May 15, 2026
This is the sixth blog in Mike’s 8-part series about how AI can become a force for clarity and acceleration across the broader software development lifecycle.
To make more progress towards actually delivering software, we need to start breaking our insights down into the actual work that should be done, and what the user experience should be.
Claude has plenty of context to help us make those decisions, from architectural documents provided by the engineering team, to our discovery synthesis.
The process of deciding what to build is historically a hard one for product teams. A tight iteration loop involving human collaboration and review ensures we get to the right package of work to provide to the engineering team. Requirements captured during problem discovery are best documented as stories in separate markdown files within our file structure. If preferred, product requirements documents (PRDs) can also be generated for larger feature narratives using similar techniques.
Initial story templates reflected traditional product management best practices and dictated small stories with high levels of detail. In an AI-accelerated delivery environment, this proved counterproductive. Rich templates consume unnecessary context tokens when architectural rules and delivery conventions already exist elsewhere.
Agentic coding also changes assumptions about granularity. Larger-grained stories – three to five times the size of traditional stories – can be delivered quickly and safely when supported by shared context.
“Incremental development remains essential, but its value lies in risk reduction, not scope reduction. If larger stories can be delivered within the same elapsed time and assumptions are validated, the value-to-risk ratio increases.”
– Mike Mitchell, Product and Strategy Principal, Equal Experts
Multiple story templates support different delivery styles. The /req command prompts the product manager for preferences, surfaces open questions, and generates stories aligned with the chosen approach.
Even with well-structured stories, refinement remains a collaborative and iterative activity.
Product managers, engineers, stakeholders and users must share an understanding of why features should be built and which assumptions require validation. Story mapping helps identify gaps and prioritise delivery for larger features. AI can assist by generating visual maps to facilitate human collaboration. It can then incorporate changes back into the written stories.
Generative AI prototyping tools further accelerate validation by producing clickable demos. They reduce the cost of exploration and speed up feedback. Screenshots or raw HTML captured from a selected prototype can be stored alongside stories and indexed for LLM access, further informing its context.
While this work only included a cursory examination of AI techniques for story mapping and prototyping, this is an area rich for deeper future exploration.
As discovery spans multiple iterations, the LLM must understand what has already been captured. A story backlog file outside iteration directories provides this orientation.
The backlog indexes unbuilt stories with IDs, priorities, and source iterations. Stories are added automatically via /req, and the LLM consults the backlog to avoid duplication and surface dependencies.
This shared awareness improves accuracy and efficiency as discovery scales across teams and iterations, and transitions into delivery.
Read the full Evolving product discovery for AI delivery eBook
Mike Mitchell is a Product and Strategy Principal at Equal Experts, helping organisations adopt modern product practices and integrate AI into their product lifecycle in practical, sustainable ways.
His current focus is on helping product teams do more meaningful strategic work with AI, both by accelerating the routine work that crowds it out, and by making AI a more proactive, multi-modal collaborator.
Mike brings 35 years of experience in enterprise software, with a foundation of engineering followed by two decades of product leadership. He is a founder of multiple startups, has advised dozens more, and has led multiple enterprise transformations. Just prior to EE, Mike served as a fractional Chief Product Officer, guiding early- and growth-stage firms to deliver value faster through stronger product strategy and operating models. He holds a degree in Computer Science & Engineering from MIT.
Mike’s next blog in this series is about making sure AI builds the right things.
All business models have their pros and cons. But, when you consider the type of problems we help our clients to solve at Equal Experts, it’s worth thinking about the level of experience and the best consultancy approach to solve them.
If you’d like to find out more about working with us – get in touch. We’d love to hear from you.
