2 Discovering and prioritizing AI opportunities

This chapter explains how to find, evaluate, and shape AI opportunities that solve real problems and fit a company’s strategy, capabilities, and risk tolerance. It frames discovery in an “opportunity space,” urging teams to resist building “AI for AI’s sake” and to focus on measurable customer and business value. The text maps common AI value themes—automation and productivity, human-AI improvement and augmentation, personalization, inspiration and innovation, convenience, and emotional engagement—while highlighting pitfalls such as automating rare, one-off decisions or pursuing use cases that require full explainability better served by simpler methods. Throughout, it emphasizes balancing quick wins with longer-term, defensible investments.

To source strong ideas, the chapter recommends combining existing knowledge and intuition with internal crowdsourcing, dogfooding, and customer listening validated against real behavior, while also scanning broader signals like technological shifts, competitor moves, regulation, and positioning. It distinguishes horizontal opportunities (broad, cross-industry) from vertical ones (domain-specific) and notes the different demands each places on data, domain expertise, and model adaptation. Three integration scenarios are outlined: AI as an add-on to an existing product (fast path but data/UX integration challenges), AI as the core value driver in greenfield products (requires solving cold-start and continuous model improvement while avoiding “solution in search of a problem”), and AI as an internal enabler that improves the product behind the scenes (guided by clear KPIs, governance, and strong internal UX).

For prioritization, the chapter proposes criteria such as user and business impact, technical feasibility, data availability, regulatory ease, scalability, and customer fit, illustrated via a music app where better discovery drives retention and is feasible with existing behavioral data and standard recommendation techniques. It cautions against overreliance on scoring matrices; instead, teams should surface uncertainties, make transparent trade-offs, and be ready to pivot. Finally, it contrasts two ways to shape opportunities: a careful “ready, aim, fire” approach suited to high-stakes or regulated contexts, and a fast “ready, fire, aim” approach that leans on rapid prototyping, real-world feedback, and iterative model improvement. Choosing the right mix helps teams deliver near-term wins while building a durable AI advantage.

This chapter addresses discovery in the “opportunity space” in the mental model of an AI system

The process for discovering AI opportunities

AI opportunity tree for a music streaming app

The AI cost equation for automation opportunities; beyond the development and operation cost, you should also estimate the cost of finding and fixing mistakes and the risk of uncaught mistakes.

Humans and AI each have different strengths which should be leveraged when building AI-driven products.

Three scenarios for AI integration

You decide to explore the personalization branch in your opportunity tree

This simple prioritization matrix evaluates AI opportunities according to user impact, business value, and technical feasibility. It is clear but potentially oversimplifies the decision.

When to be careful vs. fast: comparing “ready, aim, fire” and “ready, fire, aim”

The traditional design thinking process

• Cultivate idea flow: Create and maintain a constant flow of AI ideas by sourcing from multiple channels—user feedback, competitors, team insights, and technological developments. The more ideas you evaluate, the more likely you are to discover valuable opportunities.
• Involve your team: Engage the entire team in the idea generation and prioritization process. Use their intuition, creativity, and domain expertise for faster ideation and more robust decision-making.
• Quickly discern the “AI advantage”: Prioritize AI opportunities that solve specific user problems such as automating routine tasks, improving outcomes, innovating, or personalizing user experiences. Avoid "one-off" decisions and situations requiring full explainability or perfect accuracy.
• Use diverse sources: Go beyond customer feedback to identify AI opportunities. Draw inspiration from competitors, emerging technology, regulations, and market positioning to stay innovative and competitive.
• Distinguish horizontal vs. vertical use cases: Understand whether your AI opportunity applies to industries (horizontal) or a specific domain (vertical). Horizontal opportunities need broad market understanding, while vertical ones require deep domain expertise.
• Apply a prioritization framework: Use a structured framework to score AI opportunities based on user impact, business value, and technical feasibility. This approach ensures an objective evaluation process and helps align the team.
• Balance quick wins vs. long-term investments: Prioritize quick wins for immediate user impact while investing in long-term AI opportunities to create a competitive edge with AI.
• Choose the right development approach: Select between a careful, research-first approach ("ready, aim, fire" / design thinking) for high-stakes projects and a fast, iterative approach ("ready, fire, aim" / lean startup) to speed up time-to-value. Also consider your team's culture and risk attitude when making this choice.
• Validate and iterate: Continuously test your AI ideas through experimentation. Whether you're following a fast or careful approach, gathering user feedback early and iterating ensures better alignment with real-world needs.

FAQ