EVPick

About this app

EVPick is an AI assistant for choosing an electric car. You ask in plain words, such as “a seven-seater under £50k that charges fast”, and it answers from a catalogue of more than 1,700 EV versions rather than from memory. Answers come back as a short verdict with a sortable table, side-by-side comparisons, radar charts and price history, and every figure taken from the catalogue is underlined, so you can tell a spec from a comment.

It’s built for people shopping for an EV, with the UK and US first, and for enthusiasts who like the numbers. Around the chat there’s a free catalogue with a page for every brand and model, list prices by country, rankings such as longest range and fastest charging, 240 ready-made head-to-head comparisons, a comparison builder and a daily EV quiz with leaderboards. The site is published in nine languages.

How we built it

The problem

General-purpose chatbots answer questions about electric cars confidently, and often get the specs wrong. Car sites have the data but make you learn their filters, and prices differ in every country. We wanted the ease of a conversation with the reliability of a spec sheet, and a way for readers to check which parts of an answer are hard facts.

What we shipped

An AI assistant that answers from EVPick’s own catalogue instead of from memory. You can ask by budget, body style, seats, range, charging speed or where a brand comes from, and the answer arrives as a short verdict with tables, radar charts and price charts built from the matching cars. Every figure that comes from the catalogue is underlined in the answer, so a number the data doesn’t back up stands out.

Around the chat sits a full catalogue: a page for every brand and model, list prices by country, rankings, 240 head-to-head comparisons, a comparison builder and a daily quiz. It’s all published in nine languages with translated URLs, so each version can rank in its own market.

Under the hood

The catalogue covers more than 1,700 electric car versions, with list prices by country, official WLTP and EPA ranges and crash-test ratings. Prices are never converted between currencies: each market shows what that market actually charges.

The web app is built with Next.js, and the catalogue pages are generated ahead of time, so they load fast and are easy for search engines to read. The chat, the catalogue and the quiz all read from the same data, so a figure is the same wherever it appears on the site.