An energy switching lead generation designed, built and shipped with AI
Creating a UK energy comparison and lead generation website from concept to a live product using AI tools
Role
Business partner
INDUSTRY
Energy & Utilities
Duration
3 months
Tools
Google AI Studio, Firebase, GitHub
PROBLEM STATEMENT
UK energy comparison and lead generation sites usually run on a SaaS platform or a full engineering team to handle quoting, routing and follow-up. A Utility Warehouse referral partner needed a live website that could capture leads, qualify them and route them to the right consultant automatically, without either of those overheads. The goal was to design and build the entire product end to end within weeks.
THE TEAM & MY CONTRIBUTIONS
I designed and built Switch Hunter, and the brands that came before it, from the first wireframe to the live backend, working directly with the person running the business to define lead qualification rules, routing logic and brand requirements.
Evidence & Reasoning
Building a pipeline of warm, ready-to-switch leads
Not every enquiry is worth the same. A lead still comparing options costs the same to handle as one ready to switch, but converts at a fraction of the rate. The business needed a steady flow of leads who had already decided to move, not just traffic filling in a form, which meant the site had to qualify interest before a lead was ever routed to a consultant, filtering out casual browsers early rather than treating every submission as equally valuable.
STRATEGY & APPROACH
Discoveries & Opportunities
Quick lead handling
Manual handling meant leads could sit for hours before anyone actioned them. The platform needed to route and notify automatically, the moment a form was submitted.Guidance over comparison
Customers switching energy suppliers are often unsure and have doubts about the process. A phone conversation with a consultant reassures them in a way a comparison table can't, and it's a touchpoint our target audience responds well to.
OUTCOMES & DELIVERABLES
System & Implementation Design
The build runs on Firebase Hosting with a GitHub Actions CI/CD pipeline, so every change goes from a prompt in Google AI Studio to a live deploy without a manual release step. Google AI Studio, using Gemini, became the primary implementation tool. Instead of specifying components in Figma and waiting on a build, I described the interaction, data flow and edge cases directly to Google AI Studio, then reviewed and iterated until it matched the intended experience.
Lead Distribution Engine
Behind the forms sits a Google Apps Script backend that receives each submission and distributes it using round robin logic across the client's Google Sheets, so leads are shared fairly and immediately instead of queued for manual sorting. Designing this meant thinking beyond the interface, into how a lead record should be structured, what state it needs to carry, and how the routing logic should behave as new submissions come in.
Notifications & Partner Communication
Each submission triggers two emails, built with Google Apps Script: a notification to the partner with the lead's details, and a branded confirmation to the customer summarising their potential cashback and next steps. Writing and building both templates meant treating email as part of the product interface, not an afterthought bolted onto a form.
IMPACT
Switch Hunter has generated and stored over 10,000 leads over the past three years. Conversion varies by season, but on average 2 in 10 leads convert within a month. Google Ads campaigns run at an average CPC of £1.15, with a 10% interaction rate.
