How India’s GCCs are emerging as AI-led product innovation hubs

India’s Global Capability Centre (GCC) ecosystem is undergoing a structural transformation. What was once viewed largely as a cost optimisation strategy is now increasingly evolving into a high-value product engineering and innovation engine for multinational enterprises.

As AI reshapes software development, enterprise security, compliance, and developer workflows, GCCs are also being pushed to take on larger ownership responsibilities across product strategy, AI infrastructure, software supply chain governance, and platform engineering.

In an exclusive interaction with Express Computer, Prasanna Raghavendra, Senior Director – R&D, JFrog speaks about how AI is fundamentally changing software development pipelines, why binary-level governance is becoming critical, and how enterprises need to rethink software integrity, compliance, and developer security in the AI era.

India GCCs are shifting from execution centres to value creation hubs

According to Raghavendra, India’s GCC story is rapidly evolving from labour arbitrage to strategic innovation ownership.

JFrog India entered India through an acquisition after acquiring Raghavendra’s startup, but the larger objective was never about building a low-cost engineering base. It was always about value generation. The focus was on building leadership-driven engineering capabilities capable of owning meaningful product domains.

Initially, the India centre focused heavily on engineering before gradually expanding into products, support, and business continuity functions. “From a business continuity perspective, because a huge amount of R&D is in Israel, we started building areas where we could overlap and take over whenever needed.”

Today, he believes Indian GCCs are increasingly moving towards creating entirely new product verticals rather than simply supporting global headquarters.

AI is changing how multinational companies view Indian engineering teams

AI is significantly changing how enterprises evaluate engineering capability and domain ownership inside GCCs.

Domain expertise accumulated inside GCCs over time is becoming extremely important in the AI era. “There is a lot of local domain knowledge built over a period of time, and that is critical for reusing AI material,” he says.

This is leading Indian engineering teams to take on larger portions of development ownership across organisations. “They will start taking a much larger share of that part of the development. It could be horizontal or vertical depending on what those GCCs are building.”

AI is accelerating development, but operational complexity is exploding

While AI is undeniably improving productivity and reducing software development timelines, Raghavendra believes enterprises are underestimating the operational challenges AI introduces into modern software delivery pipelines.

One of the most immediate concerns is infrastructure scalability.

Beyond infrastructure pressure, enterprises are also struggling with AI model governance and trust validation. “There are these new animals called ‘AI models’ right now. How do you trust it? The moment you start using it and newer versions are coming, you need to go through a firewall from an enterprise perspective,” he points out. 

To address this, “We provide what is called AI curation, which can validate these models and give you a pruned list of models approved as per the company’s policy.”

At the same time, enterprises are also facing massive growth in model delivery complexity. The scale of this shift is unprecedented.

“One of the numbers from our software supply chain survey: 1.4 million new models were created in Hugging Face in 2025. 58% of these models are getting downloaded by public networks. So that would mean you need to rapidly start supporting newer things.”

AI is changing software governance from source code to binaries

Raghavendra points out that AI is fundamentally changing the software artefact enterprises need to govern. “With AI, code is becoming a not-so-important artefact. What runs is binary.”

This shift is forcing enterprises to rethink software integrity frameworks entirely.

Earlier, enterprises largely focused on Software Bills of Materials (SBOMs), which documented the components used inside software packages. Now, AI introduces an entirely new governance layer.

According to Raghavendra, enterprises now need AI Bills of Materials capable of tracking AI model lineage, prompts, evidence trails, and validation metadata.

He also asserts that AI-generated software still lacks full trust and confidence inside enterprises. “The confidence is still low. So, there is a lot of manual validation of software happening.”

However, he expects automation-driven verification systems to mature rapidly over time.

Enterprises need evidence-driven software governance frameworks

Enterprises now need continuous evidence collection frameworks across software pipelines. This means capturing verifiable metadata at every stage of software delivery. Once enterprises build evidence frameworks, governance becomes programmable and enforceable.

This also becomes critical for compliance traceability and vulnerability response.

Software supply chain security is becoming a boardroom priority

Raghavendra explains that modern software supply chain security now extends far beyond vulnerability scanning.

Enterprises now need policy-driven “firewalls” that validate every dependency entering enterprise environments.

This becomes especially important because open-source packages and AI models continue evolving even after deployment. 

He also highlights how AI systems introduce entirely new governance challenges around data trustworthiness. This creates a requirement for timestamped evidence systems capable of validating data freshness and AI compliance continuously.

Open-source ecosystems are creating new security risks at scale

“Open-source ecosystems are becoming significantly more complex in the AI era,” says Raghavendra.

He warns that threat actors increasingly target small upstream open-source dependencies capable of infecting entire downstream ecosystems.

This makes curated enterprise firewalls and operational risk assessments essential.

Enterprises now need continuous curation systems evaluating open-source tools across operational reliability, security posture, and licensing compliance simultaneously.

Developers are becoming the new enterprise attack targets

Raghavendra warns that developers themselves are increasingly becoming prime attack targets.

To counter this, they have started building firewall protections around developer IDE plugins and AI-assisted coding tools to recheck.

Is this vulnerable from an AI perspective? According to him, one compromised developer laptop can now become an entry point into entire enterprise environments.

Looking ahead, Raghavendra believes developers will increasingly evolve into “skill engineers” capable of orchestrating AI systems modularly.

Rather than writing every layer of software manually, developers will increasingly focus on designing modular AI-driven workflows.

He believes the future of development will revolve around granular modular orchestration rather than monolithic coding.

Although Raghavendra acknowledges it is still early to predict whether AI can fully replace traditional development, he believes the shift is already transforming engineering culture significantly.