Gocious: how senior product leaders should govern AI across a portfolio in 2026

Kevin Jankay published this guide for senior product leaders on April 13, 2026. Where most AI-in-PM writing focuses on individual productivity, this piece addresses organisational strategy: how product leaders should govern AI initiatives across a portfolio rather than managing them as isolated experiments.

The core argument

The guide argues that AI initiatives should compete within the same portfolio prioritisation frameworks as other product investments, not in a separate track with its own budget and reporting. Treating AI as a special category tends to produce projects that are technically interesting but disconnected from measurable business outcomes. According to Deloitte’s 2026 State of AI in the Enterprise study cited in the article, 66% of organisations now report tangible gains from AI adoption — the distinction between those that do and those that don’t is structural alignment, not the quality of the AI tools used.

The portfolio-first framework

The article introduces an evaluation matrix that assesses AI initiatives across multiple strategic dimensions: ROI potential, risk level, scalability, and competitive differentiation. Running initiatives through this matrix before funding decisions prevents the common pattern of approving AI projects based on technical novelty rather than strategic fit.

Roadmapping receives specific attention. Traditional static roadmaps cannot accommodate the pace at which AI capabilities change. The guide recommends adaptive roadmap tools that explicitly model hardware-software dependencies and support scenario planning for regulatory and cost fluctuations. The practical implication is that product leaders need both short-term delivery plans and contingency scenarios for the cases where a key model provider changes pricing or a regulatory requirement shifts.

Governance and risk management

For high-risk AI deployments, the article recommends establishing lightweight review boards rather than relying on informal approval processes. Security and compliance requirements should enter product requirements at the beginning of an initiative, not as a late-stage review. Jankay frames this as continuous feature development — AI safety work is ongoing, not a gate to pass once at launch.

The article also maps AI accountability across leadership levels, from CPO to individual PMs, specifying who owns model selection, data governance, feedback loops, and user-facing AI behaviour. This role distribution prevents the situation where everyone assumes someone else is responsible for AI quality.

Who it is most useful for

Product leaders at VP or director level who are moving past individual AI tool adoption and need a structured approach to managing AI across multiple product lines. The frameworks described are practical enough to apply incrementally, and the governance structures are scaled to organisations that do not yet have a dedicated AI governance function.