Digital transformation in manufacturing is entering a more consequential phase. What once centred on automation, connectivity, and incremental efficiency is now evolving into something more strategic, the ability to convert vast volumes of operational data into timely, intelligent decisions that directly influence performance.
In an exclusive interaction with Express Computer, the newly appointed, Global Chief Digital and Information Officer at PGP Glass Pvt Ltd, Santosh Rai, outlines a pragmatic roadmap for industrial enterprises preparing for 2026. His perspective is notably grounded. Rather than viewing AI as a standalone disruptor, Rai frames it as the natural culmination of a disciplined data lifecycle, one that organisations must architect deliberately if they want technology investments to translate into measurable business outcomes.
Manufacturers today operate under multiple pressures. They must enhance operational efficiency while managing supply chain volatility, pursue sustainability while protecting margins, and accelerate digital adoption without compromising governance. Against this backdrop, Rai believes the real differentiator is no longer data generation, it is decision readiness.
The shift from accumulation to intelligence
Over the past several years, manufacturing enterprises have invested heavily in capturing data through ERP platforms, IoT deployments, and manufacturing execution systems. Yet Rai observes that many organisations remain stuck in what he calls the “accumulation phase”.
“There is always a cycle from data to decision,” he explains. “But collecting data is only one part of the journey. The real question is how you structure it and move towards AI-driven insight.”
For Rai, this journey must follow a defined sequence. Data must first pass through strong engineering layers, typically ETL processes, where it is standardised and prepared. Only then can it be visualised, analysed for trends, and ultimately elevated through AI models capable of guiding decision-making.
Without this structure, AI risks becoming superficial, an additional layer placed on top of fragmented data rather than a capability embedded into the organisation’s operating fabric.
He notes that enterprises must therefore define a clear strategy around building an enterprise-grade data lake and integrating it with AI platforms. Treating AI as a series of isolated functional experiments limits its long-term value. Instead, it should be architected at the enterprise level, much like ERP once was.
From IoT enthusiasm to AI intent
Investment patterns across manufacturing reflect this evolution. Rai points out that as recently as 2023, most organisations were focused on IoT — connecting machines, capturing parameters, and building visibility into operations.
Today, the conversation has shifted decisively towards AI.
Organisations now recognise that connectivity alone does not deliver advantage unless the resulting data can inform faster and better decisions. AI provides that interpretive layer.
The maturity gap between enterprises also plays a role. Larger organisations often employ data scientists who can analyse datasets and translate insights for leadership. Smaller firms may lack such specialised resources, making AI an attractive equaliser.
“Once you define an AI model and feed it structured or unstructured data, it can help you take decisions quickly,” Rai notes. “That is the beauty of AI.”
However, speed must not come at the expense of discipline.
Why data sanctity becomes non-negotiable
If there is one area where Rai urges caution, it is data quality.
In many markets, including the Indian subcontinent, he sees persistent challenges around data accuracy and integrity. Poor-quality inputs inevitably lead to flawed insights, a risk that becomes magnified when AI is involved.
“First, you need to capture data sensibly,” he emphasises. “Otherwise whatever trend you plot or information you generate will be wrong.”
For Rai, digital transformation truly begins at the moment an organisation commits to trustworthy data capture. Once that foundation is in place, enterprises can prepare the data within a lake environment, apply AI models, and create a decision layer that enhances responsiveness across operations.
Notably, he believes the definition of digital transformation itself has expanded. Until recently, many initiatives stopped at IoT implementation. Now, organisations increasingly recognise an additional cycle, like the AI layer, that converts information into action.
Automation, connectivity, or AI? Think sequence, not choice
Technology leaders often debate where to prioritise investment: connected plants, automation, or AI-driven insight. Rai rejects the idea that these are competing agendas.
“They are in sequence,” he says.
Data capture remains foundational. Without it, there is nothing to analyse. Yet in environments already rich with historical datasets, even those stored in spreadsheets or PDFs, AI can begin generating value almost immediately. In such contexts, AI becomes the faster lever.
He also highlights a practical advantage: compared with large-scale IoT deployments, AI pilots typically require lower investment and can transition into production more quickly. This shorter cycle reduces financial exposure while encouraging experimentation.
AI’s early wins: Planning and quality
Among the most promising use cases Rai sees emerging is production planning, an area he describes as a “low-hanging fruit”.
