Can AI supercharge the Product Design Sprint? - Part 1: Discovery

Introduction

The need for rapid and effective product development has never been greater. The design sprint, popularized by Google Ventures and detailed in Jake Knapp’s Sprint, has become a go-to framework for solving complex problems and validating ideas in just five days. But what if this process could be enhanced even further? Enter AI—a tool that, when thoughtfully integrated into the design sprint, can amplify creativity, streamline decision-making, and drive more impactful outcomes.

In this post, I’ll explore whether AI can supercharge a product design sprint, using a sample product idea: a marketplace for crowdfunding events. We’ll walk through various sprint exercises, to see if AI, specifically ChatGPT 4o can enhance the process from discovery to validation.

For more insights into how thoughtbot approaches the different phases and exercises of a design sprint, you can also check out our Design Sprint Guide.

This is the first installment in the series, where I’ll begin by covering the Discovery Phase.

Roles in this Example Sprint

A traditional Design Sprint typically involves a diverse team, each member bringing unique expertise to the table. Roles such as facilitator, decider, designer, and subject matter expert work together to drive the sprint forward.

In this sprint, AI was introduced as an additional participant, capable of filling various roles as needed. This makes it a versatile contributor, providing insights and accelerating tasks. This integration allows for exploring more possibilities and making faster, informed decisions.

Crafting the Problem Statement

In this example, I’m focused on creating a marketplace for crowdfunding events. To ensure that the design sprint has clear direction, it’s essential for me to craft a well-defined problem statement. A strong problem statement acts as a guiding compass, helping align my efforts and drive the sprint toward effective solutions.

I typically break down the process of creating the problem statement into three steps:

1. List User Problems: I begin by identifying the specific challenges faced by both event organizers and attendees. For instance, organizers might struggle with securing funds and gauging interest, while attendees may want more influence over which events come to their area. At this stage, I would normally have every sprint participant write down the problems they can identify, adding them to a shared list.
2. Focus on the Critical Problem: Next, I prioritize the most pressing issues that the product needs to solve. This might involve deciding which problem is most important or combining similar problems to focus on a core issue, such as the difficulty of funding events.
3. Refine the Problem Statement: Finally, I collaborate to ensure that the problem statement is clear, concise, and closely aligned with user needs. After identifying and grouping the problems, I remove duplicates and refine the statement to focus on the most critical challenge.

While this process generally works well, it can encounter a few challenges, such as:

• Bias and Subjectivity: Participants may bring their own biases, leading to a problem statement that reflects individual perspectives rather than a collective understanding.
• Information Overload: Manually sorting through large amounts of data and feedback to identify problems, can be time-consuming and overwhelming, making it difficult to focus on the most critical issues.
• Time Constraints: Crafting a problem statement manually can be time-intensive, potentially delaying the start of the design sprint. When individually listing out problems, there may be a chunk of time dedicated to identifying duplicates to remove.
• Limited Knowledge: Participants who are not key stakeholders might have minimal understanding of the problem space. Despite access to provided materials, they may still overlook significant aspects due to their limited exposure to the idea and the short amount of time available to grasp its nuance.

However, with the power of AI in this example, I can speed up this process tremendously while also gaining fresh new perspectives and insights.

To start, I make sure to give ChatGPT enough context about the idea.

Note: In this example, I didn’t have any experts involved, but in a classic design sprint, talking with them is a key exercise. Experts who are close to the problem would typically be brought in to help fill knowledge gaps. In this sprint, I’ll feed ChatGPT limited information to begin with and see what output it can generate.

For this example, I simply told it:

“I am trying to create a marketplace for people to crowdfund events in their area from concerts, live events, and more.”

Second I asked it:

“Provide me a list of problems that both event organizers and attendees face. These problems will be used to help create the problem statement.”

The result was:

Event Organizers:

• Securing Funding: Difficulty in raising sufficient funds to cover event costs.
• Gauging Interest: Challenges in accurately assessing potential attendee interest before committing resources.
• Marketing and Promotion: Struggles with effectively reaching the target audience and promoting the event.
• Logistics and Planning: Managing the logistics, such as venue booking, equipment rental, and staff coordination.
• Risk of Low Turnout: Fear of low attendance, which can lead to financial loss and reputational damage.

