EVPicker - Fast EV Shortlists & Buying Tools

Update (2026)

The original idea behind EVPicker was a concierge-style EV shortlist service delivered within 48 hours. After exploring the concept and building the initial flow, the idea evolved into something slightly different: a faster, more lightweight EV discovery experience.

The core insight was that most people didn’t need a full concierge-style report. What they really needed was help cutting through the noise and narrowing down a few realistic EV options quickly.

The current direction focuses more on rapid EV shortlists and lightweight tools rather than a formal report-based service.

Background

• Created to test whether people would pay for a simple, human-reviewed EV shortlist.
• Designed to see if a quick, paid shortlist could drive decisions faster than more exploratory tools.
• Built to answer the hypothesis that wasn’t proven inside Motoru: will people pay directly for concise EV advice and concierge?

Problem

• Drivers ready to “go electric” still didn’t know which EV to pick and didn’t want another comparison tool.
• Wanted a human to look at their situation and say, “Here are the options that make sense.”
• Needed a helpful path into novated leasing/finance partners without feeling like a sales hand-off.

Approach

• Speed run in a day: shaped the offer, brand, copy and site in WordPress to move fastest.
• Added an AI chat bot on the homepage to field questions before people submitted details.
• Built a simple workflow behind the intake: review details, research, return a short, opinionated shortlist with reasoning.
• Kept the interaction tight: one clear input, one clear shortlist, with a gentle path into novated leasing/finance when someone had a favourite.
• Used the constraint to test willingness to pay and the service shape, not to prove a stack choice.
• Used Jotform to capture details and Stripe for payment so I could validate quickly without custom billing code.

Outcome / Evolution

The original version proved useful as a way to structure EV recommendations quickly, but it also highlighted an important insight: most people don’t want a long report or a concierge-style process. They want clarity fast.

That insight pushed the concept toward simpler tools and faster recommendations rather than a formal report service.

This exploration also informed later experiments around EV discovery tools and AI-assisted recommendations.

My role

• Sole contributor end to end: offer, product shape, UX, brand, content and delivery.
• Built the site myself in a day, plus intake, email and automation workflow.
• Handled every customer interaction and folded learnings into how I approach EV decision tools.

Current direction

The EVPicker idea is continuing to evolve. The current focus is exploring faster ways to help people shortlist EVs without requiring long questionnaires or manual research.

This includes lightweight tools, quick recommendation flows and experiments with AI-assisted discovery.

Stack

• Web frontend (simple UI for user inputs)
• EV vehicle data dataset (structured comparison data)
• AI-assisted recommendation logic
• Prompt-based analysis for insights and explanations

Supporting tools used during development included lightweight scripting and AI models to test recommendation logic and generate insights.

Architecture

EV Picker was built as a lightweight decision-support system rather than a traditional search tool.

Instead of asking users to filter through dozens of specifications, the system collects a small number of meaningful inputs about how the car will actually be used.

Typical inputs include:

• driving distance
• charging access
• budget range
• household needs (space, seats, etc.)
• preferences such as performance or technology

These inputs are processed by a recommendation layer that compares the user's requirements against a structured EV dataset.

The system then ranks vehicles based on suitability and returns a shortlist of recommended cars.

Recommendation Logic

The core of EV Picker is a simple scoring approach.

Each vehicle in the dataset is evaluated against the user’s inputs across several dimensions such as:

• range suitability
• charging compatibility
• budget alignment
• practical fit for household needs

Each dimension contributes to an overall suitability score.

Vehicles with the strongest overall match are returned as recommendations.

This approach allows the system to stay transparent and explainable rather than relying on a black-box recommendation model.

Role of AI

AI is used primarily to assist with interpretation and insight rather than to make the core recommendation decision.

Examples include:

• generating explanations for why a vehicle is recommended
• surfacing relevant EV considerations for the user’s situation
• helping interpret EV specifications and real-world trade-offs

This keeps the decision logic predictable while still allowing AI to add helpful context.

Implementation Notes

A key design goal for EV Picker was simplicity.

Many car comparison sites rely on complex filtering systems that overwhelm buyers with specifications.

EV Picker instead focuses on capturing a small number of meaningful lifestyle inputs and translating those into practical recommendations.

During development I experimented with several ways of structuring the recommendation logic and refining the EV dataset to ensure results remained practical and understandable.

Reflection

The project highlighted how confusing EV research can be for most buyers.

Rather than needing more information, many people simply need a clearer way to translate their lifestyle into suitable vehicle options.

EV Picker attempts to bridge that gap by combining structured vehicle data with simple decision logic and AI-assisted explanations.