From Frustration to Focus: My Journey with Arise & Shine Transporters and AI
This month, I tackled the challenge of integrating artificial intelligence into my work at Arise & Shine Transporters, starting from sheer confusion about where to begin.
The Beginning
A Frustrating Day in Thika I arrived at the office on a Monday morning and felt like an outsider watching a game I didn’t quite understand. Our sand and aggregates supply business was a jumbled mess of manual processes, from taking customer orders to managing stock levels and tracking truck performance.
"The first day wasn't pretty," I remember thinking to myself as I sifted through the crumpled paperwork scattered across my desk. "All these numbers mean nothing without digital order forms, geocoding for delivery points, automated pricing based on distance — every part of our operations needs a cohesive system."
Discovery and Ambition
The Spark That Monday marked something different inside me though. I had stumbled upon an article about how AI could transform supply chain management, sparking the first idea: what if we applied these technologies to streamline Arise & Shine's processes?
"It was as if a light bulb switched on," I admit with a smile now. "I started doing research and noticed there were solutions already out there, but implementing them would require a lot of work."
The Pivot
Migrating from OpenAI As I dove deeper into the solution landscape, one name kept coming up: OpenAI. Initially excited by their generative capabilities — article writing, report generation, analytics summaries — my initial excitement soon turned to concern.
"Running all that AI in production was going to be expensive," I realized as I reviewed our budget and cost estimates. "We needed a more affordable option."
The Pivot Point
Switching from OpenAI to Qwen With the help of some community resources, particularly someone who had previously made this switch with success, we decided it was time for a pivot.
"Porting everything over to our new LLM (Qwen) at Hetzner's wasn't as simple as I imagined," I confess. "SSH into servers, managing environment variables — the learning curve felt steep."
The Journey
Learning and Growth It was during these early days that I learned a lot about running servers remotely, SSH commands, and Linux administration.
"I started to see the value in having such control over our infrastructure," I reflect. "The abstraction provided by OpenAI had allowed us some leeway with their API calls — but now we needed something more robust."
The End of a Learning Curve
Progress and Satisfaction Even though moving from one AI provider to another was challenging, the satisfaction in seeing our system go live again, serving real users and generating value, outweighed any initial frustrations.
"I realized that being able to control every aspect of what powers your software gives you a lot more confidence," I conclude with a sense of pride. "It’s not just about building an app — it's understanding the underlying processes too."
The Future
What Lies Ahead As we move forward, we’re thinking about automating more aspects of our operations and exploring ways to improve delivery times for customers.
"The journey is far from over," I say with a sense of excitement. "But today, I feel like we’ve taken an important step in becoming better equipped for the AI-driven future."
