Meet Kamlesh, a seasoned expert in the grocery supply chain. With over 12 years of experience, he has played a key role in scaling Flipkart Grocery and Swiggy Grocery and now serves as the Vice President of Supply Chain at Jumbotail.
In this insightful conversation, Kamlesh dives deep into the challenges of the grocery supply chain, how AI is transforming everyday operations, and his advice for aspiring supply chain professionals.
You’ve worked in grocery supply chains for most of your career. What are some common challenges?
The grocery supply chain comes with extensive compliance requirements, far more than other supply chains, since it directly impacts lakhs of people who consume these products. So as an ethical organization, we have to be responsible to ensure we are meeting all the compliance requirements.
Another major challenge is demand forecasting, While there are techniques available for forecasting demand, poor accuracy can lead to wastage, especially since grocery items come with expiry dates. Also, margins are limited, so the scope of error goes down.
Additionally, the grocery supply chain operates at an immense scale, with lakhs of products moving daily. Since it is a fast moving supply chain, implementing changes in the system while ensuring all stakeholders are informed and aligned remains a significant challenge.
You were involved in the early phase of what is now Swiggy Instamart and optimized the supply chain to achieve profitability at a unit economics level. How did you do it?
I was instrumental in building and scaling Suprdaily, which was the first grocery platform in Swiggy’s ecosystem. It initially operated in the micro-delivery space—customers placed orders before 11 PM for next-day morning delivery before 7 AM. Later, I also worked closely with the Swiggy Stores team, which eventually evolved into Swiggy Instamart.
At Suprdaily, we built and scaled grocery delivery using an asset-light model for last-mile delivery. For example, in Bangalore, we operated with a central warehouse, four micro hubs, and over 100 delivery points (open spaces near apartments, petrol pumps, etc.). The key to optimization was identifying the right demand clusters to reduce last-mile delivery costs. Every morning, groceries were transported from hubs to delivery points within each cluster, such as an empty petrol bunk or an apartment gate. Delivery personnel would then pick up the orders and distribute them to other parts of the cluster. Since deliveries took place in the early morning, we benefited from traffic-free roads and the availability of free spaces like petrol bunks to station the goods. The cost structure was really low and the model was profitable at lower order value.
We later scaled this model to other urban cities, reaching approximately 2.5 lakh orders per day.
What challenges did you face when scaling to other cities?
Finding suitable facilities – Warehouses and hubs had to meet compliance standards.
Identifying the right demand clusters – Areas with high order volumes had to be mapped correctly.
Regulatory approvals – Since we dealt with food, obtaining licenses from authorities was critical.
Vendor selection – Choosing vendors willing to adopt technology and scale with us.
Are you using AI in supply chain operations? How is it solving problems?
We use AI for multiple use cases. One is route optimization for the milk run delivery. For example, if a delivery person is handling 30-40 stores a day, the system not only optimizes routes but also considers customer behavior, such as the time taken at each delivery stop to predict and prepare the most optimized route. Based on this data, we ensure deliveries are made within the promised time, allowing customers to be communicated and be prepared to avoid any delays at the customer door step. We maintain the slot promise with up to 95% accuracy today. AI models also predict delays, which are then communicated to the internal stakeholders in real time to act quickly.
Another major application is in HR. Blue-collar worker attrition is a serious challenge, as productivity is lost during the time it takes to train new employees. Our AI model predicts employees who are likely to quit and alerts HR, enabling them to proactively engage with these employees and reduce attrition.
India’s fragmented logistics ecosystem has been a biggest challenge we are trying to solve at NSCS. How are you addressing this?
This is one of the biggest challenges in our industry. In Tier 1 cities, there are organized players, but as we move to Tier 2 and Tier 3, finding organized players becomes more difficult.
On the transportation side, we have developed systems that ensure transparency in compliance and payouts. The system is built in detail to capture data, including the payout given to drivers versus what was committed, allowing us to manage operations professionally.
During the vendor selection process, we prioritize vendors who are open to learning new technologies rather than those who insist on sticking to traditional methods.
What is your advice to operational managers?
AI is now a reality. Understand its various use cases across industries and ideate on how you can leverage AI/ML to solve daily supply chain challenges.
Supply chain is known for its unpredictability and stress. How do you manage it?
Supply chain operations come with constant uncertainties and firefighting, making it a stressful field. However, technology brings control and predictability, allowing for better communication in case of delays.
Despite the challenges, I always remind my team of the impact we create for customers. Unlike other departments, we interact with customers directly and witness the smile we bring with each delivery. That satisfaction keeps us motivated through the difficulties the industry presents.
