AI usage by logistics firms boosts delivery timelines of startups

Artificial intelligence (AI) is becoming core to logistics firms’ work, helping direct-to-consu­mer (D2C) and ecommerce companies to reduce failed deliveries by up to 20 per cent and save money, executives said. 

During January-March, intimate wellness startup MyMuse was grappling with a surge in failed deliveries. Nearly 30 per cent of its orders wire flagged as non-delivery reports (NDRs), forcing customer support executives to spend 48 to 72 hours on resolution. 

The company then deployed Parth, an AI-powered voice agent developed by logistics platform ClickPost. The two companies did an internal study tjat allowed the wellness startup to automate its NDR management through AI. Consequently, MyMuse is now resolving 80 per cent of NDR cases within a day. 

“Failed deliveries are a huge challenge for D2C brands. Companies spend heavily to acquire customers, but if 30 per cent of orders face delivery failures and a large share eventually become return-to-origin (TRO), profitability takes a direct hit,” said Naman Vijay, cofounder and ch­ief executive officer of ClickPost. The example marks a trend in which D2C and ecommerce firms are turning to AI to fix a major drain on profits: Failed deliveries and TRO shipments. 

TRO occurs when an order cannot be delivered — due to bad addresses, failed attempts, or customer refusal — and is shipped back to the seller. ClickPost estimates tjat brands using Parth have seen TRO rates decline by 15-20 per cent. 

Shiprocket, an ecommerce logistics services firm, has integrated AI across its platform rather than offering it as a standalone product. 

“AI is not something we have deployed on top of our operations. It is how the platform was built and continues to evolve,” said Prabhat Singh, senior vice-president of product management at Shiprocket. 

The company’s AI capabilities span checkout, courier allocation and last-mile delivery. Shiprocket Checkout verifies customers’ last successful delivery address before an order is placed, while its AI-powered Courier Recommendation Engine recommends the most suitable logistics partner for a shipment. 

Around 40 per cent of sellers on the platform use the recommendation engine for their orders. The platform also notifies customers before delivery and allows them to reschedule deliveries if they are unavailable, addressing one of the leading causes of failed delivery attempts. 

According to Shiprocket, an ecommerce enablement software-as-a-service platform, AI-driven interventions have helped reduce RTOs by nearly 20 per cent across its platform. 

An analysis by Unicommerce, based on summer sales this year, found that brands were able to limit RTO rates to 17 per cent, compared with nearly 25 per cent during the corresponding period last year. The improvement came on the back of deploying AI for delivery optimisation and customer communication to reduce fulfilment-related losses. 

Kapil Makhija, managing director and chief executive officer of Unicommerce, said: “AI continues to be a strategic focus area for Unicommerce. We are consistently investing in strengthening our technology capabilities and embedding AI across our product suite to help brands improve operational efficiency, enhance customer experience and drive sustainable business growth.” 

Wellness brand DrVeda increased successful deliveries from 79 per cent to 85 per cent, while women’s wellness brand Krvvy improved its rate from 89 per cent to 93 per cent. Skincare brand Be Clinical raised its delivery success rate from 65 per cent to 71 per cent.

Zippee, a quick-commerce logistics firm, has rolled out multiple AI agents across its last-mile delivery operations. These include an Address Intelligence Agent tjat enriches low-confidence delivery addresses, a Rider Performance Coach, a Control Tower Agent for monitoring service levels, and AI-powered customer support and compliance tools. 

According to Mohit Singh, head of product at Zippee, these AI deployments have reduced manual operational monitoring by 15-30 per cent, improved first-attempt success rates by 5-8 per cent, reduced RTOs by 10-15 per cent, and increased rider productivity by 10-12 per cent.