The global logistics industry has spent decades relying on fragmented software stacks, spreadsheets, and human intuition to move goods across supply chains. But as geopolitical disruptions, labour shortages, and rising operational costs force enterprises to rethink how logistics works, a new category of AI-first platforms is beginning to emerge.One of the companies positioning itself within that shift is Shipsy, which this week said it has crossed $25 million in annual recurring revenue (ARR), more than doubling its business over the past year. While the revenue milestone signals growth, the larger story is about how logistics software itself is evolving from systems that merely record operations to platforms that actively make decisions.Speaking to ET, Shipsy co-founder and CEO Soham Chokshi described the transition as a structural change in enterprise software. “Legacy vendors built systems to record and track. Customers now want platforms built to act, in their industry, with knowledge of how their operations actually run,” Chokshi said.That transition has accelerated as logistics networks face increasingly unpredictable conditions. From the COVID-19 pandemic and the Suez Canal disruption to the Ukraine war and recent tensions around the Strait of Hormuz, supply chains have been forced into a constant state of adaptation.According to Chokshi, many logistics organisations still operate with more than a dozen disconnected systems spanning ERP modules, spreadsheets, ticketing systems, and operational dashboards. The result is a heavy dependence on manual interventions and reactive decision-making.“Before COVID, things were running the same way for years. People were taking decisions largely on intuition,” he said. “But once disruptions started happening globally, companies had to become data-driven very quickly.”That urgency has also widened the scope of where AI can be deployed inside logistics companies. Earlier, software buying decisions often revolved around feature checklists. Now, enterprises are increasingly focused on measurable operational outcomes.Shipsy recently launched what it calls “AgentFleet”, a collection of AI agents designed to automate operational tasks across customer support, dispatch operations, invoice validation, dispute resolution, and logistics coordination. Chokshi argues that the rise of AI agents is also blurring the boundaries between traditional enterprise software categories.“In logistics, solving a support problem vertically now makes more sense than using a generic horizontal platform,” he said. “A logistics technology company can solve logistics support better because the intelligence is tied to operational context.”The company demonstrated several operational AI use cases during the interaction, including an address intelligence system designed to reduce last-mile delivery failures. In many regions, particularly outside highly standardised Western markets, incomplete or inconsistent addresses remain a major source of inefficiency.“In parts of the Middle East, customers sometimes just share a landmark or a phone number and expect the delivery company to figure it out,” Chokshi said. “That creates repeated calls, delivery failures, and significant productivity loss.”The AI system attempts to clean, validate, and enrich address data using historical delivery memory and operational context, reducing the need for repeated manual intervention by delivery personnel.But while AI agents are becoming more capable, Chokshi stressed that enterprises often underestimate the operational discipline required to make them work reliably at scale.“AI is fundamentally personalised,” he said. “You have to onboard it like an employee. If your organisation itself lacks clarity in processes or SOPs, the AI will reflect that confusion.”Shipsy says this became evident during deployments with enterprise customers, where AI systems analysing support tickets and operational workflows identified large gaps between documented standard operating procedures and actual on-ground behaviour.In one case, the company found that operational issues expected to be resolved within hours were in practice taking nearly a full day, leading to higher detention costs for logistics operators.The company has also started building “supervisor agents” that monitor other AI agents in real time to reduce hallucinations and enforce operational guardrails. According to Chokshi, such oversight systems are becoming essential as enterprises move from experimentation to mission-critical AI deployments.The broader market environment is also influencing demand. Shipsy says Europe, the UK, Australia, and the Middle East are currently among its fastest-growing regions, partly due to labour shortages and rising logistics costs. The company now counts enterprises such as Coca-Cola, Heineken, Aramex, Nippon Express, and Decathlon among its customers.The disruptions are not limited to physical logistics operations alone. Chokshi also pointed to recent infrastructure incidents, including outages affecting cloud deployments in the Middle East, as a reminder that resilience and disaster recovery are becoming central priorities for enterprise customers.“Now everybody is actively thinking about disaster recovery and contingency planning,” he said. “One minute of disruption in logistics is still a disruption.”That changing mindset is increasingly benefiting technology vendors focused on operational optimisation. As enterprises attempt to prepare for future disruptions rather than merely react to them, AI systems capable of monitoring risks, triggering workflows, and coordinating operational responses are beginning to move from experimentation into core infrastructure.For logistics, an industry historically associated more with physical movement than software innovation, that may represent one of the most significant technology shifts in decades.