In conversation with CIO&Leader, Anand Deodhar, Group CIO at Force Motors shares how IT needs to transition AI from experimental pilots to enterprise-grade solutions that deliver real business impact. From predictive maintenance to GenAI-powered assistants, Anand Deodhar breaks down the strategy, architecture, and workforce transformation that helps embed AI into the core of many manufacturing operations.
CIO&Leader: How do you define success when transitioning an AI initiative from pilot to production?
Anand Deodhar: In manufacturing, success isn’t defined by shiny dashboards or isolated pilots. It’s when AI genuinely improves plant performance, be it uptime, efficiency, or bottom-line savings. We need to look at hard KPIs like OEE improvements, reduction in downtime, or process optimization that holds up under real-world pressure.
For example : predictive maintenance model . It takes approximately two months of rigorous piloting—including stress testing and failure scenario analysis; before we can deploy at plants and this also helps any org to cut unplanned downtime by 15%. For me, success is when the shop floor, IT, and leadership teams all see value and start relying on AI as part of their daily decision-making—not as a side experiment but as core to how many org should operate.
CIO&Leader: What are the core pillars of manufacturing enterprise AI strategy?
Anand Deodhar: AI strategy is grounded in four solid pillars—built from lessons learnt on the shop floor and in the boardroom alike:
- Business-Aligned Use Cases: We need to start with what matters to the business—targeting bottlenecks with high impact and low resistance.
- Data Governance & Security: With GenAI, ensuring data privacy and compliance isn’t optional—it’s foundational for me
- Scalable Infrastructure: A hybrid cloud-edge model ensures that we can scale AI with both speed and efficiency, without disrupting legacy operations.
- Skill Enablement: AI should be made as learning priority not just for IT, but for operations, quality, and procurement teams too.
To me These aren’t just strategy buzzwords; they’re working principles that help many org , scale AI without losing sight of business value.
CIO&Leader: What are the key AI use cases that can be successfully moved into production, and what measurable impact they can deliver ?
Anand Deodhar: AI is no longer a pilot exercise in setup; it’s already implemented in multiple cases, across many organisations. Some of the key use cases now embedded in production across orgs.
- Predictive Maintenance : which brings down maintenance costs and improved uptime as well
- Machine Vision for Defect Detection : Helps to cut down rejections significantly
- AI-Driven Demand Forecasting: Improves procurement planning and helps to reduce inventory by freeing up working capital.
Each project should normally start with a business challenge, not a tech pitch. That’s what ensures real-world impact.
CIO&Leader: What infrastructure or architectural changes are necessary to scale AI effectively with many auto manufacturing organisations?
Anand Deodhar: To scale AI sustainably, manufacturing companies need to rethink their IT backbone—because models alone don’t deliver results. The focus should be on
- Hybrid Cloud & Edge Stack: For seamless model training and fast real-time inference.
- Kubernetes & Containerisation: Ensuring models can run across diverse environments smoothly.
- Modular AI Libraries: Reducing time-to-market by reusing proven components.
- Dedicated VLANs & GPU Workstations: Ensuring low-latency AI even in bandwidth-constrained areas.
These upgrades required are not just tech decisions, but they were enablers of future-ready operations for any organisation.
CIO&Leader: What are the biggest challenges for IT in operationalizing AI, and how do we need to address them?
Anand Deodhar: Every AI initiative meets roadblocks. The major ones, which are generic to manufacturing are given below
- Data Silos: Build standardised ETL pipelines with tagging so data stays clean and consistent.
- Skill Gaps: Invest in reskilling, bring in vendor support, and create cross-functional squads.
- Trust & Governance: From the start, ensure built-in explainability and compliance checks to win stakeholder confidence.
Having said this, we should never treat these as obstacles but treat them as part of the journey toward enterprise-grade AI.
CIO & Leader: How do we need to prepare the workforce for scaled AI adoption, and what organizational shifts across manufacturing are required?
Anand Deodhar: Technology can only go so far, but people make or break transformation. Some of the solutions are
- Plan 2,000+ hours of AI/ML training across functions.
- Appoint an AI champion in all plants as practitioners, who will bridge IT and operations.
- Make AI performance reviews part of monthly meetings across various functions in any org.
This approach shifts AI from being “someone else’s project” to something everyone owns.
CIO&Leader: Looking ahead, what does the manufacturing AI roadmap over the next 12–18 months look like — especially in terms of GenAI or foundation model deployments?
Anand Deodhar: Looking ahead, most of the manufacturing AI roadmap will be focused on scaling further and tapping into GenAI:
- GenAI Assistants: To help technicians quickly access SOPs, logs, and manuals.
- Digital Twins: for real-time simulation and production line optimisation.
- GenAI in Supply Chain: To improve planning precision and reduce supplier risk.
- LLM Governance Frameworks: Ensuring responsible and secure GenAI usage.
- ERP/MES Integration via APIs: embedding AI deeper into workflows, targeting a boost in line-level productivity.