New Delhi: Till quite recently, Venkat (who goes by a single name ), an operator at Flipkart’s sorting centre in Soukya, Bengaluru, would sort out and assign all packages manually. The task was challenging and time-consuming. The deployment of 100 automated guided vehicles (AGVs) at the centre last month has allowed him to do a lot more with ease. The AGVs can sort and assign packages to the pin codes of the customer just from the encoded information on the package. Flipkart claims it can process up to 4,500 shipments in 60 minutes and, when required, scale up to five times with slight adjustments. There is an algorithm that guides the AGVs on how and what to sort.

“Supply chain for e-commerce is tight. Sorting plays a very critical role from the speed as well as liability perspective, because we have sellers all over the country. Robotics can help us increase the throughput and as it is modular, we can deploy more bots if we need higher throughput during sales days," says Pranav Saxena, head of robotics and automation at Flipkart Internet Pvt. Ltd.

However, shipment sorting is just a part of the larger supply chain wheel. For instance, to quickly estimate the delivery time of a product to a customer, Flipkart has deployed a “promise engine", which can calculate an accurate date and time of delivery after factoring in elements such as customer location, seller’s inventory location and the type of product.

Once this is done, the firm has to ensure that the order reaches the customer on time and for that they rely on an Uber-level route optimization network. It allows Flipkart to address factors such as how the product will reach from one city to another by a given day, what if there is a disruption, and what if the product reaches late at night. It ensures that the shipment reaches on time and is delivered within the promised date, adds Saxena.

Sorting addresses is another challenge in e-commerce as customers often give incomplete information. Flipkart’s address intelligence system, which has been built over a period of time using machine learning, can predict the address more accurately. “Once the address is sorted, we run a route planner to optimize the route. The key problem it solves is that how many shipments can a given fulfilment executive (delivery person) handle on the same route and reach out to more customers in the same time," notes Saxena.

Flipkart also has a mobile app for its fulfilment executives, which keeps them informed about all processes.

Saxena feels technology-based solutions are very important from both an infrastructure and consumer point of view as they help build trust and allow companies like Flipkart to offer multiple things at scale.

Close