Early AI commerce prototype
Automart: Making spare parts discovery and ordering easier through WhatsApp
Automart explores how spare parts retailers and distributors could improve ordering accessibility and increase revenue by reducing the manual work required to identify parts, check availability, and process WhatsApp inquiries.
Independent prototype built outside of employment. Business and user validation has not been completed.
- Role
- Independent build — product concept, experience design, and prototype development
- Timeline
- 2026
- Team
- Independent build
The problem
Spare parts ordering still runs on informal messages and manual catalog searches.
Spare parts inquiries arrive through WhatsApp, calls, and informal messages. Staff search catalogs and inventory by hand, while customers may contact several sellers before finding the part they need.
Automart explores how retailers and distributors could reduce that manual work — making it easier to identify parts, check availability, and process WhatsApp inquiries — and in turn improve ordering accessibility and revenue.
The model
A B2B2C service, built for businesses and their customers.
Primary customers
Retailers and distributors — the businesses whose catalogs, inventory, and ordering workflows the experience is built around.
End users
Mechanics, repair shops, and vehicle owners — the people trying to find the right part and get a clear response quickly.
Capabilities
What works today, and what's planned.
Working today
- Informational storefront website
- Product and business information
- Natural language catalog search
- Semantic matching for related parts or vehicle models
Planned
- Image based identification
- Clarifying questions
- Branded and generic recommendations
- Live inventory confirmation
- Unavailable item request capture
- Registered user ordering
- Order system integration
- Business analytics
- Marketplace capabilities
Future workflow
From a WhatsApp message to an order in the business system.
- 01WhatsApp inquiry
- 02Text or image input
- 03Clarifying questions
- 04Catalog and inventory search
- 05Matching branded and generic options
- 06Availability response
- 07Registered user order
- 08Order sent to the business ordering system
When inventory information is incomplete
If complete inventory information is unavailable, the agent clearly states what could not be verified and captures the request for follow up.
Validation status
Honest about what's validated.
Early prototype validating natural language catalog discovery and semantic matching.
No external business or user validation has been completed.
Next validation steps
- 01Test real inquiries
- 02Validate search relevance
- 03Measure successful part matching
- 04Test inventory data availability
- 05Measure inquiry to order conversion