AI-Powered Vehicle Light Selector Chatbot

Job ID: 40548038

Budget: $250 – $750 USD

Before applying, please explain in 5–10 lines how you would prevent the AI from recommending an incorrect product fitment. Generic AI chatbot proposals will not be considered.

We are looking for an experienced developer or development team to build an AI-powered fitment chatbot for our WooCommerce websites.

We sell vehicle lighting products, mainly for agricultural tractors and machinery. The chatbot needs to help customers find the correct lights, kits, connectors and accessories for their specific machine.

This is not a basic FAQ chatbot. It must work as a technical fitment and compatibility assistant using our own structured product and fitment data.

Our websites are built on WordPress/WooCommerce and hosted on SiteGround.

Example questions the chatbot should answer
What lights fit my John Deere 6155R?
What replaces OEM part number AL234085?
Is this light plug-and-play?
Which connector do I need?
Do I need an adaptor?
What tractor kit do I need?
Can I upgrade from halogen to LED?
Main requirement

The chatbot must use our own fitment data and WooCommerce product data. It must not guess compatibility.

The LLM/OpenAI should be used to write helpful customer-facing answers, but the actual fitment decision must come from our database.

Required features for version 1
WooCommerce website chatbot pop-up
Connection to OpenAI or similar LLM
Fitment database built from spreadsheet/CSV data
WooCommerce product sync
Product recommendation cards in the chat
View product buttons
Add-to-basket option if practical
Admin area to upload/edit fitment data
Chat logs
Failed search reporting
Customer service handover when fitment cannot be confirmed
Basic internal staff search tool for customer service
Data we can provide

We can provide:

Product SKUs
Product names
WooCommerce product catalogue
Fitment spreadsheet/CSV
OEM part numbers
Machine makes and models
Connector information
Customer service notes
Known fitment issues
Product descriptions
Safety rules

The system must follow strict fitment rules:

Never say a product fits unless confirmed in the fitment database
If unsure, ask the customer for more details
If there are multiple versions, explain the options
If connector fitment is uncertain, ask for a photo
If no match is found, offer customer service handover
Do not invent product SKUs
Do not invent OEM references
Do not recommend products purely because the wording sounds similar
Preferred architecture

We are open to recommendations, but our preferred structure is:

WordPress/WooCommerce on SiteGround
→ chatbot pop-up plugin/widget
→ external chatbot API/server
→ fitment database + WooCommerce product data
→ OpenAI/LLM
→ answer returned to customer

We do not want the full AI system to slow down or put risk on the main WooCommerce website.

WooCommerce integration

The chatbot should pull live product data from WooCommerce, including:

SKU
Product name
Price
Stock status
Product image
Product URL
Categories
Short description
Attributes
Related products
Internal staff version

We also need a simple internal version for our customer service team.

Staff should be able to search things like:

John Deere 6155R front lights
AL234085 replacement
UTV359 connector
New Holland T7 work light kit

The staff tool should show confirmed products, fitment notes, connector notes, suggested customer replies, product links and stock status.

Future features

Not all required for version 1, but we may want these later:

Customer photo upload
Better add-to-basket flow
WhatsApp/Facebook Messenger integration
Multi-site support for UK, Ireland, EU and Canada websites
Offline fitment lookup for staff
Sage or stock system integration
Warranty assistant
Multilingual support
Please include in your proposal

Please answer the following:

Have you built anything similar before?
What architecture would you recommend?
Would you build this as a WordPress plugin, an external app, or a hybrid?
How would you stop the AI from guessing fitment?
How would you connect to WooCommerce products?
How would the fitment CSV upload/edit area work?
What database/vector search approach would you use?
Estimated cost for version 1
Estimated timescale for version 1
Ongoing support/maintenance cost
Required skills
WordPress
WooCommerce
PHP
JavaScript
API development
OpenAI API or similar LLM integration
RAG/vector search
Database design
Chatbot development
Product search
Admin dashboard development
Important note

We are looking for someone who understands that this is a technical product fitment assistant, not just a normal chatbot.

The system should work in this order:

Search our fitment database
Check WooCommerce product data
Apply our rules and notes
Use the LLM to write the final customer-friendly answer

The AI must not decide fitment by itself.