White-Label Client Management Tool

2025

Researched, designed, built and deployed an internal tool with AI-assisted coding that replaced a manual white-label onboarding process. No developers involved.

WL Hub clients view: summary counts for total, complete, in-progress, blocked and pre-sales clients above a filterable client table

Context

The company offers white-label versions of its platform to partner organizations. Every new client meant the same work, and there was no system for any of it.

Branding assets
Logos, colors and files collected for every new client.
Every surface
Applied across the portal, the mobile app, cards and emails.
Review and sign-off
Every client checked the result before development.

Problem

I interviewed the person who ran the process day to day. Requirements lived in spreadsheets, and nothing was checked before it reached design and development.

50%+
of clients lacked branding materials
Some sent logos scraped from their own website.
Assets everywhere
Files arrived through emails, Slack, whatever channel the client preferred.
Format back-and-forth
Constant rounds over file formats.
Spreadsheets
Requirements tracked by hand, with no checklist.
No validation
No standard way to check what had been collected before handoff.

My role

A side project at 10% of my time over three months: I ran the research, designed the tool, built it with AI-assisted coding and deployed it to production. No developers were involved.

Propose
Wrote the brief and got buy-in from leadership.
Research
Mapped the white-label process from pre-sales to maintenance.
Design
Validated the tool with the person who’d use it daily.
Build and ship
Built and deployed it myself with Claude Code.

Process

Mapping the journey

Pre-sales, asset collection, design, client review, development, testing, handoff and maintenance each had their own pain points. The bottleneck wasn’t one step: it was the missing structure across all of them.

From pain points to solutions

A “How Might We” question for each key problem, then a brainstorm with the team. The solutions fell into four groups: clearer expectations with clients, a standard checklist, an internal tool, and a new way to assign work.

8
process phases mapped
4
solution directions
3
months, as a side project
10%
of my time

Simplified customer journey map of the white-label process: presale, then implementation steps from handoff and asset collection through design, client review, development, UAT testing and maintenance, each with columns of notes — Mapped end to end, the bottleneck wasn’t one step but the missing structure across all of them.

How Might We brainstorming boards, one per process phase, with rows of sticky-note ideas and the prioritized picks grouped below each board — One question per pain point, then prioritized: the internal tool was one of four answers, not the only one.

Solution

A workspace for every client

Challenge: Assets arrived through any channel, and nobody could see what was missing or how it would look before development.
Solution: A tab for each branding surface collects only the assets it needs and previews them in place. Brand colors set once cascade to every tab, and a dashboard shows each client's phase, owner and blockers.

Brand colors are set once here and cascade to every other branding tab.

Each surface collects only the assets it needs and previews them in place.

Email branding tab: an account-verification email template rendered with the client’s button color, next to the image settings — Seeing the email in the client’s colors replaces a round of review before development.

Pre-sales previews in minutes

Challenge: Every prospect needed a designer to mock up a branded version of the product: a day of work each time.
Solution: Sales upload the prospect's logo, pick colors and show the branded product during the conversation, without waiting on design.

Sales can show a prospect their own branded product during the conversation.

Built without developers

I didn’t hand designs to a development team. I built the tool myself with Claude Code and took it to production.

Database to deployment
Built end to end with Claude Code.
A stack to ship fast
I researched the options and chose one.
Plan for the CTO
I presented the deployment plan myself.
In daily use
A production tool, not a prototype.

Outcome

Replaced an entirely manual process.

Asset collection, progress tracking and branding previews moved from spreadsheets and email threads into one system.

Pre-sales previews went from days to minutes.

Sales generate previews themselves during client conversations, with no designer in the loop.

A designer-led initiative, built and shipped without developers.

From spotting the problem to deployment, no engineering resources were used. It changed how the team thinks about what a designer can deliver.

What I learned

Small scope, full ownership
At 10% of my time, owning it end to end is what gave it an outsized impact.
Designers can drive technical decisions
Recommending a realistic path to production kept it from dying as a Figma file.
AI coding changes what a designer can ship
The research, design and domain thinking were mine; the code came from a conversation with AI.