AI tools
How to use artificial intelligence in design thinking so that it supports learning and leaves the thinking to you.
How to use AI in design thinking
On this page you will find the principles of working with AI, tips for each stage of the design thinking process and a list of tools.
Use AI as a tool, keep the thinking yours
AI is becoming a tool as common as a word processor or a search engine. Used well, it can give the design thinking process real support. Used badly, it can harm the learning process itself, which is after all the main goal of this method in university teaching. What matters is the difference between two ways of using AI.
Consciously using it as a tool that supports your work, for example to look for information, analyse data or generate ideas. It is a partnership in which the final decisions and the responsibility stay with the human.
Handing it the responsibility for the creative process without reflection and accepting generated content uncritically. It leads to shallow solutions and leaves students without the key competences.
It is like moving heavy objects: technology is invaluable when the goal is to transport something efficiently. When the goal is to build muscle, as in the gym, you have to do the lifting yourself. The same goes for education: if the goal is to gain competences, relying too heavily on technology can hold that process back.
Three questions before using AI
- 1Does AI support learning?Consider whether the tool supports the learning process or replaces it.
- 2Can I judge the result?Do I have the competence to assess AI’s output critically?
- 3Do I know the limits?Do I understand what this technology cannot do?
How to ask: the structure of a prompt
A well-described task gives better answers than a general question. A simple structure with four elements helps.
CONTEXT: Who you are and what you do
TASK: What exactly you need
FORMAT: What the answer should look like
CONSTRAINTS: What to avoid
Example. CONTEXT: I am an economics student working on a project about sustainable development. TASK: Identify 5 problems that link the theory of the tragedy of the commons with urban policy in my city. FORMAT: A list with a short description (2-3 sentences) and an example of a specific location. CONSTRAINTS: Focus on problems that students could realistically solve.
AI at each stage of the process
At each stage AI helps with a different task and carries a different risk. It pays to know both.
Discover
How AI can help- Quick research and synthesis of knowledge, also as a podcast made from chosen sources
- Root-cause analysis with the “5 Whys” method
- Simulating interviews with personas in voice mode
- Hallucinations and the illusion of accuracy: check key information in the sources
- No real empathy: contact with people is what a simulation cannot give you
- Shallowness: ask detailed questions with context
Define
How AI can help- Qualitative data analysis: themes, quotes and categories from transcripts
- First drafts of empathy maps and personas
- Variants of the How Might We question
- “Falling asleep at the wheel”: the final interpretation belongs to people
- Confirmation bias: ask neutral questions
- Loss of nuance, such as sarcasm and hesitation
Ideate
How AI can help- Creative techniques: SCAMPER, reverse brainstorming, analogies
- A virtual panel of experts and innovators
- Connecting distant fields
- The first-idea trap: start with a “human” brainstorm
- Homogenised solutions: use AI as a provocateur
- Decision paralysis: ask for the 10 most unconventional ideas rather than 100
Prototype
How AI can help- Visuals and storyboards from a description
- Generating code (vibe coding) for interactive mockups
- Sample content for prototypes
- Code that looks professional and hides technical problems
- Polishing the prototype rather than the solution: go back to the How Might We question
- Scope creep: a prototype should answer one or two questions
Test
How AI can help- Usability test scenarios for different personas
- Analysis of “think-aloud” transcripts and surveys
- Simulated tests when access to users is limited
- Limited grasp of context and emotions
- Bias in analysis: questions can steer towards your own hypotheses
- A false sense of completeness: real people are what a simulation cannot replace
Opportunities and risks in brief
- Fast analysis of information and synthesis of data.
- Simulating perspectives and practising empathy.
- Support for creativity and idea generation.
- Automating and speeding up prototyping.
- Democratised access to advanced tools.
- Shallowness and factual errors (hallucinations).
- Loss of authentic contact with people.
- Over-reliance on AI and weaker critical thinking.
- Data privacy and copyright.
- Reinforcing existing biases.
Tools we recommend
AI tools change fast, which is why their list lives on this page.
An AI platform built to support research, making it easier to work with academic papers. It helps with literature reviews, understanding complex studies and finding related publications.
The best-known AI tool. Voice mode allows a natural conversation and can be used to simulate interviews. Custom GPTs let you create a specialised assistant tailored to specific tasks in the design process.
A large language model that stands out for its advanced understanding of natural language and for answers of high, “human” quality. In the design thinking process it can be used to simulate a panel of experts, apply creativity techniques and visualise ideas as diagrams or mind maps.
An app made for working with documents and analysing larger bodies of text. Its distinctive feature is that answers rest solely on the materials the user uploads, which is useful for analysing interview transcripts or reports. It can also create audio and video summaries of uploaded materials, as well as mind maps.
An AI tool for data analysis and visualisation that lets users ask questions in natural language. It simplifies working with data by automating charts and reports with no coding needed. It handles large data sets.
An AI platform that creates user interfaces and web apps from text descriptions. It enables rapid prototyping of sites and apps.
More design tools