EasyEV AI : A Voice-First EV Buying Assistant

EasyEV AI helps people explore, compare, and take the next step toward buying an electric vehicle through natural conversation. It combines a voice-controlled virtual showroom, an interactive AI consultation room, EV comparison debates, and buyer-support workflows in one platform.

Built by Team CodeHackers for EchoSphere 2026, the project uses Agora as its core real-time conversational layer.

THE PROBLEM

Choosing an EV involves more than comparing its listed price. Buyers need to understand range, charging access, running costs, vehicle suitability, and what to do next.

These questions are often spread across specification pages, calculators, charging maps, and dealership conversations. EasyEV AI brings these activities into a shared conversational experience, where a question can produce both an explanation and a useful visual result.

OUR SOLUTION

A buyer can describe their needs, explore vehicles, ask follow-up questions, and compare options while the interface updates alongside the conversation.

For example, a buyer can ask to compare two EVs, examine a vehicle, estimate ownership costs, and find nearby charging stations. The consultation retains useful findings in a Decision Passport/report and provides pathways to booking assistance or a human specialist.

KEY FEATURES

1. Voice-controlled virtual showroom

Explore supported vehicles through interactive 3D/360° views and vehicle-specific voice assistance. Spoken navigation commands control supported showroom views while the AI explains the selected vehicle.

2. Live AI consultation and Smart Stage

Ask questions naturally and receive relevant visual results during the conversation. The Smart Stage presents vehicle comparisons, images, ownership-cost estimates, charging information, and decision reports. Clickable actions provide an alternative way to trigger consultation tools.

3. EV comparison debate

Hear competing perspectives on two vehicles through a spoken AI debate, helping buyers examine trade-offs in a more engaging format.

4. Ownership and charging assistance

Estimate ownership costs using the buyer’s inputs and look up nearby chargers with location permission. These outputs support exploration and are presented as estimates rather than guaranteed purchase outcomes.

5. Decision Passport and reports

Retain consultation details and findings in a structured summary that buyers can review and download.

6. Booking and human assistance

The project includes test-drive booking workflows and human escalation. A specialist can join the buyer’s existing Agora channel, allowing the consultation to move from AI assistance to a human conversation.

7. Dealer and CRM workflows

Voice-assisted dealer registration reduces manual form entry, while CRM workflows support the management of buyer information and follow-up activity.

HOW WE USE AGORA

Agora is central to the live buyer experience:

• Agora RTC carries real-time microphone and agent audio.
• Agora RTM and the agent client toolkit deliver transcripts and conversational state.
• Server-created conversational agents power consultation, vehicle-specific assistance, and the comparison debate.
• Interruption handling allows users to redirect the conversation.
• In public HTTPS mode, Agora’s MCP connection exposes EasyEV decision tools to the consultation agent.
• Human specialists join the same RTC channel for handoff.

Showroom visual commands use a local command interpreter driven by transcripts. Local development also supports a browser-to-server tool bridge. Dealer registration uses a custom conversational flow rather than the full Agora agent pipeline.

TECHNOLOGIES

• Frontend: HTML, CSS, JavaScript, interactive vehicle assets
• Backend: Node.js
• Real-time conversation: Agora RTC, RTM, Conversational AI, and agent client toolkit
• Tool integration: Model Context Protocol and application APIs
• Language and speech: OpenAI, with optional Sarvam Bulbul v3 for Hindi/Hinglish and the existing speech-provider fallback
• Supporting integrations: charging lookup, CRM, calendar/booking, and email services where configured

WHAT MAKES IT DIFFERENT

EasyEV AI connects conversation to a visible decision-making workspace. Buyers can move between exploring a vehicle, comparing alternatives, calculating costs, and requesting assistance without repeatedly starting over in separate tools.

The combination of a virtual showroom, visual consultation, spoken comparison debate, and human handoff makes the experience useful across multiple stages of the EV buying journey.

ENGINEERING CHALLENGES

A key challenge was keeping spoken requests, tool execution, and visual results synchronized across consecutive questions. The implementation combines conversational agents with direct action controls and structured result rendering.

Other challenges included handling Hindi/Hinglish speech, maintaining vehicle context, preserving session information during handoff, and keeping internal tool identifiers out of user-facing responses.

CURRENT LIMITATIONS

• This is a hackathon prototype with a curated vehicle catalog and limited visual-asset coverage.
• Vehicle prices, specifications, charging information, and ownership estimates may differ from current local conditions.
• Live voice features require network access, browser permissions, and configured provider credentials.
• Booking, CRM, and email delivery require their respective external services to be configured.
• Public Agora MCP operation requires an HTTPS endpoint reachable by Agora.
• Human handoff requires an available specialist.
• End-to-end behavior depends on the configuration of the submitted deployment.

FUTURE SCOPE

Expand verified vehicle and dealer coverage, strengthen live inventory and booking integrations, improve regional-language support, and evaluate recommendation quality and usability with real buyers.

TEAM CODEHACKERS

Shivam Mahendru — Team Leader
Satvik Kesarwani
Jatin Vats

PROJECT LINKS

Live prototype: 
Demo video: https://www.youtube.com/watch?v=NSrUyxcUV40 
Source code: https://github.com/shhhivam12/easyev