Over Christmas I finally put words to it: I’m not frustrated with design as a craft, I’m frustrated with what our industry keeps rewarding. The incessant need to justify our existence. The chronic underfunding. The endless discourse about personas vs. archetypes. The constant stream of “UI best practice” posts on Medium, as if the world’s biggest problems are waiting for the right UI checklist.
I stopped going to UX conferences because too many talks orbit the same small set of deliverables. I stopped reading most “UX internet” content for the same reason: it’s either another method primer, another hot take about AI, or another thread of UI optimisation tips dressed up as insight.
I think we need to be honest with ourselves about what that looks like from outside. A profession asking to be taken seriously while arguing over its own tools.
This has to stop. As soon as UX entered the public realm “seamless” stopped being the harmless default.
Best practice became dogma
For the past two decades the governing paradigm for digital products has been “seamlessness.” Reduce friction, raise efficiency, get people to the thing they came for as fast as possible. We treat conflict as a “pain point” to be designed or engineered away.
This is great when you’re helping someone buy trainers or book a train.
However, as digital tools increasingly mediate democracy, public wellbeing, and economic participation, the limitations of “UX best practice” become glaringly obvious.
Infinite scrolling that gets people addicted, algorithmic information bubbles that amplify political polarisation, public services where decisions are made “behind the scenes” with no user involvement.
We’ve internalised that “good design is invisible.” We celebrate low cognitive load. We measure success with conversion, time-on-task, and completion rates. Eliminate friction at all costs, but friction is the not the enemy. Sometimes friction is the moment a person realises, “hang on, who decided this? On what basis? What are my options? What if this is wrong?”
The uncomfortable truth
The uncomfortable truth is that mainstream UX is optimised for profit. A large portion of what we call “UX maturity” is actually commercial maturity. Our best practices are standardised because they scale revenue. Our metrics are legitimised because they map cleanly to growth dashboards. Our case studies get celebrated because they show uplift. The kind of story that reliably gets you interviews: I built a thing and it resulted in £Xm in revenue uplift.
Now, there is nothing inherently wrong with designing for commerce. The problem is what happens when a discipline is optimised for selling is applied to policing, credit scoring, immigration, public health, and urban planning.
In these contexts, “seamlessness” can become a black box: a denied loan, a risk score, a decision about eligibility. A filtered view of reality—with no explanation, recourse, or accountability—presented as objective rather than design choices, politics, incentives, and biases.
When a system is seamless it often has no meaningful place to question it, no way to appeal or disagree. Seamless UI strips users of their agency.
Introducing agonistic design
Usually I just moan. This time I found a solution. I thrive in adversity, honestly, I kick myself for not coming across this sooner.
Agonistic design is an approach that makes room for friction, the struggle, fosters community and tries to make spaces for conversation between users and systems. It’s based on a book Adversarial Design (DiSalvo, 2015) and articulated further in a journal article The Right to Contestation (Collins et al., 2024). I’ll attempt to summarise.
Design should do the work of agonism—from the Greek word ‘agon’ or ‘struggle’, which is central to the political idea that all productive democracies are built on dissent and conflict—by creating spaces, artifacts, and interactions that allow for the expression of dissensus. It shifts the role of the designer from a “problem solver” who harmonises conflicting needs to a “provocateur” or “mediator” who reveals conflict and allows stakeholders to navigate it productively.
This means that in certain contexts (particularly those involving AI, public services, and high-stakes decision-making) our goal should not be to make the system easy to use but to make it “contestational.” We must design systems that allow users to question the system itself.
I love this idea.
Illustrative example
A few months ago I was arguing with my wife. The content is irrelevant and the reason is lost to time. I went to my favourite LLM and, since I wasn’t thinking all that rationally, asked it for advice on the situation. It completely validated my feelings. I went to another LLM. The output was the same.
The third one called me out. It pointed out what I was doing, named the problematic assumption in my framing, and forced a pause. That was a lightbulb moment. Systems that are optimised for agreement and “delight” can quietly train us away from self-critique. Sometimes what we need isn’t a smooth experience. Sometimes we need a moment of friction that creates space for accountability.
Agonism in public service
I highly recommend reading this paper in full. I’m probably going to do a terrible job summarising all of its intricacies and nuances.
In a two-year design ethnography with the City of Atlanta’s Mayor’s Office of Immigrant Affairs, researchers Corbett and Le Dantec (2021) developed the “Trust-as-Distance” framework. They were asked to design a system to help immigrants navigate housing code violations.
The “smart city” logic dictates that trust is built through efficiency. By making government services fast, digital, transparent, and by essentially removing distance. However, the researchers found that for vulnerable populations (undocumented immigrants) “distance” was actually safety. Digital records could create a trail that could lead to deportation.
Instead, trust was built by maintaining distance where it protected people, while creating closeness through human relationships. Through a combination of digital and non-digital components, like a Code Guide (a pamphlet used to explain rights and violations) and a Repair Request Webpage (no app or account needed).
By resisting the “Neoliberal Design Logic” of sticking everything into a seamless and efficient app and by using technology to mediate work rather than delegate it the researchers built trust, accountability, and equity into the service.
The paper closes on an idea that as HCI research (and User-Centred Design by proxy) moves deeper into the civic sphere, designers must recognise that trust is not a simple metric to be optimised but a complex political grant of power that carries the risk of legitimising injustice if misplaced. By adopting agonistic principles designers can move beyond transactional efficiency to focus on relationships.
How should designers understand (and when necessary) resist desires to make things “more efficient”? How should we view the role of design interventions: as mediators or delegators?
Closing thoughts
If UX feels stagnant it’s partly because we treat it as a set of deliverables and templates. Agonistic design isn’t the same as antagonism. It’s about designing legitimate ways to disagree with a system and be heard. It’s designing conditions for disagreement, explanation, and recourse; especially where these decisions affect people’s lives.
Here are a few ways to start, even inside a commercial design role:
- Add a “contestation check” to decision points (eligibility, ranking, refusal, enforcement)
- Design friction intentionally (confirmation pauses, “why am I seeing this?”)
- Make reasons legible (inputs used, “would would change the outcome”)
- Build recourse paths (appeal, correction, human review)
- Broaden your reading diet (HCI, Design Research, STS)
This is why I’m optimistic for 2026. The most energising and inspiring thing I read this year didn’t come from another UX trend cycle. It came from outside the bubble, through Design Research Society’s proceedings of Nordes 2025 conference.
So let’s stop arguing about the furniture arrangement of our methods and start building a mature design practice that can stand in the real world.
It’s by looking outside of UX that we get fresh ideas. It’s by having discussions about the nature of design that we grow as a field. It’s by educating new generations of UX designers on more than just “best practice” that we mature. It’s by creating spaces for discourse and dissent with the way things are right now that we build a better world.
We need dissent. We need discourse. We need agonistic design.
