I had a great conversation with Kike Peña about how systems thinking can make design more meaningful and effective, especially as AI speeds up our work.
Unlike traditional workflows, moving faster with AI often comes with a tradeoff: speed can skip over the deeper thinking that makes products meaningful. Kike reminds us that AI is a tool, not a replacement for human judgment. The best outcomes happen when teams combine AI’s creative speed with systems thinking and human insight to keep work grounded in real outcomes…not just fast outputs.
Watch my full Glaringly Obvious interview with Kike Peña:
Key Takeaways:
AI speeds momentum → Explore and test ideas fast.
See the whole picture → Map users, goals, constraints.
Depth with systems thinking → Make solutions reliable and lasting.
AI + human insight → Empathy and judgment keep it relevant.
Good questions, better outcomes → Ask why, who, and how.
“It’s fascinating to see how the entire technology world is being impacted and enchanted (in my opinion) by AI. This impact can bring us multiple benefits, but also challenges that can lead us to forget the essence of what works. AI in design is a promise of ‘speed,’ and this attribute can’t now be the sole quality standard for digital products. That’s why it’s essential to view AI as a means rather than an absolute truth.”
When paired with UX metrics like clarity, confidence, and completion rate, this approach helps teams see whether designs truly work, where they confuse, and where to improve before launch.
Discussion:
How are you combining AI speed with systems thinking so your team doesn’t skip over the deeper questions in design?
@Helge got me thinking on “what to why” …so I expanded on my conversation with Keke!
He framed in his “What to Why” post, the challenge is moving past what happened into why it matters. I believe with AI tools, integrations, and agents, the questions pile up.
What we found in our work is that design signals (as you frame them) work if they are based on research / knowledge that gives them context where they become meaningful.
e.g. we found measuring ‘time-on-site’ on it’s own and using it to assume any kind of value can be misleading.
But, when our research told us that physicians wanted two things from our experiences:
To learn
To learn it as fast as possible
The we could use time-on-site as a measure and assume that if #1 had happened (they learned something) then reducing #2 would mean increasing value to our customers.
So I would agree, if the design signals are given meaning first through underlying research.
This is really inciteful Helge. One of the difficulties we have with ‘analytics’ is that few people build a system to understand how to interpret them. What you’ve just described provides meaning to the numbers, which have no inherent value in and of themselves.
Yep, context matters! But there are more pieces to a design signal that create value beyond an insight.
In your example:
To learn what? In what way? In what frequency and with what tools, when. People describing what they want in dialog, it often allows them to fantasize in ways they will never act. If you did this repeatedly in comparative scenarios, you can contrast what physicians actually mean by learning. UX metrics help to add color.
Lot’s a way to use UX data that goes beyond analytics!
I like to question the current state of design. I am convinced that we can be more prominent in companies and, from there, make fundamental changes in how we offer products to users.
Love this connection between user needs and business outcomes! I’m curious @Helge what type of research did you conduct to determine that physicians want those two things above all?
@MoData the research was very traditional qualitative interviews. In the Netherlands if I remember correctly. We found the same insights later in both Germany and the US.
Our approach was to use qualitative research to uncover insights that we could try to get to with creative use of experience design and quantitative measurements.
We got the idea from this article and tries to build on it / our own approach:
We wanted to eventually exchange the qualitative interviews with thick data collected from intentionally designed digital experiences. But didn’t get to the project before I left.
What I quickly found when joining the pharma-industry in 2017 is that most of their market research is essentially product response or brand response research. They are trying to figure out how their product is perceived or used.
But to your point Bryan, we wanted to find out e.g. how physicians are learning .. and we had almost no research on this.
We eventually found some research that we could deconstruct and reconstruct in order to pull out that e.g. physicians get a great deal of energy from learning, and they want to learn as a network, not as an anomaly etc..
My point though, is that if we are going to use design signals to learn, we can’t just collect them. We need to design our experiences to produce the right environment and context to learn the right thing.
And this is where I’ve felt the gap running back to my own beginning with the Internet: we’re designing experiences to help people do things or achieve things (which is good), but we are not designing them to learn explicit things.
Especially given the Internet’s exceptional quality of avoiding many of the traps of traditional research (because people are doing not answering, and because they don’t know they are being observed (Hawthorne effect)). It is a vastly untapped learning and insights resource for most organizations.
I like how this article outlines a bridge between CPIs and KPIs. That’s what we’ve been trying to do with UX metrics: create a system of leading indicators (CPIs) that can be mapped to lagging indicators and business goals (KPIs).
Absolutely agree, design must use learning loops to inform the business. Design signals become leverage that can be used to drive the business to better outcomes.
I dove into this topic to explain the idea.
The difference between insights and design signals are their application. Design signals must be used to drive better decisions in the work.
Discussion:
