🚗 About shopSenseAI

⚡ Inspiration

Running Pink Chicken Speed Shop as a mechanic and solo entrepreneur quickly highlighted how managing customer interactions while actively working in the shop became a significant challenge. Constant disruptions from calls and texts seeking quotes or appointment scheduling not only interrupted workflow but also risked customer dissatisfaction if responses weren't prompt.

I imagined a seamless, intelligent assistant capable of instantly handling these interactions—providing accurate, professional quotes and scheduling services with minimal human intervention. shopSenseAI was born from this necessity, offering a scalable, responsive, AI-powered solution tailored specifically to automotive repair shops.

📚 What I Learned

• Direct API Integration: Transitioning from no-code tools to directly hosting my own endpoints expanded my understanding of backend API integrations, webhook handling, and robust data parsing.
• AI Prompt Engineering: Crafting precise, consistent prompts for GPT-4o was crucial. Small adjustments in prompt structure significantly impacted response quality, necessitating extensive experimentation.
• Full-Stack Application Development: Leveraging bolt.new, I streamlined the coding process dramatically. Bolt's intuitive platform accelerated the development of the endpoints and provided immediate deployment feedback, significantly improving the workflow efficiency.

🛠 How I Built shopSenseAI

My refined, streamlined tech stack replaced the initial reliance on Zapier with a more robust, scalable approach:

• OpenPhone Webhooks: Directly trigger my custom endpoints upon receiving customer SMS messages.
• bolt.new Platform: Enabled me to rapidly build, debug, and deploy secure, high-performance server endpoints. Bolt’s integrated development environment simplified testing and iterating the webhook handlers.
• OpenAI GPT-4o API: Powers intelligent quoting, intent recognition, and response generation. Carefully crafted prompts yield accurate and contextually appropriate customer replies.
• OpenPhone API: Directly sends responses back to customers seamlessly and reliably, maintaining instant communication without third-party delays.

Using bolt.new was central to this project's success, providing powerful yet accessible tools to build and iterate quickly, enhancing my coding workflow immensely compared to previous no-code/low-code solutions.

🧗‍♂️ Challenges Faced

• Direct API Development: Transitioning from Zapier automation to direct webhook handling presented new learning curves around HTTP request handling, error catching, and data security.
• Prompt Optimization: Extensive trial-and-error was required to perfect GPT response consistency and accuracy. Ensuring stable parsing and response formats proved challenging but rewarding.
• Billing and API Quotas: Navigating OpenAI's billing system and API quotas presented unexpected hurdles, requiring creative solutions like enabling API data sharing to access additional credits.
• International Payment Issues: Being based in Canada meant dealing with unexpected billing challenges, requiring alternate approaches for service activation and API access.

🥇 Achievements and Outcomes

Today, shopSenseAI provides a robust, efficient, and professional service, dramatically reducing the manual labor involved in customer communications. It handles incoming inquiries autonomously, delivers consistent, accurate estimates instantly, and schedules appointments seamlessly. Integrating bolt.new's streamlined development environment allowed for quicker iterations, resulting in a more reliable and responsive solution.

This project highlights the transformative power of combining AI, direct webhook integration, and modern development platforms like bolt.new to automate and enhance customer service operations for automotive businesses.

Built With

• bolt.new
• git
• github
• javascript
• json
• node.js
• openai-gpt-4o-api
• openphone-api
• openphone-webhooks
• render