In this exclusive conversation with CIO&Leader, Satej Revankar, CIO at FIAT India Automobiles shares a grounded view of what it truly takes to move AI from pilot to production at scale. From redefining success metrics and enabling infrastructure to fostering cultural change and aligning AI with business KPIs, he outlines a practical roadmap for enterprises navigating the complex journey of AI adoption.
CIO&Leader: How do you define success when transitioning an AI initiative from pilot to production?
Revankar: Success from pilot to production can be assessed if intended transformation is visible & measurable, if atleast 50% of the outcomes are achieved in first 3 -6 months post go live because AI models need time to learn & unlearn based on patterns, if the end customers stop the manual option replaced by AI and if more demands on the same or similar AI use case start coming from the business within first 3 months of any AI deployment.
CIO&Leader: What are the core pillars of your current enterprise AI strategy?
Revankar: Technology , Partners, data driven AI culture, internal Skills are core pillars of AI strategy.
CIO&Leader: What key AI use cases have successfully moved into production, and what measurable impact have they delivered?
Revankar: We have deployed vision inspection for cars at the final rollout stage of our assembly production line. The use case deploys AI platform to check online exterior & interior images and detects the feature anomalies and gives inspection results. The deployment has improved accuracy of inspection process , eliminated the human errors in manual inspection process, improved the cycle time of car rollout and reduced the direct labour cost.
CIO&Leader: What infrastructure or architectural changes were necessary to scale AI effectively within your organization?
Revankar: AI applications are data hungry & latency dependent. Hence very strong internal network equivalent to 5G from machine level to gateway is necessary. Alongwith it, strong OT layer security is essential when AI deployment goes to Shopfloor. In addition, the strong compute & cloud adoption is essential as we scale up and infra scale up is required as we grow.
CIO&Leader: What are the biggest challenges you’ve faced in operationalizing AI, and how have you addressed them?
Revankar: Biggest Challenges faced so far are managing business expectations on ROI and adoption across business team in accepting the decision & execution triggered by AI deployment. Managing Patience in AI maturity cycle is also significant challenge..The AI use case needs to be mapped to measurable benefits and to be mapped to business KPIs. This helps in structuring the ROI visibility. The deployment also needs to be broken into small parts each having some visible outcome. This helps in managing the expectations of end users & business teams. Also the AI learning phase timelines & errors can be adjusted before the next step to be resumed.
CIO&Leader: How are you preparing your workforce for scaled AI adoption, and what organizational shifts have been required?
Revankar: We are encouraging business teams to explore ChatGPT for their business challenges and are reviewing their proposals & project outcomes to identify the business scope & team members to be supported with skills & budgets. Organisation need to allocate seperate AI budgets and training budgets to support the AI roadmap. The assessment of AI deployments also need to be subjective rather than purely ROI based. The leadership teams need to assess the overall acceptance and cultural shift in initial 12 months .
CIO&Leader: Looking ahead, what does your AI roadmap over the next 12–18 months look like — especially in terms of GenAI or foundation model deployments?
Revankar: We are planning to explore more GenAI deployments in various manual inspection processes for engine assembly & car assembly and some safety prevention use cases as well as predictive maintenance of critical equipment. The overall focus will be to cover most of the business teams and facilitate enterprise wide AI culture and foster cross functional AI inter-dependent workforce.