Route Optimization MVP Built in 21 Days, Seed Round Closed in 60

The Situation

the client was built around a clear insight: mid-market logistics companies spend 20–35% more on fuel and driver time than necessary because their route planning is done manually in spreadsheets. The founders had 15 years of combined logistics operations experience and had already lined up three pilot customers. What they didn't have was a product.

Their Series A fundraising timeline was fixed — they had a window of investor interest they needed to capitalize on. They approached Kovil AI with a 40-page spec document and a 4-week deadline.

The Challenge

Route optimization is a technically interesting problem. The naive approach — just using Google Maps for each stop — doesn't actually optimize anything. Real vehicle routing involves:

• Solving variants of the Vehicle Routing Problem (VRP), which is NP-hard
• Handling time windows, vehicle capacity constraints, and driver shift limits
• Presenting results in a way that field dispatchers can actually use — not just an algorithm output
• Integration with real fleet data (vehicle types, depot locations, stop lists)

The MVP also had to look credible to investors — not a prototype, but something that felt like a real product a dispatcher would use every day.

Our Approach

We started with a scoping call where we cut the feature list in half. A lot of the spec document was "nice to have" — we identified the 20% of features that would deliver 80% of the demo value and investor confidence. Everything else went to the backlog.

For the optimization engine, we used Google's OR-Tools — the same vehicle routing library used by Google's own logistics products. It handles VRP variants efficiently and is well-maintained. This let us skip writing optimization algorithms from scratch and focus on product quality instead.

We used Mapbox GL for the interactive route visualization — it supports custom overlays and smooth animations that made the demo genuinely impressive in the room.

The Solution

The MVP delivered in 21 days included:

• Stop management: Upload stops via CSV or enter manually; geocoding handled automatically
• Fleet configuration: Define vehicles by type, capacity, depot location, and shift windows
• Route optimization: OR-Tools VRP solver with configurable constraints; typical solve time under 8 seconds for 50-stop routes
• Interactive map view: Color-coded routes per vehicle, drag-to-reorder stops, time window visualization
• Dispatch export: PDF turn-by-turn route cards per driver, CSV export for existing TMS systems
• Comparison view: Side-by-side comparison of optimized vs. current manual routes — the feature that made every demo land

Results

the client used the MVP in five investor meetings. Three cited the demo as the primary reason for investing. The seed round — $2M at a $9M pre-money valuation — closed 60 days after launch. Within 90 days, they had signed four enterprise pilot agreements, including one with a regional grocery distribution company operating 140 vehicles.

The route comparison feature showed an average 34% efficiency improvement over manual planning in real pilot data — a number that went into every subsequent pitch deck.