Traditional ERP-driven planning processes often demand extensive manual input and time. By contrast, AI models trained on existing data can generate plans within minutes.
“If you pull the data from ERP and use an LLM-based model, planning can happen in fifteen minutes,” he explains. Several organisations are already seeing significant time savings through this approach.
Quality management offers another high-impact opportunity. AI systems can analyse historical defects and alert shop-floor teams whenever a similar product enters the production line. Inspectors can focus on known risk areas while directing more energy towards new challenges.
The result is stronger first-time quality, a metric that resonates deeply in both B2B and B2C environments.
Beyond efficiency, these capabilities also contribute to sustainability. AI-enabled calibration of machines and furnaces can optimise fuel consumption and electricity usage, helping organisations operate more responsibly while controlling costs.
Governance in an AI-first era
As AI adoption accelerates, governance inevitably moves centre stage. Rai points to established frameworks such as ISO 27001 and ISO 20000 as foundational but acknowledges that emerging regulations, including the Digital Personal Data Protection (DPDP) Act, add new layers of accountability.
Compliance, however, cannot remain the sole responsibility of IT or governance teams.
“It is everyone’s responsibility,” he says.
Consider HR functions managing candidate data. If consent timelines expire, that information must be deleted. Similarly, enterprises must establish master data management protocols that ensure a single source of truth feeds downstream systems. Such centralisation simplifies actions like data purging while strengthening overall control.
Master data as the backbone of trust
Maintaining data quality requires structured accountability. Rai advocates defining clear data stewardship roles so that ownership is never ambiguous.
“Master data governance helps data quality, and AI can help validate data as it is entered,” he notes, describing a reinforcing cycle in which strong data supports AI and AI, in turn, protects data integrity.
For organisations seeking to scale analytics, this interplay becomes critical. Trusted data ensures that AI outputs remain reliable enough for operational decision-making rather than merely advisory.
Linking technology to business outcomes
When Rai speaks about the year ahead, he does so through the lens of measurable outcomes rather than technology adoption alone.
Improving overall equipment effectiveness, strengthening first-pass product quality, and advancing sustainability rank high on his agenda. Each objective ties directly to operational performance, a reminder that digital investments must ultimately serve business goals.
Yet achieving these ambitions requires more than platforms and processes.
“The real challenge is people,” Rai says.
The talent gap manufacturing cannot ignore
Across the industry, demand for expertise in data engineering and data science continues to outpace supply. Rai attributes this partly to the limited integration between academia and industrial realities.
While skill-based education is gaining traction, he believes students need greater exposure to real manufacturing environments, where AI models must control machines, diagnose faults, and function within operational constraints.
Building a small chatbot may demonstrate technical competence, he suggests, but the true test lies in applying AI within complex industrial settings.
Greater collaboration between academia, industry, and policymakers will be essential to close this gap.
The next frontier: Convergence of AI and industrial tech
Looking beyond immediate priorities, Rai anticipates deeper convergence between AI and advanced sensing technologies.
Vision systems, for example, can combine cameras with AI algorithms to inspect every product leaving a facility, ensuring it meets quality standards in real time. If anomalies arise, operators can be alerted instantly.
Similar integrations are emerging across optics and sensor-driven applications, signalling a future in which intelligence is embedded directly into operational workflows.
For Rai, 2026 is likely to be defined by this fusion, the blending of AI with cutting-edge industrial technologies to create smarter, more autonomous factories.
A message to technology leaders: Architect before you scale
Asked what guidance he would offer peers planning for the year ahead, Rai’s advice is refreshingly direct.
Do not replicate another organisation’s blueprint, he says, because every business operates under unique constraints. Avoid fragmented deployments, and instead focus on building an enterprise-wide AI framework.
Equally important is workforce readiness. Leaders must educate teams across the AI spectrum, from data engineering to data science, so that the organisation develops a shared understanding of how intelligence drives value.
Ultimately, Rai returns to a principle that feels increasingly relevant in the age of AI: structure must precede scale. Enterprises that treat AI as an enterprise capability, grounded in trusted data, strong governance, and skilled talent, will be best positioned to convert information into advantage.
For manufacturing leaders navigating the next phase of transformation, the takeaway is clear. Competitive differentiation will not belong to those who generate the most data, but to those who design the clearest path from data to decision and from decision to action.