In this exclusive conversation with Rajeev Ranjan, Editor, Digital Terminal, Venkat Moganty, Co-founder and CTO of Enmovil, explores how AI-driven orchestration is transforming supply chain planning and execution across Indian enterprises.
Rajeev: How are Indian enterprises approaching AI adoption in supply chains, and what are the biggest barriers preventing them from translating AI investments into measurable business outcomes?
Venkat: Indian enterprises are approaching supply chain AI with a practical question: which operational decisions can we improve? Priorities include demand forecasting, inventory availability, vehicle utilisation, delivery reliability and logistics costs. The next step is connecting these individual capabilities to the decisions teams make every day.
The biggest barrier is the gap between generating a recommendation and achieving an operational outcome. A better forecast creates limited value if procurement, production and dispatch continue working with different assumptions. A shipment alert helps only when teams can assess its impact, evaluate alternatives and implement a response.
Fragmented data, inconsistent master records and limited integration between enterprise systems make this difficult. Equally significant are unclear business ownership and pilots measured primarily on model accuracy rather than improvements in cost, service or productivity. Users also need confidence that recommendations reflect operational constraints and that they can question or override them.
Enterprises should begin with a recurring business decision, establish a baseline and assign an accountable owner. They should then connect the necessary information, validate recommendations with operational users and embed the process into daily work.
Success should be measured through indicators such as freight cost per unit, inventory days, service levels and planning time. Sustainable value emerges when AI improves how decisions are made and followed through, with people responsible for the resulting business outcomes.
Rajeev: How can companies introduce AI-driven decision-making across planning, logistics and warehousing without replacing their existing ERP, TMS and WMS infrastructure?
Venkat: Companies can introduce AI through an intelligence and orchestration layer that connects their existing Enterprise Resource Planning (ERP), Transportation Management System (TMS) and Warehouse Management System (WMS). These applications continue managing transactions and operational records, while AI helps teams coordinate decisions across them.
Consider a sudden increase in customer demand. Responding requires teams to assess inventory, material availability, production capacity, warehouse throughput and transport options. An orchestration layer connects these capabilities through approved interfaces, allowing information to flow between systems and authorised changes to be submitted.
Connections can use Application Programming Interfaces (APIs) and Model Context Protocol (MCP), a standard through which AI applications access tools and information from connected systems. Configurable agents, workflows and approval gates coordinate these interactions, with permissions governing which capabilities users and agents may invoke.
The division of responsibility matters. Language models interpret requests and coordinate tasks; forecasting, optimisation and business services provide calculations and check operational constraints. Proposed changes to enterprise records should pass through the appropriate validation and approval steps.
Implementation should begin with a focused use case, dependable data and clearly defined ownership. Companies can introduce visibility and recommendations first, then enable approved actions as the process is validated.
The objective is to increase the value of existing infrastructure by connecting information, decisions and execution steps that currently require manual coordination.
Rajeev: As enterprise AI evolves from copilots to agentic systems, how can organisations ensure AI agents execute supply chain decisions reliably while maintaining human oversight and accountability?
Venkat: As AI moves from assisting people to executing tasks, organisations need explicit boundaries around its authority. They must define what an agent may do, whose permissions it uses and which decisions require human approval. Reliability depends on governance, validation and operational supervision throughout the process.
Agents should operate with assigned tools, access permissions and execution limits. For consequential changes, proposed actions should be held for review. Approval should be bound to the stored action and its specific parameters, so execution proceeds against what the person reviewed. Changes to that proposal should require renewed approval.
Duplicate-action controls help reduce the risk of repeated transactions when a workflow retries or resumes. These controls need to work alongside checks on business conditions, confirmation from the receiving system and procedures for handling incomplete or uncertain outcomes.
Human oversight must remain practical. Approvers need supporting evidence, proposed changes and relevant business impact. Named owners should be able to reject proposals, request changes, pause workflows or stop execution when circumstances require intervention.
