I wanted to share something I’ve been building called Spana Time (https://spanatime.com). It’s an AI-powered conversational commerce platform that handles everything in the automotive ecosystem—dispatching mechanics, sourcing spare parts, towing, and rentals—entirely through WhatsApp. We are currently operating in Uganda.
The Problem
The automotive repair and parts industry here (and in many emerging markets) is deeply fragmented. When your car breaks down or you need a specific part, the last thing you want to do is download a new, single-use app, fill out complex forms, or drive around to five different shops to compare prices. Furthermore, the industry suffers from a massive "trust gap"—it’s incredibly hard to know if you're getting a fair price or a reliable mechanic when you are stranded.
How it Works
Instead of forcing users into a native app, we built the entire pipeline on top of WhatsApp, which is already installed on almost every smartphone here.
Intent Extraction: A user texts our bot a problem (e.g., "My car stalled, need a tow" or a photo of a broken brake caliper). The AI parses the intent, extracts vehicle details, and identifies parts.
Geospatial Broadcasting: It then pings verified local mechanics, tow trucks, or parts shops within a specific GPS radius of the user.
Bidding/Quoting: The providers receive the lead and reply with their price. The user gets a list of options right in the chat and picks the best one.
Solving the Trust Issue
The biggest thing I noticed while building this is that standard 5-star rating systems are practically useless for weeding out bad mechanics; they are too easily gamed.
To solve this, we built a trust engine that goes deeper than just aggregate stars. The system analyzes the actual sentiment of post-job written reviews to calculate a dynamic trust score for each provider. We also generate automated weekly performance reports for the mechanics, showing them their job completion rates and customer feedback highlights. If a provider's trust score drops too low, the system automatically stops broadcasting leads to them. This ensures that genuinely top-tier mechanics are rewarded with the most work, rather than just whoever has the biggest marketing budget.
I’d love for you to poke around the site (you can see live examples of the chat scenarios on the landing page). I'd deeply appreciate any feedback on the concept, the onboarding flow, or the approach to parsing conversational intent for physical services without forcing an app download.
I'll be hanging out in the comments to answer any questions!
Hi HN,
I wanted to share something I’ve been building called Spana Time (https://spanatime.com). It’s an AI-powered conversational commerce platform that handles everything in the automotive ecosystem—dispatching mechanics, sourcing spare parts, towing, and rentals—entirely through WhatsApp. We are currently operating in Uganda.
The Problem The automotive repair and parts industry here (and in many emerging markets) is deeply fragmented. When your car breaks down or you need a specific part, the last thing you want to do is download a new, single-use app, fill out complex forms, or drive around to five different shops to compare prices. Furthermore, the industry suffers from a massive "trust gap"—it’s incredibly hard to know if you're getting a fair price or a reliable mechanic when you are stranded.
How it Works Instead of forcing users into a native app, we built the entire pipeline on top of WhatsApp, which is already installed on almost every smartphone here.
Intent Extraction: A user texts our bot a problem (e.g., "My car stalled, need a tow" or a photo of a broken brake caliper). The AI parses the intent, extracts vehicle details, and identifies parts.
Geospatial Broadcasting: It then pings verified local mechanics, tow trucks, or parts shops within a specific GPS radius of the user.
Bidding/Quoting: The providers receive the lead and reply with their price. The user gets a list of options right in the chat and picks the best one.
Solving the Trust Issue The biggest thing I noticed while building this is that standard 5-star rating systems are practically useless for weeding out bad mechanics; they are too easily gamed.
To solve this, we built a trust engine that goes deeper than just aggregate stars. The system analyzes the actual sentiment of post-job written reviews to calculate a dynamic trust score for each provider. We also generate automated weekly performance reports for the mechanics, showing them their job completion rates and customer feedback highlights. If a provider's trust score drops too low, the system automatically stops broadcasting leads to them. This ensures that genuinely top-tier mechanics are rewarded with the most work, rather than just whoever has the biggest marketing budget.
I’d love for you to poke around the site (you can see live examples of the chat scenarios on the landing page). I'd deeply appreciate any feedback on the concept, the onboarding flow, or the approach to parsing conversational intent for physical services without forcing an app download.
I'll be hanging out in the comments to answer any questions!