Apollo Tyres, in the analyst call following the Q4 FY26 results announcement, said, “We are scaling the use of Artificial Intelligence (AI) across business functions, with several initiatives underway in manufacturing, logistics, and customer service to drive efficiency and cost optimization.”
“The overall investment has been highly cost-effective. The key advantage has been the availability of extensive open-source capabilities,” said Hizmy Hassen, Chief Digital Officer, Apollo Tyres speaking with FE FUTECH.
The AI Playbook
Over the last three to five years, multiple consultancies and solutions providers were evaluated, the tyre maker ultimately chose not to adopt an off-the-shelf AI platform. Instead, it worked backwards from specific problem statements, grouped them into different categories of operational gaps, and built its own AI environment using a combination of open-source and paid technologies.
While specific numbers were not disclosed, Hassen said its “AI platform and AI solutions have been managed in an extremely economical manner.” Apollo Tyres has also focused on optimizing the use of generative AI components, GPUs, and tokens efficiently. “Most of the work has been carried out internally by a strong in-house team.”
For manufacturing, the manufacturer operates a hub in Hyderabad comprising permanent staff as well as enthusiastic interns. “As a result, both platform and operational costs have remained significantly lower compared to what many other companies are reportedly spending.
The costs on cloud operations have also been managed very efficiently. Most of the manufacturing-related AI work has been developed entirely in-house, while certain functions such as warranty management and logistics involve external partners bringing solutions to specific business problems. The company added that the overall AI spending remains a very small percentage of its total IT budget.”
In-house Built Micro-Language Model
The manufacturing journey of using AI to drive efficiency, quality improvement, and safety started nearly three years ago, when all the manufacturing machines were connected to the cloud. With these highly capital-intensive machines, they come with a lot of sensors. For example, in tyre building and compound mixing machines, there could be more than 100 sensors. They have been connected in near-instant fashion to the cloud, where the firm has built an extensive data lake over the last three to four years with a vast amount of data.
A lot of the use cases where generative AI capabilities have really been successful involve the use of machine data. For example, an in-house ‘micro-language model,’ has been designed using which, operators can use natural language to understand what is impacting production and why efficiency is being lost. The AI bot provides this information, which then enables engineers and shop-floor teams to address bottlenecks much faster.
Improvement in Production Throughput
One key success of AI in manufacturing is that it has significantly reduced the time required to identify the root cause of issues in production. As a result, they are now addressed much faster than in the past. “Therefore, periods of inefficiency have come down significantly, helping increase overall efficiency and throughput across the plant,” said Hassen.
AI is also looking across various plants and identifying which plants, cycles, compounds, or production specifications are performing better, and therefore suggesting benchmarks to operators to help close performance gaps faster.
In the last few years, the focus has been on reducing process variability, and through that, “the company has achieved successful increases in factory throughput.”
AI Enabled Target Quality Within First Iteration
The tyre major is now creating digital twins of machines, and they are feeding AI systems giving AI much deeper insight into what is happening inside the machines, initially to suggest improvements, and increasingly to take control of machine parameters.
In the tyre industry, one of the reasons for process variability is that the input compound or raw material naturally has some variation, even within an accepted quality band. This variation then leads to variability in the output.
What the company is now doing with AI is enabling it to understand exactly what input is going into the machine for a given production batch, and allowing it to take charge of machine settings to ensure that the output achieves the desired quality parameters the first time. The company refers to this as a ‘closed-loop’ initiative. “This is another initiative that has been underway over the last few months, and it is beginning to show results — not just in quality improvement, but also in increasing ‘right first time’ production. In other words, the company is reaching target quality within the first iteration, or maybe one or two iterations.”
IT-OT Integration for Creating Self Service Capabilities
The first step the company took, more than three years ago, was to connect the OT network with the IT network to start streaming data into the cloud. All data from every single batch, along with sensor data, is now connected to the cloud, and this is what the company uses for analytics.
The company has also trained the majority of its production staff on various technology tools to visualize OT data. Through this, employees have what the company calls self-service capabilities, enabling them to independently access and analyze data to address operational bottlenecks.
By connecting OT data to the cloud, the OT network is exposed to the external world. Therefore, the company has implemented extensive cybersecurity safeguards to ensure network security. This has been a major focus area in the last few years, and the company continues to enhance these safeguards.
The integration is not just about pulling data in one direction. The IT network, where the AI systems operate, is now also taking control of the OT network for certain machines by controlling machine parameters — whether through SCADA systems or by directly interacting with PLCs to manage machine settings.
Gains from AI in FY25
Apollo Tyres saw major gains from the use of AI solutions in FY25. The B2B e-commerce platform, powered by AI/ML, enabled near-instant order tracking, warranty processing, and claims management — reducing turnaround time from days to minutes, said Apollo Tyres in the annual report for FY25.
In manufacturing, “AI-powered MES systems delivered tangible outcomes — including a 30% reduction in scrap, 3–5% productivity gains, and deeper supply chain visibility. Over 750 IoT-enabled data points across global machines feed a central cloud data lake, enabling predictive analytics and production loss detection via custom micro-AI models,” the company said in FY25.
