How Biocon is Leveraging AI to Reshape Both Drug Discovery and Manufacturing

Bengaluru-headquartered Biocon, one of India's largest biopharmaceutical companies, is quietly reconfiguring how it discovers, develops, and manufactures medicines, not through a single sweeping overhaul, but through the gradual embedding of artificial intelligence across its operations.

In a conversation with AIM, Mandar Ghatnekar, the company's chief technology officer, talks about a deliberate two-track approach. 

"AI at Biocon operates across two clear and complementary dimensions: AI for Enterprise, focused on productivity and operational excellence, and AI for Science, focused on deeper, transformation-led innovation in R&D," he notes.

That distinction, operational AI versus scientific AI, reflects a broader tension playing out across the global pharmaceutical industry as companies attempt to separate what technology can deliver today from what it might eventually promise.

The Science Track

Biocon's primary business is biosimilars—medicines that are near-identical to biological drugs whose patents have expired. The development pathway is demanding. Biosimilars must demonstrate comparable safety and efficacy to their reference products, a process governed by strict regulatory and clinical requirements across the US, Europe, and other markets where Biocon operates.

The company recently received US Food and Drug Administration (FDA) approval for two denosumab biosimilars and has been among the first to win FDA approval for biosimilars of trastuzumab, pegfilgrastim, insulin glargine, and aflibercept, among others. The commercial pipeline gives AI applications in R&D a clear target: reduce the time and cost of bringing these complex molecules to patients.

Ghatnekar says AI is being applied to molecule design, analytical characterisation, comparability assessments, and process development. "AI helps our scientists connect large and often fragmented datasets, identify meaningful patterns, predict critical quality attributes, and design more efficient development strategies," he adds.

However, when it comes to claims of saving time, he remains measured. "It is still early to quantify the full impact on overall timelines, but we are already seeing encouraging improvements in development efficiency and scientific decision-making." The real aspiration, he affirms, is shifting R&D "from a largely iterative process to one that is increasingly predictive, data-driven, and precise."

Across the sector, artificial intelligence is compressing drug discovery timelines, and regulators are beginning to accommodate this shift. In January 2025, the FDA released draft guidance on the use of AI to support regulatory decision-making for drug and biological products. In 2023, the Indian Council of Medical Research also published ethical guidelines for the application of AI in biomedical research and healthcare.

The Manufacturing Floor

Biological drug manufacturing is inherently complex: small variations in temperature, pH or culture conditions can affect product quality. The cost of a failed batch—in wasted materials, lost time, and regulatory scrutiny—is significant. At Biocon, AI has had better success in manufacturing than in labs.

"AI enables predictive process control, helping reduce variability, improve yields, anticipate deviations before they occur, and drive greater consistency at scale," Ghatnekar says. In quality operations, he notes AI has the potential to shift organisations "from reactive quality management to real-time quality assurance."

Asked where AI had delivered the bigger impact, Ghatnekar asserts the two areas are not directly comparable. "AI for Enterprise is delivering more immediate, structured operational gains today, while AI for Science represents the larger long-term opportunity. Over time, the latter could fundamentally reshape the entire healthcare ecosystem."

A 2025 survey of 400 senior life sciences managers across the US, the UK, Switzerland, and India found that 54% of industry leaders ranked AI and smart technology integration among their top priorities for future laboratory environments. They also want AI in real-time monitoring, predictive analytics, and workflow automation.

Regulation and Accountability

However, the use of AI in pharmaceutical manufacturing also raises some questions: when an algorithm flags an anomaly, or recommends a course of action, who carries responsibility for what follows?

Ghatnekar is unambiguous. "Accountability does not shift from humans to algorithms. AI serves as a decision-support tool to help teams identify patterns, flag anomalies, and generate insights more efficiently. But all critical decisions continue to be made by qualified personnel."

He asserts that consequential processes like batch disposition and product release decisions remain with designated quality authorities. "The responsibility for ensuring product quality, patient safety, and regulatory compliance always remains with qualified people," the executive shares.

In a regulated biopharmaceutical environment, any AI application must operate within the existing GxP framework—a set of good practice guidelines that govern pharmaceutical manufacturing, laboratory operations, and clinical trials. "AI models and digital tools are subject to established validation methodologies, ensuring they are reliable, transparent, and fit for regulatory scrutiny," Ghatnekar notes.

Regulators are building frameworks to formalise these requirements. The European Medicines Agency published the first version of GMP Annex 22 specific to AI in pharmaceutical manufacturing. In the US, the International Society for Pharmaceutical Engineering released an updated GAMP guide addressing AI validation in GxP environments. Both frameworks emphasise data integrity, model transparency, and human oversight as non-negotiable requirements.

The Workforce Question

Beyond scientific and manufacturing outcomes, AI adoption is reshaping what pharmaceutical employees do from day to day, sparking concern in some quarters about the longer-term implications for jobs.

At Biocon, Ghatnekar describes the change as a reallocation of human effort rather than a reduction in it. Tasks such as data compilation, documentation, routine trend analysis, and workflow management can now be automated or accelerated through AI tools. 

"This reduces administrative burden and allows scientists and analysts to devote more of their time to activities where they add the greatest value: interpreting results, investigating deviations, solving complex problems, and making scientifically informed decisions," he explains.

But he also acknowledges that AI has created a new layer of responsibility. "In a regulated biopharmaceutical environment, AI-generated outputs cannot simply be accepted at face value. Laboratory professionals must assess whether the insights are scientifically sound, conform to established methods, and determine the appropriate course of action."

The industry more broadly is grappling with similar dynamics. Research published by Intuitionlabs in 2025 found that quality-control specialists are increasingly expected to interpret machine-learning outputs, and that bench scientists in drug discovery must now evaluate the output of AI design algorithms and decide on experimental next steps. 

For now, Biocon's priority is to ensure its workforce evolves alongside the technology through continuous learning and capability building. 

The Long View

Biocon's broader context gives these decisions added weight. The company has grown from revenues of ₹164 crore in the early 2001–2002 to close to ₹16,780 crore in the financial year ended March 2026, with the biosimilars business—which contributed 59% of total revenue in the last reported financial year—as the primary engine. It has also expanded its US manufacturing footprint, inaugurating a facility in Cranbury, New Jersey.

The company is now positioning itself to evolve from a biosimilar-led drugmaker to a broader innovation-led biopharmaceutical company, with AI-led biologics cited among its future growth priorities.

"The real opportunity lies in shifting R&D from a largely iterative process to one that is increasingly predictive, data-driven, and precise," Ghatnekar says.

Whether that shift arrives as quickly as the industry hopes will depend on how well AI tools perform against the rigour that pharmaceutical development demands.