Scaling a B2B Workflow for Enterprise
As founding UX engineer on a B2B sustainability platform, I redesigned a core service workflow, shifting consultants from back-stage setup to client-facing deliverables and helping the business close a large deal with a global electric vehicle manufacturer.
Problem
A core task that broke at scale
The platform's main workflow requires users to connect thousands of items to the right reference sources. Think of it like tagging every row in a massive spreadsheet, one by one. For small projects, this worked. It wasn't built for enterprise scale.
Through session recordings in PostHog, demo observations, and stakeholder interviews with the consulting team, I identified mapping as the service delivery bottleneck, the back-stage work gating every client engagement:
rows in the largest enterprise projects
of manual configuration per project
every row configured individually, no bulk actions
A key requirement for closing a large deal with a global electric vehicle manufacturer: the prospect needed confidence the platform could handle their scale before signing.
Key Decision
Replacing repetition with a single action
The same materials appear across dozens of projects, yet every new project required connecting them from scratch. I designed mapping presets: reusable templates that operationalise expert judgement, turning tacit consultant knowledge into shared service infrastructure the whole team can apply in seconds.
Outcome
Two weeks → one day.
reduction in project setup time
more projects delivered with the same team
saved per client each year in consultant hours freed for billable work
The workflow that was blocking enterprise adoption became a selling point in demos and helped the business close a large deal with a global electric vehicle manufacturer. Consultants who spent days on manual setup now start delivering value immediately, turning hours of configuration into recoverable billable time.
With the bottleneck gone, consultants now spend their days on the actual deliverable: impact breakdowns, hotspot analysis, and supplier comparisons that clients pay for. The redesign rebalanced the service: hours of back-stage configuration moved into front-stage analysis.
*Estimated at £75/h × ~72 hours saved × 3 projects per client.
Process
How I built it
I owned research, PRD (Product Requirements Document), and design end-to-end, shaping the work with product and engineering, and contributed directly to front-end development and code review. Before committing to build, I put working prototypes in front of consultants to test the matching logic on real project data.
The research became design principles the team still uses for any automation: never hide what the system did, always show what matched, what didn't, and why, and keep every automated decision reversible. Those principles became interface patterns and UX writing, from interface copy to help centre guides for customers and customer success.
I built an AI-native design workflow: Claude skills, codebase rules and design system context, with Figma MCP and Claude Design as tools, so AI builds with our real components and Figma, Storybook and production don't drift apart. To keep the feedback loop running after launch, I set up a CX agent that connects Fireflies, Granola and PostHog, turning customer calls into tickets, alerting the team when people get stuck, and surfacing UX gaps for the roadmap.
