chapter four

4 Understand what users and stakeholders really want

 

This chapter covers

  • Selecting the right interviewees and research methods
  • Building a deep contextual understanding of key users and stakeholders
  • Leveraging analogies to uncover insights from similar situations
  • Crafting a clear and actionable AI problem statement

Now that we have aligned on the Idea Statement, and framed the opportunity for the enterprise, we must focus on user and stakeholder needs to develop a deep understanding of what truly matters to those who will use and influence the AI solution we are planning. This stage requires careful planning and skillful execution: who to interview, which investigation methods to use, and by whom—so we can do it right the first time and gain a clear, complete understanding of what matters for stakeholders and for the solution itself.

When the opportunity is highly innovative, analogies help us not only learn from other sectors with more experience in applied AI, but also better connect with interviewees and generate new ideas.

All of these elements come together in a structured AI Problem Statement, crafted and validated with the project team. This statement becomes the foundation upon which successful AI solutions are built—and the filter that ensures you are solving the right problem before investing in the wrong one.

4.1 Fighting fraud in energy distribution

4.1.1 A serious challenge in energy distribution

4.1.2 Investigating from scratch

4.2 The challenge of understanding people’s needs

4.2.1 Identifying key users and stakeholders

4.2.2 Understanding context

4.2.3 Uncovering hidden needs

4.2.4 Dealing with unchartered territories

4.3 Preparing for a successful investigation

4.3.1 Selecting the right sample

4.3.2 Selecting the right investigation method

4.3.3 Who should be the interviewer?

4.4 Capturing and documenting user and stakeholder needs

4.4.1 Understanding context

4.4.2 Reviewing the internal user’s workflow in detail

4.4.3 Building an evidence-based empathy map to communicate insights

4.4.4 Don’t reinvent the wheel: Use analogy

4.5 Practical tools for a successful investigation

4.5.1 Interview plan

4.5.2 AI prompt to speed up the identification of analogies

4.6 How we investigated the NTL problem at Gridvia

4.6.1 Interview planning

4.6.2 Understanding NTL

4.6.3 Learning from fraud control and crime prevention

4.6.4 Learning from the field

4.7 Crafting an AI problem statement that delivers

4.7.1 What makes a good AI problem statement