FleetMind was inspired by a very simple observation: trucking is one of the most important industries in the economy, but many dispatch decisions are still made manually through phone calls, memory, and spreadsheets. A dispatcher often has to call multiple drivers, check Hours of Service, compare locations, and make a decision under pressure. At the same time, a fleet owner can see costs going up, but may not have a clear explanation for why it is happening or what to do next.
When we looked at the hackathon problem, it became clear that the real gap is not missing data. The real gap is missing decision support. That is what led us to build FleetMind. We wanted to create something that does not just display information, but actually helps people make better operational decisions in real time.
We built FleetMind as an AI-powered fleet operations platform that feels like the missing intelligence layer for Trucker Path COMMAND. The project starts with a single login page where users choose whether they are logging in as a dispatcher/admin or as a driver. From there, each role gets a workspace designed for their part of the workflow.
On the dispatcher side, we built a live dashboard that shows incoming loads, driver recommendations, task updates, and a running activity log. The goal was to help answer the most important question in dispatch: who should take this load right now? Instead of relying on guesswork, the system helps the dispatcher make a faster and more informed decision.
Behind that dashboard, we built the backend automation flow. The dispatcher enters the source, destination, and shipment delicacy, and the system filters and ranks drivers based on factors like HOS left, driver rating, reliability, delivery history, and trip feasibility. Once the best candidates are identified, the system can automatically contact drivers one by one using AI-assisted calling. As soon as one driver accepts, the trip is assigned automatically and their remaining HOS is updated. Every step is logged so the whole process stays visible and traceable.
We also wanted the workflow to continue after the load is assigned, so we built a driver-side experience that acts like an AI co-driver. This includes route-aware support, HOS-aware decision help, fuel and rest-stop intelligence, nearby assistance, emergency tools, and a simpler driver workflow. That made the project feel more complete because it supports both the dispatcher making the decision and the driver carrying it out.
Technically, we built the project as a combination of frontend experience, backend decision logic, and automation. On the frontend, we designed a dark command-center-style interface inspired by the kind of product Trucker Path could realistically use. On the backend, we created a ranking engine that first applies hard constraints, like whether a driver has enough HOS to complete a trip, and then scores eligible drivers based on overall fit. We also integrated Twilio and ElevenLabs to reduce the manual burden of calling drivers one by one and turn that into a smarter dispatch workflow.
One of the biggest things we learned is that logistics problems are really decision problems. At first, it seems like the challenge is to build a dashboard, but once we got deeper into it, we realized the real value comes from helping users decide what to do next. We also learned that a strong operations product has to work across multiple roles. It is not enough to build only for the dispatcher or only for the driver. The real value comes from connecting decision-making, automation, and execution into one system.
Another important thing we learned is that AI works best here as an assistant, not as a black box. In logistics, trust matters. People still want visibility and control. So instead of trying to replace the dispatcher, we built FleetMind as a copilot that reduces manual work, explains recommendations, and keeps the workflow structured and understandable.
One of the biggest challenges we faced was balancing ambition with a hackathon timeline. We were trying to connect the dispatcher dashboard, backend automation, and driver-side portal into one complete flow, and that meant we had to constantly decide what was most important to build first. Another challenge was making the system feel dynamic instead of static. We wanted it to feel like a real operational system, with changing tasks, updated recommendations, and logs that reflect what the algorithm is doing.
We also spent a lot of time thinking about how to translate real-world dispatch tradeoffs into a believable ranking system. In reality, the closest driver is not always the best driver. A driver may be nearby but low on HOS, or reliable but too far away, or available but not ideal for a fragile shipment. Turning those messy tradeoffs into something that feels useful and realistic was one of the most interesting parts of the project.
A final challenge was building in a way that stayed honest about integration. We aligned our workflows closely with Trucker Path concepts like routing, fuel planning, preferred lanes, and deep-link support, but we did not pretend to have access to internal systems that are not publicly available. That pushed us to build something realistic, integration-friendly, and honest rather than something flashy but misleading.
We think FleetMind matters because it addresses a very real gap: trucking already has a lot of operational data, but not enough intelligence that helps people act on it. Our project shows what happens when you connect smarter dispatch decisions, less manual outreach, better driver support, and clearer operational reasoning into one workflow. Instead of just building another dashboard, we tried to show what fleet operations could look like if the system could actually help people decide, explain, and act.
Log in or sign up for Devpost to join the conversation.