Operational interface design
Back-office and admin screens where people spend their working day. Density, keyboard paths and error recovery matter far more than visual novelty here.
A dispatch desk where the twentieth order is as fast as the first.
Capability
Interface design for systems people use all day, where a saved click compounds across a thousand repetitions. Flows mapped against the real task with the edge cases drawn, a component library your developers can build from, and accessibility checked rather than assumed.
From $3,000 · 2–4 weeks · fixed scope, agreed in writing
What your team gets
Not a strategy deck. A running system, the permissions around it, and the code in your repository.
What we build
Weighted towards the unglamorous end — the interfaces people are paid to use, where small frictions are expensive because they repeat.
Back-office and admin screens where people spend their working day. Density, keyboard paths and error recovery matter far more than visual novelty here.
A dispatch desk where the twentieth order is as fast as the first.
Tokens, components and the rules for using them — so a feature designed next year looks like the one designed today without anyone policing it.
One component set feeding a web app, a marketing site and a mobile app.
End-to-end design for a customer-facing product, including onboarding and the states nobody remembers to draw.
A member portal from first sign-in to renewal.
Pages designed to convert, built as templates so the tenth page costs what the second did.
A capability page template that scales to a dozen services.
Fixing what is measurably wrong rather than restyling everything. We start with where people actually fail.
A checkout reworked around the two steps that lost the most people.
An audit of an existing interface against WCAG 2.1 AA, with fixes ordered by how many users each one affects.
Keyboard traps and contrast failures found and fixed in one pass.
How we build it
The gap between a design and what gets built is where most design budget is wasted. Everything here exists to close it: decisions made as tokens and components rather than as pictures, and the awkward states drawn before a developer has to invent them.
We sit with people doing the task before drawing anything. What they actually do and what the process document says differ every time, and the difference is where the design earns its money.
Empty, loading, partial, error, offline, too-much-data. These are most of real usage and almost none of most design deliverables, which is why developers end up inventing them at 5pm.
Colour, type, spacing and radius defined as named values that map one-to-one onto code. A change to the scale then propagates rather than being reapplied screen by screen.
The handoff is a component library with states and props, not a folder of flattened screens. It is what makes the build match the design without a week of back-and-forth.
Contrast, focus order, target size and keyboard paths checked with tooling and by hand. Retrofitting accessibility costs several times what designing it in does.
Consumer design optimises for first impression. Operational design optimises for the thousandth repetition, and the two pull in opposite directions. More per screen, fewer clicks, keyboard paths for the frequent actions, and generous targets only where a mis-tap is expensive. We design for the person who has been using it for six months, not the one seeing it for the first time.
Design tokens are the one deliverable that survives. When colour, type and spacing exist as named values in both the design file and the codebase, a change is made once. Without them, every design decision is re-litigated in code, and six months later the product has four greys that were meant to be one.
We do not produce a hundred-slide brand exploration, and we do not redesign something that is working to make it look current. If the measurable problem is two steps in a checkout, we fix those two steps. Most of the value in this discipline is in restraint.
Where it lands
Member portals where the users are not frequent and the task must be obvious.
Order desks where the same screen is used several hundred times a day.
Checkout and catalogue flows, redesigned around where people actually drop.
Interfaces where accuracy matters more than delight, and accessibility is not optional.
Dense data interfaces where a misread number is expensive.
Listing and enquiry flows across desktop and phone.
Technology
We pick the boring option unless there is a reason not to — the framework is the part most likely to be abandoned before your system is.
How the engagement runs
Every phase ends in a named deliverable. You can stop after any of them and keep what has been built.
Who uses it, how often, and what currently goes wrong. If the answer is "it looks dated" we will ask what that is costing before agreeing it is the problem.
Time with the people doing the task, the real flow mapped including the awkward paths, and agreement on what the design is actually meant to improve.
Tokens, key screens and the components behind them — designed as a system so screens that do not exist yet will still fit.
Empty, loading, error and edge states drawn, contrast and focus order checked, keyboard paths defined for the frequent actions.
A component library with props and states, responsive rules, motion specs, and a walkthrough with the developers who will build it.
Why Devs Core
Operational interfaces are judged on repetition, not first impression. Density and keyboard paths beat novelty.
Empty, loading, error and partial states are most of real usage and almost none of most design deliverables.
Named values mapping one-to-one onto code, so a change is made once instead of reapplied screen by screen.
A library with states and props, not a folder of flattened screens. It is why what ships matches what was designed.
Tooling plus keyboard and screen-reader testing to WCAG 2.1 AA. Retrofitting it costs several times more.
If the measurable problem is two steps in a flow, we fix those two steps rather than selling you a rebrand.
Questions
Time with the people who do the task, the real flows mapped including the awkward paths, design tokens for colour, type and spacing, key screens, a component library with states and props, responsive and motion specifications, an accessibility check to WCAG 2.1 AA, and a walkthrough with the developers who will build it.
UI/UX design starts at $3,000 over two to four weeks for a fixed scope agreed in writing. The variables are the number of distinct flows, whether a design system is being created or extended, and whether user research beyond stakeholder interviews is needed.
Yes. The handoff is a component library with a specification, deliberately built so another team can implement it — including your in-house developers. A design that only we can build is a design that has failed.
By delivering decisions rather than pictures. Tokens map one-to-one onto code variables, components are specified with their states and props, breakpoint behaviour is written down, and there is a walkthrough with the build team. Most divergence between design and build comes from states nobody drew.
Part of it. Contrast, focus order, target size and keyboard paths are checked with axe plus manual keyboard and screen-reader testing against WCAG 2.1 AA. Designing it in costs a fraction of retrofitting it, which is the entire argument.
We do contextual enquiry as standard — watching people do the actual task, which is where most of the useful findings come from. Formal usability testing, surveys and funnel analysis are available where the decision genuinely needs them, and we will say when it does not.
Yes, and usually should. We take an existing brand into tokens and components rather than proposing a new identity. Where the brand genuinely blocks usability — contrast failures being the most common — we will show you the specific conflict rather than argue for a rebrand.
Not on its own, and we will ask what it is costing before agreeing. If the answer is a measurable drop-off, a support burden or a sales objection, that is a real brief. If it is only unfamiliar to look at, a restyle of the components is usually cheaper and safer than a redesign.
Thirty minutes, no preparation, no deck. You describe what keeps going wrong and we tell you whether AI is the answer — including when it is not.
Prefer email? contact@devs-core.com