear. The read names the structural load the deployment is creating or amplifying, the review and escalation gaps it is exposing, and the next decision worth making. Output is multi-form: a written read for argument and sharing, a working decision session with the team where the findings get tested live, and named tripwires for what to monitor afterward. Compressed timeline - weeks, not months; final scope agreed during discovery.
• AI Implementation Stand-up. A forty-five-to-ninety-day operating- model design pass for AI surfaces being stood up - before the system has produced its first real output, while the surface is still flexible. Two engagement forms: third-party AI / AI-powered tool implementation, and internal agentic-system implementation. Advisory cadence (eight-to-twelve hours per week of focused work, named touchpoints), not embedded participation.
• Agent Governance Install. A sixty-to-ninety-day done-with-you operating-discipline package for AI-enabled work that is already running. Two engagement forms: third-party AI use governance (vendor tools, copilots, AI-powered SaaS embedded in workflows) and internal agentic-system governance (in-house copilots, custom agents, machine-decisioning systems). Five disciplines: review, escalation, handoff, evidence, trust rituals.

The work is operating-model design and discipline, not vendor selection or developer implementation. If the question is which tool to buy or how to build the agent, that is a different conversation.