The AI-native GCC has a new test, and it is not headcount

In a factory, an AI recommendation is only as valuable as the decision it helps an operator make. A system may analyse machine data, identify a potential problem and suggest a corrective action. Getting a plant to trust that recommendation enough to act on it is a different challenge altogether.

That tension is becoming central to industrial AI, where the consequences of a wrong decision can extend beyond software performance to production losses, product quality and plant operations. It also raises a larger question for global capability centres. As AI accelerates engineering and allows smaller teams to accomplish more, will a GCC’s value be measured by the number of people it employs or by the complexity of the problems it can solve?

At Augury, an industrial AI company focused on machine health and manufacturing operations, that question is playing out in both product development and the company’s expansion in India. For Ganesh Iyer, Senior Director of Engineering and India Site Leader, the shift is visible in how quickly a product can move from concept to customer deployment.

When engineering output matters more than headcount

Iyer describes a development cycle that took Augury from a product vision and customer identification to execution, delivery and an initial proof point, before scaling from one customer to 10 within nine months.

“That is only possible and we didn’t increase our budget,” Iyer says, attributing the pace to an AI-native development approach.

The distinction is important. AI is not simply being added to an existing product. It is also changing the process through which products are built. The question for engineering leaders is whether teams can use AI to shorten development cycles without compromising the quality and reliability expected of enterprise technology.

Elan Greenberg, CEO of Augury, sees industrial sectors as particularly ripe for this shift. The company has spent 15 years working in industrial environments, where machine performance, production efficiency and unplanned downtime have direct commercial consequences.

Greenberg describes the current period as one of the fastest phases of product and engineering innovation he has experienced. For him, GCCs offer an organisational advantage because they can work closely with teams across geographies while remaining connected to the company’s wider product and engineering priorities.

That integration is becoming more significant as AI changes the relationship between workforce size and output. But industrial applications introduce a complication that software engineering metrics alone cannot capture. A product must work within the specific conditions of a factory, where even small changes to a process can affect yield, quality and profitability.

The challenge is turning data into decisions

Manufacturing plants have long generated substantial amounts of operational data. The difficulty has been making that information useful to the people responsible for production. Machine-health data, process information and operating procedures often sit in different systems, leaving workers to connect the dots.

“There’s more data than operators can ever make use of on their own,” Greenberg says.

Augury combines machine-health information with process and operational data to help its customers interpret what is happening on the factory floor. Its industry copilot is designed to help manufacturing workers make decisions, whether they are monitoring production, investigating machine behaviour or trying to improve yield.

Greenberg uses the analogy of driving a car. Traditional machine data can resemble a speedometer or odometer, showing what is happening without necessarily explaining what should happen next. AI agents can go further by making recommendations about operational adjustments, such as changing a temperature or conveyor speed to help meet production targets. “The difference can sometimes just be a couple of different points of yield or unplanned downtime,” he adds.

For low-margin manufacturers, such differences can determine whether a facility remains profitable. The commercial case for industrial AI therefore depends less on the novelty of the model and more on whether its recommendations improve production outcomes.

Anoop Mohan, Chief Product and Technology Officer at Augury, identifies the connection between fragmented information sources as a key part of the problem. Historically, people had to gather data, manuals and standard operating procedures before making a decision. Agentic systems can connect these sources through what he describes as a context graph, allowing the system to interpret information in relation to the wider operation.

The ambition is to move from detecting a problem to helping resolve it. Yet the transition from recommendation to action brings its own set of questions.

Why autonomy depends on trust

In industrial environments, an AI system may be technically capable of triggering an action through a plant’s control systems. Whether a customer is prepared to allow it to do so is another matter.

“The difference between autonomous and autonomous action is more the trust level for the customer,” Mohan says.

Augury can, for example, push a recommendation into a Rockwell control environment where a programmable logic controller can execute the action. The remaining question is whether the customer wants the system to act independently or prefers a person to review the recommendation first.

That distinction makes transparency and accountability central to industrial AI deployment. Plant operators need to understand why a system recommends a particular intervention and whether its reasoning reflects the realities of their facility.

“You can build trust year after year after year after year. And you can lose it very quickly,” Greenberg says.

The concern is particularly relevant as AI development accelerates. Faster product delivery creates opportunities, but speed cannot come at the expense of quality when systems influence physical operations.

Greenberg says the company must ensure that new solutions meet the standards customers expect from technology deployed in their plants. AI, he argues, should improve both development velocity and quality control, rather than forcing companies to choose between the two.

The last mile still needs people

This is also why industrial AI does not eliminate the need for people who understand the customer environment. Augury uses forward-deployed engineers who work with plant operators and leaders to adapt solutions to individual facilities.

Even plants belonging to the same manufacturer are not necessarily identical. Their equipment, processes and accumulated operational knowledge can differ. Some of that knowledge exists in documentation, while some resides with experienced workers.

Forward-deployed engineers help account for these differences, adapting the final 10 to 20 per cent of an agent-based solution to the needs of a particular facility. They also bridge the gap between the executives who approve a technology purchase and the operators who must use it.

The role reflects a wider challenge for industrial AI. A system may be built centrally, but its effectiveness depends on how well it fits the environment in which it is deployed. Automation can reduce the burden of gathering and interpreting information, allowing experienced workers to spend more time applying their judgement.

Greenberg does not expect people to disappear from manufacturing. Instead, he sees their roles changing as technology gives them greater capacity to make decisions. Some facilities may become increasingly autonomous, although the extent will vary by industry and application.

Bengaluru and the new talent equation

Augury is expanding its Indian operations, with Bengaluru being developed as a second major hub. The company already has a team in Delhi and plans to bring more of its India-based workforce closer to the Bengaluru operation over time. At the time of the interaction, the company had around 50 employees in India and more than 30 open positions.

The hiring challenge is not simply finding AI engineers. Iyer says the company needs people who combine AI capabilities with software development skills, an understanding of industrial applications and the ability to work in a startup environment.

“The talent pool is huge, but we have a unique intersection looking for AI and industrial and the right attitude to work in a startup like that,” he says.

That combination narrows the field. The objective is not to build a large centre for its own sake, but to recruit people who can take ownership of products and contribute to a globally distributed engineering organisation.

“We’re very clear that you’re not going to become a 300-member GCC,” Iyer says.

The remark captures the broader question facing GCC leaders as AI reshapes engineering. Larger teams may no longer be the clearest indicator of a centre’s contribution. But smaller teams will still need deep expertise, operational understanding and the discipline to deploy technology responsibly.

For industrial AI, the measure of progress will ultimately be what happens after the code is written. A system must work in the factory, earn the confidence of the people using it and deliver results that justify the trust placed in it.