When UX Design Meets Unforgiving Complexity
Stéphanie Walter spends her working life inside some of the most demanding digital environments imaginable: enterprise applications for banks, dashboards that untangle dense regulatory processes, and interfaces where a user might be a lawyer, an engineer, or a customs officer. In a conversation with Smashing Magazine’s Vitaly Friedman, Walter explains how she got there, what it takes to design for users who use the product all day, every day, and why so much conventional design wisdom just doesn’t apply.
From Linguistics to Complicated Forms
Walter’s path into UX wasn’t linear. She holds a master’s degree in design and languages, a combination she admits is “a little bit strange,” and after an internship in Germany she was asked if she wanted to work on mobile apps and said yes without having done it before. She found early documentation in Apple’s human-computer interaction guidelines, which pushed her toward usability. But the real turning point came later, during her time working with a company that had contracts in many different sectors: that’s when she started to focus on enterprise UX, or as she describes it, projects that involved “ugly, complex dashboards” and deeply layered information architecture.
One assignment, a customs form with multiple levels of hierarchy, was so complicated she printed it on physical paper “and just drew lines to understand the hierarchy.” Numbers stacked on numbers, where every digit in a tax identification code carried meaning. When the form ran to more levels than there are HTML heading elements (<h1> through <h6>), she understood the scale of the challenge — and embraced it.
Rather than chase flashy marketing sites or immersive campaigns, Walter realized she felt most at home in the space most designers avoid. “I like complex, challenging, super heavy information architecture and solving problems for people who have to work with a tool on a daily basis,” she says. Part of the reason so little content exists about such projects, she suspects, is that they’re awkward to showcase. You can’t easily present a configurator for truck routes or a tool for customs processing in a portfolio. It’s less glamorous work, but no less important.
Tackling Legacy, Politics, and Wrong Questions
Walter is quick to point out that enterprise UX isn’t reducible to a single stereotype. “It necessarily depends,” she says. Sometimes you arrive at a project with nothing existing and no legacy code. Yet you still face a mountain of meetings, because the business rules are genuinely intricate and meeting time is the only way to understand them.
The bigger enemy, in her view, comes from shifting strategic fashions in the industry, which swing between building internal teams and buying off-the-shelf packages. “The business is so complicated — we bought a package and now we discover it doesn’t fit our needs, so let’s rebuild internally,” she describes. “But building things internally costs a lot of money, so let’s go back to packages.” That cycle, amplified by consulting firms large companies listen to, repeats every few years.
On top of that high-level churn is an even more insidious problem: finding out what users truly need. People with an “old-school” business analysis mindset may collect requirements by asking users what they want, rather than identifying what would solve the underlying problem. Walter offers an almost textbook example of this failure mode.
A user copies and pastes an entire table into Excel because there’s no export button, then deletes inactive rows. The takeaway isn’t that users need an Excel export; it’s that they need to filter out inactive items without leaving the browser. An analyst hearing only “give us an export to Excel” would miss that entirely. “You come here and you’re like, actually no, we’re not going to do the export,” Walter says, or at least the export is not the right solution for this specific need — a filter built into the table is.
From Server Logs to Seat-of-the-Pants Speed
Walter is honest that drawing sharp lines between redesigns of old interfaces and completely new builds can prove tricky. Internal tools, however, make discovery work faster than in external, B2B scenarios. Since she works in the internal IT department of the European Investment Bank, her users are colleagues. That means she can easily obtain a list of people who use the tool and schedule time with them. “When you go to people and say we’re going to talk about the tool you use in order to improve it, most of the time people are super happy,” she says.
The process starts with analytics like server logs, which indicate which pages are visited most frequently and should be migrated first. But numbers only tell so much: the team then matches visits to specific user tasks and runs discovery interviews structured around those tasks rather than around job titles. Walter dislikes relying on personas, noting if interviews produce 300 of them, they cease to be useful. Instead, the lens is the task itself: “We don’t discriminate based on persona — we discriminate and do the user research on specific tasks.”
Her toolbox in this phase includes open-ended interviews in which she covers topics even beyond the immediate one, keeping notes for possible future conversations on those areas. She also performs light observational study using screen sharing — asking users to show her how they perform a task on the spot and narrate their reasoning. The think-aloud protocol, as she explains it to participants, involves them saying out loud what they see and why they are clicking what they do. When people get silent, she nudges without guiding: “You stopped — what’s happening? Could you explain why?”