Attendees:

• Lack of Influence: Frustration with having little to no say in which events are brought to their local area.
• Discovering Relevant Events: Difficulty in finding events that align with their interests or are happening nearby.
• Event Uncertainty: Concern that events they are interested in might be canceled due to insufficient funding or interest.
• Ticket Costs: Issues with high ticket prices, especially if they are not confident the event will deliver value.
• Limited Community Engagement: Desire for more opportunities to connect with others who share similar interests before or during events.

After it provided me that list, I ported them over into FigJam to organize it into cards seen below:

After reviewing all the problems identified by both the AI and myself, I chose what I believe are the most critical issues faced by each user group.

For Event Organizers, I identified Securing Funding as the top challenge. Without sufficient funding, organizers struggle to realize their event ideas, leading to potential financial losses and unfulfilled objectives.

For Attendees, the most pressing issue is their Lack of Influence over local events. This often leaves them frustrated with the limited options available that match their interests.

From these insights, I used ChatGPT to craft problem statements that clearly define these challenges.

“Please create a problem statement based on the issues we’ve identified: one focused on securing funding for the event organizer and another addressing the lack of influence for the attendee.”

This is what it gave me:

For Event Organizers: “Event organizers struggle to secure the funding needed to plan and execute events, risking financial loss and potential cancellations.”

For Attendees: “Attendees are frustrated by their lack of influence over local events, leading to a disconnect between what they want and what’s available.”

Feeling satisfied with the problem statements provided by ChatGPT, I felt confident in moving forward with the rest of the design sprint, equipped with a clear and solid direction to guide us.

Creating the Critical Path and How Might We Questions with AI

With our problem statements in place, we were ready to create the critical paths for both user groups: Event Organizer and Attendee. The critical path outlines the key steps each user takes, from recognizing their problem to finding a solution, ensuring we cover every important interaction along the way.

I decided to experiment with how ChatGPT might develop detailed critical paths for both user groups. My goal was to ensure that each step aligned with the problem statements and addressed the users’ most critical needs. In addition to the critical path, I asked it to provide some How Might We (HMW) questions to each step so that it can be a useful point of reference when ideating later on.

I used this prompt to generate the critical path for Event Organizers and Attendees.

Create two tables with the critical path for both the event organizer and the attendee. The first column should list each step (e.g., Problem Recognition, Explore Solutions). The second column should provide details of the step (e.g., “The attendee looks for ways to influence local event offerings or find events that align with their interests.”). The third column should include at least two HMW (How Might We) questions for each step.

Critical Path for Event Organizers:

Note: I’ve included a sample of the raw output from ChatGPT (below) and the chart I created in Figjam (above) based off that output.

Problem Recognition:

• The organizer has an event idea but lacks the necessary funding or is seeking new event ideas to fulfill.

HMW Questions:

HMW help organizers quickly identify viable event ideas or funding needs?

HMW simplify the recognition of funding challenges?

---

Critical Path for Attendees:

Note: I’ve included a sample of the raw output from ChatGPT (below) and the chart I created in Figjam (above) based off that output.

Problem Recognition:

• The attendee wants more control over local events or desires certain events that aren’t available.

HMW Questions:

HMW enable attendees to express event preferences effectively?

HMW allow attendees to influence event offerings easily?

I was impressed with the results ChatGPT provided, especially considering how quickly it generated two distinct flows for the Event Organizer and Attendee. While the details weren’t exhaustive, it offered enough structure to outline the end-to-end flow, which will be invaluable during the ideation phase. The HMW questions it generated at each step were particularly insightful, prompting deeper consideration of potential solutions. With the Critical Path established, I concluded the first phase and moved on to the next crucial step: the Assumptions Board.

In the next part of this series, I’ll explore the Assumptions Board and Storyboarding, where we’ll dive deeper into validating ideas and visualizing the product journey.