Audit records should capture the request, tool activity, approvals and execution results, making it possible to investigate what happened and who authorised it.
These measures manage risk; they do not eliminate errors. My view is that autonomy should expand only after behaviour has been validated under realistic conditions, with human accountability, exception handling and recovery procedures retained throughout.
Rajeev: With strong adoption across India’s automotive, manufacturing and FMCG sectors, which supply chain challenges are creating the greatest demand for AI-led orchestration solutions?
Venkat: The greatest demand comes from decisions that span departments and systems. Enterprises may have capable planning, transport and warehouse applications, yet still depend on calls, spreadsheets and individual experience to coordinate a response when conditions change.
In automotive, the priority is synchronising component availability, production requirements and outbound movement. A delayed inbound shipment can affect manufacturing and customer commitments. Teams need to understand the consequences quickly and compare feasible responses, including alternate stock, revised schedules or expedited transportation.
In manufacturing, material availability, production capacity, storage and dispatch must be considered together. Increasing output at one plant may look efficient until it creates warehouse congestion or higher freight costs. Companies need decisions that account for the wider operation.
In fast-moving consumer goods, demand variability, promotions, regional consumption and shelf-life constraints put pressure on forecasting and replenishment. Excess stock in one location can coexist with shortages elsewhere, requiring coordination between inventory allocation, supply and transport.
Across these sectors, truck utilisation, load consolidation, detention, delivery exceptions, proof of delivery and freight settlement also create demand for connected workflows.
The opportunity is to coordinate these activities around shared business priorities. Governed workflows can connect analysis, approvals and actions across existing systems, with people retaining responsibility for consequential choices. The practical goal is to shorten the time between detecting a problem, deciding what to do and carrying out an approved response.
Rajeev: What will it take for Indian supply chain technology companies to build globally competitive AI platforms, and where do you see the biggest growth opportunities for Enmovil over the next three to five years?
Venkat: Indian supply chain technology companies have an opportunity to build globally competitive AI platforms because they work with complex networks, diverse operating conditions and demanding cost constraints. The challenge is to turn that experience into products that deliver consistent value across markets.
From my perspective, competitiveness rests on four foundations: domain expertise, interoperability, dependable engineering and measurable business outcomes. Platforms must accommodate different enterprise systems, operating rules and data requirements without extensive redevelopment for every customer. Security, governance and implementation quality must be designed into the architecture.
Technology choices also need discipline. Forecasting, mathematical optimisation, simulation and language models serve different purposes. The advantage comes from combining them effectively around a business decision, with evidence supporting recommendations and appropriate controls governing execution.
For Enmovil, I see three major growth opportunities over the next three to five years. First, connecting planning and execution across demand, inventory, production, warehousing and transport. Second, helping enterprises manage disruptions through coordinated decisions across departments. Third, expanding internationally through repeatable industry solutions, enterprise integrations and strong implementation partnerships.
A further opportunity is making advanced decision capabilities accessible to operational teams without requiring them to become AI specialists.
Ultimately, global growth will depend on demonstrated customer value: better service, improved asset utilisation, faster decisions and lower operating costs. Companies that can reproduce those outcomes across customers and geographies will earn lasting trust.
𝐒𝐭𝐚𝐲 𝐢𝐧𝐟𝐨𝐫𝐦𝐞𝐝 𝐰𝐢𝐭𝐡 𝐨𝐮𝐫 𝐥𝐚𝐭𝐞𝐬𝐭 𝐮𝐩𝐝𝐚𝐭𝐞𝐬 𝐛𝐲 𝐣𝐨𝐢𝐧𝐢𝐧𝐠 𝐭𝐡𝐞 WhatsApp Channel now! 👈📲
𝑭𝒐𝒍𝒍𝒐𝒘 𝑶𝒖𝒓 𝑺𝒐𝒄𝒊𝒂𝒍 𝑴𝒆𝒅𝒊𝒂 𝑷𝒂𝒈𝒆𝐬 👉 Facebook, LinkedIn, Twitter, Instagram