EXPERTISE · 2 SPECIALTIES · 5/7 PROVEN

Design.

Design at ARCHITEKT covers the interface and the process behind it: UX, UI and the design system that keeps them consistent, and the lean rework of the business process the interface is a window onto. It is applied inside the client organisation, in any of six purchasing forms.

State of the proof

counted from the ledgers below
7sub-capabilities
5proven · a published case carries a sourced figure
2held · carried by a named operator, no case published
0declared · in scope, no published proof today

What this is not

the boundary, stated

It is not a brand or campaign studio. ARCHITEKT designs the product a business runs on, not the advertising around it, and does not take identity or communication work.

It is not a redesign delivered from outside and handed over as files. The mandate installs the practice — the system, the conventions, the review — with the client's own team, because a design system nobody inside the company owns is a folder that ages.

"Product design" as a phrase lives here and only here. The Product expertise does not use it: what Product covers is discovery, delivery, quality and improvement, which are decisions about what to build rather than how it looks and behaves.

UX / UI / design system

Research, interface design and the shared component system that stops every squad reinventing the same button.

The work starts from evidence rather than from a moodboard: what users actually do, where they abandon, and which parts of the interface are absorbing support cost. That research produces a small number of decisions, and those decisions produce the interface — not the other way round.

The design system is the part that survives the mandate. Components, tokens, states, accessibility rules and the conventions for extending them, held in one place and versioned like code. It exists to remove a specific cost: without it, every squad re-solves the same button, the product drifts into five dialects of itself, and the drift is only noticed once it is expensive to reverse.

It is bought when a product has grown past the point where consistency can be held by memory — several squads, several surfaces, or a brand consolidation putting formerly separate products in front of the same user. It is the wrong purchase when there is one team and one surface: a system for a product that has not yet found its shape is overhead with a maintenance cost and no compounding return.

Accessibility is inside the system rather than beside it. Contrast, focus order, target size and reduced motion are properties of the components, so a compliant interface is what the team gets by default instead of what an audit asks them to retrofit.

Done is measurable: a component library the client's own engineers extend without asking, a contribution path someone outside the original team has used, and an interface where a new surface can be assembled rather than drawn. Until those three hold, the system is a deliverable rather than a practice, and the mandate is not finished.

Product is where a strategy becomes a decision about what exists. The werk covers the discovery that finds the right thing, the management that sequences it, and the design system that stops ten teams from inventing the same button.

CX and UX design proven

Louvre Abu Dhabi · VERMEG

Product discovery proven

Louvre Abu Dhabi

Product management proven

Spoticar · EBP Informatique · Kantar XTEL

Design systems and product operations proven

EBP Informatique · Opella

Growth and SEO held

E-commerce revenue share driven from 4.7% to 7% at Celio; no isolated growth case published.

Product analytics held

Decision flows and product architecture mapped at Kantar XTEL.

Business model reframing proven

EBP Informatique

Lean business process

The process the interface exposes, reworked before it is redesigned — value-stream mapping, waste removal, and the measure that proves the change took.

An interface is a window onto a process. Redesigning the window while the process behind it stays broken produces a faster route to the same wait, and the improvement shows up in a satisfaction score and nowhere in the operating figures.

So the work maps the value stream first: every step from the trigger to the outcome the business actually cares about, with the real elapsed time against the time anyone is working. The gap between those two numbers is almost always the finding. Steps that exist to compensate for an earlier step, approvals nobody can name the risk for, and re-keying between systems are removed before a single screen is drawn.

The deliverable is the reworked process, the measure that proves it moved, and the owner who holds it afterwards. Without a baseline taken before the change, an improvement is a claim: the mandate takes that baseline first, even when it delays the visible work by a fortnight.

This is bought alongside a redesign, not instead of one. It is the wrong entry point when the process is sound and the interface is genuinely the constraint — mapping a value stream in front of a UI problem adds a month and no clarity, and the collective says so rather than selling the wider scope.

The commonest finding is not a slow step. It is a step that exists only because an earlier one is unreliable — a reconciliation, a second approval, a spreadsheet kept in parallel. Removing the compensation without fixing what it compensates for makes things worse, which is why the mapping goes upstream before anything is cut.

Where the reworked process crosses a system boundary, the mandate names who owns each side before the redesign starts. A process improved on one side of a handover and unchanged on the other moves the queue rather than removing it.

In scope, with no published case behind it today. The specialty is carried on mandates; nothing here claims a sourced figure until one is published.

How it is bought

six forms, not all six by default

Any expertise can be bought in six forms: fixed price, time and materials, short placement, long placement, permanent placement or named seat. Each specialty carries the forms that suit it.

Questions

from the mandates behind this expertise
What does the collective bring to a stalled product?
A reframing before a rebuild. At EBP Informatique a stalled EUR 1M-a-year SaaS bet was reframed onto a EUR 2M economic base, and a design system was unified across more than ten product lines — rather than launching another initiative beside the ones already there.
How is discovery run when the audience is not a user base?
By putting the real audience in the room. The Louvre Abu Dhabi mobile application was framed in a design sprint with 20 participants and structured around a three-phase visitor journey — before, during and after the visit — because a museum's product question is a visit question, not a screen question.
What does a first full product cycle look like?
Spoticar was the first end-to-end product cycle ever implemented at Groupe PSA: a used-vehicle marketplace carried from framing to market as one continuous mandate. The deliverable was not the marketplace alone but the proof that the cycle could be run at all inside that organization.
Where does product stop and design operations begin?
Where one product becomes ten. At Opella the answer was to make design a governance function across a ten-brand ecosystem, which is an operating-model decision that happens to be made by product people.

The other expertises

four neighbours