Malayali firm which won Google Maps innovator award also helps US govt in crowd control

When you send your mother flowers or a cake for her birthday, you want them to arrive exactly on time — not a day later. Delivery delays and rescheduling are a hassle for both customers and businesses in today’s last-mile delivery landscape. Dispatch AI, an artificial intelligence-powered assistant developed by Torcue Digital, recently won First Place Gold in the ‘Best of AI’ category at Google Maps Platform Awards, for solving these problems.

Founded by Labby George in Florida in 2016, Torcue Digital has its software development centre in Kochi.

“When we did a market study, we found that out of every ten deliveries, only six were successful,” says Labby George. “The US was our benchmark because roads and Google Maps there are well-defined. Businesses lose heavily from failed deliveries — not just customers, but money. For example, for a $10 item, companies lose around $15–20 when they have to redeliver or return it. This prompted us to find a solution.”

Initially, the team assumed that traffic or weather caused most delivery failures. But they soon realised that 60% of failures were actually due to address-related errors and poor customer-driver interactions.

“The AI acts as a personal assistant for the driver to solve these issues— like a voice or text-based guide,” explains Labby. “It analyses customer data and advises the driver. For instance, ‘This customer has been ordering for three years; use Gate A, not Gate B.’”

Another major problem Dispatch AI tackles is rescheduling deliveries. “Currently, customers have to call Amazon or D2C companies to request changes, which takes 10–15 minutes. We built an agent AI system that automates this entire process,” she says.

Dispatch AI won the prestigious award after competing against more than 400 global submissions including leading universities such as UC Davis and the University of Texas, as well as major geospatial companies. “Each team solved different problems using Google Maps,” Labby notes. “Our edge was that we approached it from the customer’s perspective, not just the technology side.”

The cutting-edge technology was developed by a 20-member team in Kochi. Torcue’s developers were initially based in Bengaluru, but the company shifted its base after COVID-19.

“Bengaluru had high attrition and a toxic work culture,” says Labby. “We formally established the Kerala office in 2024. We focus on building a strong, passionate team, hiring local college graduates and training them rigorously in problem-solving, not just coding. Within six months, our developers reach the skill level of someone with two to three years of experience in a major IT firm.”

The system comprises three components:

• Customer agent: Allows users to say, “Hey AI, can you reschedule my delivery?” It reads their profile — whether they’re at work or travelling — and adjusts accordingly.
  
• Driver agent: Provides real-time updates, alerting drivers instantly if a package has been rescheduled.
  
• Warehouse agent: Manages shift schedules and truck loading automatically, matching deliveries to routes, vehicle capacity, and driver preferences.
  

“All this reduces human effort drastically — what once took hours can now be done automatically,” says Labby.

Dispatch AI is expected to hit the market soon, with pilot rollouts planned for next year. Apart from Dispatch AI, Torcue has built ERP and POS systems, IoT-based patient inflow systems for the Cleveland Clinic, AI-driven vendor bidding platforms for New York County, and disaster management and crowd control software for the US federal government.

While AI has prompted layoffs at several large firms, Labby believes smaller companies can adapt faster. “There’s no rigidity here,” she says. “Unlike earlier, customers are now two steps ahead of us, and we’re just catching up. It’s essential to stay open to ideas and adapt quickly.

“There’s so much noise in AI today — hundreds of tools every day. But not everything is meaningful. The trick is to use AI only if needed. Don’t use AI if you can solve the problem with a pen and paper.”