If the team has confidence a feature won’t produce many usability problems, they prototype quickly or go straight to implementation, then test on the live version. For riskier areas, low-fidelity prototypes precede a usability test. After launch, new users already go through the new interface with a “user diary” — a bank-appropriate Excel sheet — logging every defect, missing content, or frustration they encounter for a month. This feedback loop keeps product teams refining long after the official go-live date.
Finding Surprises, Despite Experience
After hundreds of such sessions, some user behavior becomes predictable, especially on dated interfaces with obvious pain points. “But no — sometimes on new things we have interesting results, and you’re like, yeah, that makes sense, we should’ve thought about this,” Walter admits.
Small details make a huge difference. When her team turned the user’s name into a mailto link that auto-fills an email with the correct project title, users were delighted because they no longer needed to copy and paste email addresses. Yet on another page displaying the same kind of person’s name without the link, a user immediately asked why the link wasn’t there. Walter says it was “a quick win,” but also proof that even after you know a space well, users still reveal what they rely on.
Surprising moments still happen inside customer journey maps, too. Walter underlines the maps aren’t a thing you make to have the map; they are artifacts you produce from research. “It’s as if I said I didn’t mention that I use wireframes — it’s the same thing,” she says. One map featured a triggering event labeled “human notification,” an assistant calling to say, “You need to validate this before 6 tonight. Could you please do it?” The human element remains the decisive trigger inside a digital process.
Why Enterprise UX Needs Different Skills
People who want to move into enterprise UX should work on specific capabilities. Walter points first to information architecture and making sense of data. Abby Covert’s writing on IA — making sense of messy content — along with her companion book on diagrams, resonate with this workflow. “You get a lot of information thrown at you in enterprise,” Walter adds, so you need to manage it carefully and present it in ways that align with how the user thinks.
But designers who cling to classic best practices often need to rethink their assumptions. “You need to understand that you might need to let go of all the UI principles taught in mainstream articles, like make the font bigger, put more white space. No — I have places who want a small font, who want as much data as possible on the screen, who don’t want to scroll.” The lessons that hold on marketing sites and blogs don’t carry over to dense internal web apps.
Working in this space has its compromises. Walter appreciates her manager and colleagues who shelter the team from enterprise politics so they can concentrate on users. Even so, she frequently sees wider gaps between what decision-makers expect and what a proper user-centered workflow would require. Yet she also sees the world differently after interactions with banking users: “We asked to invent changes people won’t use because they don’t solve the real problem.” That’s the core challenge she fights daily.
Listening to Articles and Making Time for Creativity
Given her intense role as UX expert, author of the customer journey mapping book, and teacher at universities in Strasbourg and Luxembourg, Walter remains a voracious learner. Her secret: she doesn’t read many long-form articles. She has her browser narrator read them aloud. “I put the headset on my ears and listen to the article while cooking, cleaning dishes, and doing manual labor,” she explains.
For consistent social media output, she batches her reading on weekends and then uses tools that allow scheduling posts on both Twitter and LinkedIn at once. The trade-off there, she admits, is that LinkedIn doesn’t always capture a person’s LinkedIn username correctly, forcing her to edit posts afterward.
Between engagements, Walter also paints digital illustrations — but as a hobby. Most often, she’s following Domestika craft courses that Instagram ads tease her with, ranging from illustration to pottery and woodworking. “I could not be a professional illustrator, because I don’t know how to illustrate something someone else tells me to do,” she says. This hobby protects her creative energy without the pressure of a client-facing portfolio.
Looking back, she wishes she had formal training in psychology, which would help her understand cognitive biases more deeply. But as she notes, the user-experience career path barely existed during her studies in France. Instead, she learned that field on the job — from academic HCI work, from UX agencies, and from her own side research.
Suppose she could pick a dream project. In that case, she’d want to work on service design or physical industry tools, such as equipment factory machines or cockpit interfaces — actual space-based UI design. “Airbus was looking for an intern, and I thought, I would’ve loved to be an intern there starting out,” she recounts, because such projects combine interface design with service design and process design. For now, however, she’s where she plans to stay, teaching designers how to tame complexity and quietly improving the tools people rely on for their most serious work.
Her parting thought belongs to the changing patterns around web3 — whether interfaces themselves are necessary, and how accessibility fits in. She cites a talk by Geri Reid she recently shared, noting Reid offered one of the few forward-looking conversations grounded in reality rather than marketing hype. As for the listener who made it to the end of a dense episode about design, Walter’s wish is simple and unsentimental: “Stay safe.”
Walter is on Twitter at @WalterStephanie, on LinkedIn, and at her website stephaniewalter.design.
