Shipped programme work: live constraints, not a deck. The test was whether registration could read as credible when research, UI, copy, and build moved through one disciplined pipeline. The bar held—calm on first read, then consistency through navigation, onboarding, and reassurance (invoices, contracts, returns) without trust snapping or the flow sounding like a retrofit.

The constraint was explicit: ship the registration story, field model, and product surface with research setting pace—Hotjar to expose failure modes, Useberry to validate flow calls, Claude for label and error drafts—so quality came from iteration, not volume.

Industrial e-commerce is judged in milliseconds: field order, spacing, and error copy either read as control—or noise. Every block had to earn space: headline promise, territory capture, states that behave like a desk handoff, and rows that carry proof without theatre.

The working stack was replay, prototype, and script—evidence first, then design calls: which failure repeated, which route shipped, what still reads intentional when someone files a ticket after hours.

What shipped is one coherent path on a live URL: a single voice for Distrelec, a registration thread from first entry to active account, and patterns that respect caution in slow buys—without the journey feeling like varnish.

Distrelec sits where catalogue depth meets procurement caution: parts, tools, infrastructure, six-region contracts, and the reality of reconciling purchase numbers. The brief was not prettier chrome; it was fewer duplicate identities, clearer proof, and speed finance can stand behind.

Strategy, UX, and UI moved in one lane—reviewed with stakeholders, revised against tickets and transcripts—so the voice could repeat without sounding canned. Same signal at landing, in the steps, and at service handoff. No ornament. Only accountable steps.

Research can pile up fast; trade UI is mostly editing: what you cut, how you stage fields, which words appear when the part number is real but the channel is glass. The discipline was anchoring loops in Hotjar (funnels, rage taps, form health), Useberry for flow checks, light AI for microcopy only where it shortened recovery—and humans holding the final string so nothing reads synthetic.

Each pass narrowed to one voice—compact type, calm spacing, completion paths (email first, territory next, contracts, returns) with obvious orientation. Cohesion meant the same operational register from programme headline to invoice line, not repetition copy-pasted.

Industrial buyers weigh certainty first—lead times, compliance, traceability—then velocity. The UI needed to read like a counter people trust: scan-friendly hierarchy, labels that line up with paperwork, prompts that sound like operations rather than retail charm.

I treated reassurance as spacing: rhythm with room to read, grouped fields, microcopy that states facts. It should scan light—while territory, tax, and permission rules stay explicit enough for audit.

Launch is where promise meets backlog. Flows were set for regional programmes, acknowledgements where consent matters, and recovery that does not live in a footer maze. The frame stayed blunt: state the registration story plainly, keep evaluation steady, and let support proofs carry the move into an active account.

Secondary proof—invoices, contract routing, returns wording—sat at commit, not buried as annex copy. The UI stayed simple: one narrative spine, calm review, and service evidence supporting the next action.

The palette borrows from catalogue UI practice: a dark shell for chrome, off-white for fields, colour reserved for error, success, or urgency—not decoration.

Neutrals carry SKU strings without dulling numerals; success and error carry equal weight so no one hunts for state. Every swatch had a job.

The visual system leans on a controlled wordmark—Distrelec with regional treatment where needed—rather than extra marks. Icons stay practical: validation, wayfinding, service entry that appear when friction surfaces.

Headlines stay sans for discovery; tabular styles carry prices, SKUs, and microcopy—so tone stays light, weight stays on legibility and count.

Registration touchpoints ("Verify your email", "Choose your region") were written as programme beats—dated, plain, no discount theatre. The same lines fed internal scripts so sales, support, and procurement were not improvising different stories.

AI widened draft options; people picked lines that fit a buyer who audits before they approve.

Confirmation and ticket language appeared where buyers need paper confidence: acknowledgements that match finance wording, next steps that mirror what support says on the phone, no surprise tone shifts.

Continuity was the measure: the same steady voice on the form should carry into emails and tickets, from credential check through to first PO.

AI can compress drafts, variants, and labels; design still owns risk, tone, and truth. Distrelec stayed a case in tight passes—fewer claims, steadier signals, and a registration route people reuse because it reads considered.

The live path is the proof: walk regions, permissions, and confirmation, and watch restraint carry the work.