Design thinking
In product work, it is a discovery technique and a cross-cutting capability, not a phase prior to development, and what is learned through it feeds into the Product Backlog. A problem with a known cause and clear success criteria does not need Design Thinking; it needs good engineering or good management.
Origin and evolution
It has its roots in industrial design and business innovation. Herbert Simon described design thinking in The Sciences of the Artificial (1969). IDEO, led by Tim Brown, popularized the methodology in the 1990s by applying it to different industries. Stanford's d.school systematized it as a five-phase framework and made it accessible to all kinds of teams.
Key principles
- User-centered: the aim is to understand people's needs and problems in depth, not to assume they are already known.
- Iterative and experimental: multiple tests and adjustments are carried out. Failing early and cheaply is part of the process, not a failure.
- Collaborative and multidisciplinary: different perspectives and disciplines produce richer solutions.
- Learning through error: experimenting without fear of failure is a necessary condition for innovation.
The five phases
The five modes of Stanford's d.school (empathize, define, ideate, prototype, test) are the best-known representation, and the d.school itself put them into perspective in 2016: "they were just a first recipe, a suggestion to get started." Since then, it has taught eight design abilities (among them navigating ambiguity and designing your own design work). Other current frameworks describe the same territory: the Design Council's Double Diamond (discover, define, develop, deliver, with its organizational and systemic extensions), IDEO's three spaces, LUMA's 36 methods, and IBM's loop. All of them are best read as maps for choosing practices according to what the team needs to understand, imagine, make tangible, or learn; none of them is a mandatory sequence.
The process is not linear: you can go back to earlier phases when learning calls for it.
Empathize
Understand the problem from the user's perspective through direct observation, interviews, and context analysis. The goal is to understand who the users are, what they feel, and what they need. Empathy maps or user personas can be used as synthesis tools.
Define
Synthesize the findings from the empathy phase into a clear problem statement. Identify patterns, frame the challenge, and summarize the key needs in a sentence that is specific yet open to creative solutions. Separate symptoms, possible causes, constraints, and implicit assumptions. There is no "real problem" simply waiting to be dug up: there are better and worse frames, and reframing seeks to find out whether there is a better problem to solve (the slow elevator fixed with a mirror on the landing). The point of view (user, need, insight) and the "How might we…?" question are the artifacts of this phase, and the definition is revisited throughout the cycle.
Ideate
Separate generation from evaluation and seek variety, not just quantity: a session with 80 ideas from four families has diverged less than one with 30 from 12. Evidence accumulated since 1958 shows that oral group brainstorming produces fewer and worse ideas than the same people working separately; the approach that works is alone, then together: each person generates ideas silently and in writing (quick sketching, brainwriting), and then the group shares, builds on, clusters, and selects ideas using explicit criteria. Two or three distinct ideas are taken to prototype, not just the one the majority accepts.
Prototype
Build quick, inexpensive representations of ideas to assess their feasibility. A prototype is any representation made to answer a design question (Houde and Hill, 1997). Fidelity is set by the question, not by how mature the idea is, and it is raised only in the dimension that question requires: real content with rough visuals to test comprehension, faithful interaction with gray screens to test a flow. Low and high fidelity detect the same usability problems, and the more finished a prototype looks, the less people listen to the feedback. Services, processes, and policies are prototyped too, not just screens.
The key is to iterate: refine with each test.
Test
Put the prototype in users' hands to gather real feedback. Usability tests, interviews, and, where applicable, A/B tests are carried out. It is essential to let users discover the prototype on their own, listen actively, and not justify design decisions.
Prototype as if you were right, and test as if you were wrong: testing teaches about the solution and also about the problem, and it may reveal that the problem was poorly framed. The closing question is not "Did they like it?" but "What surprised us?"
Design Thinking and AI as a Tool
Generative AI is affecting different phases of the process:
- Empathize: language models can analyze large volumes of reviews, support tickets, or written user feedback to identify thematic patterns. However, they cannot replace direct observation or the building of real empathy: AI can summarize what users say, but it cannot capture the emotional and situational context observed in an in-depth interview.
- Ideate: AI can generate variants and combinations of ideas at great speed, which can be useful as a stimulus for the team's ideation. The risk is that AI-generated ideas tend to converge on "average" solutions that already exist: the most innovative ideas usually emerge from the diversity and direct experience of team members.
- Prototype: generative AI is significantly reducing the cost of the prototyping phase. Text, image, interface, and code generation tools make it possible to build higher-fidelity prototypes in less time.
Using AI in design work is a process decision: at each moment you decide who leads, the person or the AI, and what is preserved (capability, originality, and ownership of the decision). In ideation, AI comes in after the human round, because it raises the individual average and reduces the variety of the set as a whole. AI-generated "synthetic users" are useful for rehearsing a test script, not for validating a concept or as evidence about real people. A conversational prototype tested with people must disclose that it is a machine (EU AI Act, Article 50).
Common mistakes
Skipping the empathy phase or doing it superficially. It is the most uncomfortable phase (it requires getting out of the building, talking to real users, and tolerating ambiguity) and the one most often cut short. A team that defines the problem without having observed and listened to users is solving the problem it thinks the user has, not the real problem. All subsequent phases depend on the quality of the empathy work.
Empathizing without data. An empathy map or a persona filled in inside the meeting room, without observation or interviews, captures the team's assumptions dressed up as research. The answer is not more imagined empathy, but real contact with, and participation of, the people affected; the d.school itself has removed the empathy map from its Bootleg and speaks of "care" rather than "empathy."
Resources
🏦 Skill Design ThinkingState of the art and development guide · Skill Arena
🔧 Desafío de empatía [Empathy Challenge, in Spanish]Tool · Scrum Manager
🔧 Definición de problemas [Problem Definition, in Spanish]Tool · Scrum Manager
📄 Apuntes: Design Thinking. Lidera el presente. Crea el futuro [Notes: Design Thinking. Lead the Present. Create the Future, in Spanish]Scrum Manager · PDF
📄 Apuntes: Design Thinking Playbook [Notes: Design Thinking Playbook, in Spanish]Scrum Manager · PDF
📄 Tres técnicas de Design Thinking [Three Design Thinking Techniques, in Spanish]Scrum Manager Blog · Sep 2024
📄 ¿Un mapa de empatía puede ayudar a identificar resistencias al cambio? [Can an Empathy Map Help Identify Resistance to Change?, in Spanish]Scrum Manager Blog · Feb 2025
References
- Brown, T. (2008). "Design Thinking", Harvard Business Review.
- Brown, T. (2009). Change by Design. HarperCollins.
- Carter, C. (2016), "Let's stop talking about THE design process", d.school.
- Houde, S. and Hill, C. (1997), What do prototypes prototype?
- Lewrick, Michael; Link, Patrick; Leifer, Larry. Design Thinking Playbook.
- Serrano Ortega, Manuel; Blázquez Ceballos, Pilar. Design Thinking. Lidera el presente. Crea el futuro [Design Thinking. Lead the Present. Create the Future, in Spanish].
See also
Want to advance in agility? You can search for course and exam calls, or go at your own pace by becoming a member of the Agile Club. This membership includes exclusive resources, e-learning classrooms, and access to Skill Arena: a space to practice and measure your agile skills at your own pace.