As India’s logistics ecosystem evolves into a digitally orchestrated, always-on network, transportation companies are increasingly reimagining how intelligence, automation, and real-time decision-making can redefine operational efficiency at scale.
For organisations managing thousands of daily freight movements across sprawling physical infrastructure, the challenge is no longer limited to moving parcels from one point to another. The larger imperative is building a data-driven operational fabric that can dynamically optimise routes, warehouse flows, customer interactions, workforce productivity, and service quality in real time.
At Safexpress, this transformation is unfolding through a multi-layered AI and digitalisation strategy spanning digital twins, enterprise data platforms, conversational interfaces, intelligent customer engagement, and agentic automation.
In this exclusive interaction with Express Computer, Sandeep Dewangan, President & Group Chief Information Officer, Safexpress Private Limited, discusses how the organisation is leveraging AI to orchestrate one of India’s largest logistics operations, why digital twins are becoming critical for warehouse optimisation, and how conversational and agentic systems could fundamentally reshape enterprise applications in the years ahead.
Building a fully digital logistics backbone
According to Sandeep, the company’s AI transformation journey began with a much larger digitalisation effort focused on operational visibility and standardisation.
“As you know, Safexpress is one of the leading B2B logistics and transportation providers in the country. We are talking about millions of parcels, thousands of tonnes every day. We have 14,000 plus trucks running non-stop 365 days a year and 830 plus pickup and delivery centres,” he points out.
Managing such a large physical operation requires continuous visibility into parcel movement, fleet operations, warehouses, and delivery networks in real time. “In order to get the data, we first undertook a big journey of digitalisation. So much so that we are today 100 per cent digital in our operations. We work paperless in every department and location,” he explains.
That transition laid the foundation for data-led operational decision-making across the organisation. However, he agrees that digitalisation alone was not sufficient. “We realised that in order to become far more efficient, far more responsive to our customers, and drive service execution levels much higher, we needed quality, efficiency, and speed.”
AI is becoming the operating layer of logistics execution.
Sandeep says the company approached AI as a core operational capability rather than as an isolated technology layer. “For us, it was how do we leverage all these possibilities and abilities of data, AI, and automation together and make that a pivot on which we do our business.”
Unlike many AI initiatives that remain confined to digital environments, logistics presents a unique challenge where every decision impacts physical movement of goods, vehicles, facilities, and people. That is where AI must prove its value in the real world.
At the centre is the data lake strategy — our single source of truth. We consolidated operational, financial, customer, partner, and service data streams into a unified information layer.
The company then started introducing machine learning systems across operational workflows involving employees, field teams, and partner ecosystems.
One of the earliest AI-led deployments involved digitising proof-of-delivery validation processes. Several AI-led interventions have already delivered measurable improvements in operational productivity, turnaround times, service quality, and customer responsiveness.
Sandeep believes deploying AI use cases at a national scale brings its own level of operational complexity. “Whether it was day or night, thunderstorm or rain, in Kerala or Ladakh, all these gave us a lot of learning. It taught us how complicated it can be to make one simple use case become a mainstream business process.”
Digital twins are helping optimise warehouse operations in real time.
A major focus area for Safexpress now is the development of digital twin-led warehouse orchestration systems. Sandeep explains that warehouse environments involve continuous operational decisions across truck movement, loading capacity, routing priorities, and workforce allocation.
The challenge becomes even more complex when route dependencies, delivery timelines, and worker availability are factored into the system simultaneously. “How do we optimise workload on each worker? How do we minimise queue time?”
For Sandeep, these are fundamentally mathematical optimisation problems to be applied at a massive scale.
The broader goal is creating a synchronised logistics ecosystem capable of continuously optimising itself. “End-to-end optimization is the big vision that we are working on. The outcome will reflect significantly higher efficiencies, utilisation, and sustainability factors.”
Customer engagement is becoming increasingly conversational.
As customer expectations evolve, Safexpress is also redesigning how customer interactions are handled across both retail and enterprise segments. According to Sandeep, customer engagement models vary significantly depending on the customer category.
“We have large businesses, medium-sized businesses, and another significant bank of customers, which is retail,” he says, adding that for retail and self-service workflows, conversational automation is already proving highly effective. “Digital engagement can be highly positive for our retail customers. For self-service customer support, directly engaging with an AI-driven virtual agent gives much more speed.”
Sandeep points out that automation becomes especially valuable in reducing wait times and improving responsiveness.
At the same time, agentic AI systems are also beginning to reshape enterprise-facing operational workflows internally. “Today, a single person is handling multiple customer accounts. A lot of their transactional activities will be fully taken over or replaced by agentic solutions.”
As repetitive workflows get automated, customer-facing teams will increasingly move towards higher-value responsibilities. Their value additions will become far more meaningful and helpful.
AI adoption cycles are moving faster than enterprise systems.
Sandeep also highlights the unprecedented pace at which enterprise AI ecosystems are evolving. “Six, seven, eight years ago sounds like generations behind. Even three years ago was far different from what it is now.”
For enterprises, the real challenge is no longer access to technology but operationalising continuously evolving capabilities fast enough. “The speed at which highly powerful features are coming out is super exciting. At the same time, it is puzzling on how to continuously absorb this and create enterprise-grade solutions before the next update comes,” he says.
This has forced the organisation to move away from traditional waterfall-style implementation models. Instead, Safexpress now follows an iterative rollout approach built around pilots, real-world testing, and continuous refinement.
Conversational systems becoming the future enterprise interface
Looking ahead, Sandeep believes enterprises are rapidly moving towards a future where conversational and agentic systems will become the primary interface through which users engage with enterprise applications.
Safexpress itself has already begun moving in that direction internally. “We have also decided to adopt Slack as the primary interface for everyday operations.”
According to him, over time, traditional applications may become increasingly invisible to end users, with conversational and agentic interfaces emerging as the dominant interaction layer while core systems continue to provide business logic, workflows, and data services.
At the operational level, they continue to expand AI-led orchestration capabilities across logistics operations, routing intelligence, and disruption prediction systems. A digital twin is one large program. It encompasses a lot of data science-led developments. Whether it is route optimisation, prioritisation of parcel movement, or estimating disruptions anywhere, all of these are the main themes we are working on,” he says.
As AI adoption expands, Safexpress is also investing in governance frameworks covering data quality, security, compliance, explainability, and human oversight to ensure responsible deployment at enterprise scale.
Sandeep concludes with an observation that India’s logistics sector is moving from asset-centric operations to intelligence-centric operations and that the competitive advantage will increasingly come from how effectively organisations can integrate data, automation, AI, and execution into one continuously responsive and near-autonomous operating environment.