🏬 Co-Working Customer Journey Mapping

Preamble

I gave myself the opportunity to experiment with a range of AI tools to explore how they might support early-stage UX research; specifically in generating preliminary personas and customer journey maps. The objective was to critically examine the assumptions these tools might introduce, and to assess the strengths and limitations of AI-assisted methods in a UX context.

While this remains a highly debated and evolving topic, my goal was to identify potential biases in the AI-generated outputs. Spoiler alert: I wasn’t able to spot any overt biases. However, this finding is far from conclusive. In hindsight, a more rigorous approach would have involved conducting a parallel, traditional UX research process "by hand" and then directly comparing it with the AI-generated results. Only then could a meaningful evaluation of accuracy, bias, and nuance be made... I think. What do you think?

Persona

To build the persona, I used ChatGPT, Claude3 and QoQo AI; a UX generative tool as a way to test how AI interprets user motivations, goals, and frustrations. I then expanded the research using custom prompt flows to simulate qualitative interviews, behavioral segmentation, and emotional journey mapping.

Customer Journey Mapping

To simulate a high-fidelity research environment, I designed and iteratively refined a prompt engineering workflow that guided the AI through key research stages: from persona development and behavioral analysis to touchpoint mapping and opportunity identification.

LLM/AI used to create CJM:

• ChatGPT
• Claude3
• QoQo

The AI's was instructed to:

• Generate hypotheses and synthesize secondary data,
• Simulate qualitative research outputs (e.g., interview themes, sentiment patterns),
• Construct detailed journey maps based on user goals, emotions, and actions.

Source: ChatGPT Deep Research prompts & project

QoQo instructions for CJM

Persona

• ⏱️ Age: 28-35
• 🇸🇪 Nationality: Swedish
• 💒 Marital Status: Single
• 🚺 Gender: Female (she/her)
• 📍 Location: Stockholm, Sweden
• 👶 Children: No children
• 💼 Profession: Marketing Strategist, remote-first, SAAS company
• 💶 Income: €45,000 - €60,000 / year
• ♿️ Accessibility: Colour blind

Scenario

"Seeking a vibrant and well-equipped co-working space in Stockholm to host client meetings, focus on strategic planning away from home distractions, and connect with other digital professionals while working remotely."

Conclusion

This project functioned as both a deep UX exploration and a reflective meta-analysis of AI’s evolving role in human-centered design. It illustrates how AI can be used to rapidly generate artefacts; such as personas or journey maps, that serve as valuable conversation starters with stakeholders.

Me having worked primarily in startup environments, I recognise that this approach has practical value in fast-paced settings like the startup scene. When clearly communicated that the data is synthetic and assumption-based, these artefacts can help bridge knowledge gaps. They provide a starting point for stakeholders to share domain-specific insights with the design team, while also allowing the design team to introduce and advocate for UX principles, methods, and user-centered thinking. In that sense, the exercise can become a mutual learning opportunity; aligning internal understanding around both product context and the UX process.

---

P.S Big shout to Nick Qurle who created this Glassmorphism 2.0 effect that I used for the cover image. I really love the texture. I modified it a little bit to give it a 'milky' vs. clear glass feeling.