The Hardest Part of Industrial AI Isn’t Adoption—It’s On-Site Deployment

Augury, a New York-based industrial AI company, is betting on forward-deployed engineers (FDEs) to help manufacturers move beyond AI pilots. 

As it expands its Bengaluru engineering hub, the company plans to grow its FDE team to 25–30 people over the next two months, with engineers working directly with customers to adapt AI agents to the realities of factory operations. 

The new location expands Augury’s global footprint, complementing its existing hubs in the United States and Israel. The expansion builds on the company’s $75 million funding round in 2025, which supports product development and global expansion. The new hub will focus on advancing agentic capabilities within the Industrial AI Workforce, which includes role-based AI agents engineered for reliability, maintenance, and operations teams.

The strategy reflects a shift in enterprise AI deployment: companies are increasingly expected to deliver business outcomes, rather than simply provide software that customers must integrate and operationalise themselves. 

“The fundamental difference is the FD kind of knows and breathes the customer,” Anoop Mohan, Chief Product and Technology Officer at Augury tells AIM.

The model is also tied to a more outcome-oriented commercial arrangement. Augury plans to deploy teams comprising an industrial subject-matter expert, a general engineer, and a data engineer to work with customers over 90 days.

“Only when you have an outcome, you pay us. Otherwise, you don't pay us,” Mohan says.

The arrangement raises the bar for AI vendors. They must demonstrate that their systems improve operations, rather than relying on model performance or successful deployment as proof of value.

Why Industrial AI Needs FDEs

Manufacturing environments present a different challenge from deploying AI in conventional office workflows. Data is often distributed across multiple systems, while critical operational knowledge sits with experienced engineers and plant workers. 

An AI agent may be technically capable of analysing data, but its usefulness depends on whether it can interpret the industrial context and provide recommendations that operators can trust.

Augury is developing AI agents for manufacturing teams, including an operations agent designed to consolidate information from disparate systems, identify emerging problems, and recommend corrective action.

“One example involves declining production yield. When output begins to fall, operators may need to examine several screens, escalate the issue to a process engineer, and wait for a diagnosis before making changes. This process can take hours or even days,” Elan Greenberg, CEO of Augury, states.

The company says its operations agent can analyse the problem and recommend changes to production parameters, helping engineers respond faster. It also aims to identify warning signs before an incident causes unplanned downtime.

For manufacturers, the commercial stakes are significant. Unexpected stoppages can disrupt production schedules, delay orders, and increase waste. 

Augury says some customers have reported savings worth tens of thousands of dollars over weekly or daily periods, while the company has cited returns ranging from five times to 50 times the investment in selected deployments. These figures are company-reported and were not independently verified.

Gilad Bornstein, VP of Engineering and Israel site lead, Augury discusses the operational use cases and says the impact of such systems could extend beyond efficiency gains.

“Because the margins in heavy industry and manufacturing and energy tend to be somewhat slim, some of these outcomes these gentlemen are talking about could be the difference between a plant that is deeply unprofitable and needs to close and let people go. Or the difference between a plant that's very profitable and successful in hiring more people and supporting the community and being a center of commerce and trade,” Bornstein explains.

However, such returns will depend on whether AI recommendations translate into sustained improvements on the production line. That is where customer-specific adaptation and on-site engineering become central to Augury’s approach.

Difficult Hiring

Augury currently has around six people hired in its FDE team, with a target of 25–30 over the next two months. The company has approximately 300 employees globally, including around 175 in product, R&D and technology, and about 55 employees in India. 

The hiring challenge is finding people who can combine AI and software engineering skills with an understanding of industrial processes. Mohan says the company will initially focus on lateral hires, particularly candidates with three to six years of experience, while exploring longer-term internship and campus-hiring opportunities.

“You cannot be an FDE if you don't understand the domain,” Mohan states, adding that candidates must also understand the product and how it integrates with other industrial systems.

The combination of skills is difficult to find. The role cannot simply be filled by relabelling conventional software engineers or professional-services employees. Engineers need to work closely with end users, understand the constraints of production environments, and take responsibility for whether the deployed solution works in practice.

Bengaluru Engineering Centre

Augury’s Bengaluru expansion is intended to give the Indian team ownership of end-to-end product and engineering capabilities rather than confine it to execution work.

May Cohen Zack, People and HR lead, Augury says that the company’s decision to expand in India was driven by the adaptability of local talent, experience gained from previous work in the country, and the presence of strategic partner Rockwell Automation.

“We are building here now. It's an extension. Peers to other offices that we have at Augury, full state-of-the-art kind of development taking place,” Ganesh Iyer, Senior Director of Engineering and site leader India, describing the company’s approach to its Indian engineering operations.

The company is also looking to deepen its collaboration with Rockwell Automation, whose industrial technology and customer relationships could help Augury bring AI capabilities to manufacturing facilities.

Rockwell Automation India Managing Director Dilip Sawhney describes the broader industry shift as a move “from automation to autonomy”, arguing that industrial systems will increasingly need AI to detect, diagnose, and respond to problems before they disrupt operations.

The partnership could give Augury access to industrial customers and domain expertise that would be difficult to replicate through software development alone. 

Scaling Beyond Individual Deployments

The central question for Augury is whether FDEs can help scale AI adoption without turning every customer deployment into a labour-intensive consulting project.

Mohan says that the company does not intend to scale in a linear fashion by adding more engineers for every additional project. Instead, it wants to use its AI agents to augment the work of FDEs and demonstrate value more quickly.

If the agents can handle more of the diagnostic and analytical work while FDEs focus on domain-specific adaptation, the approach could make industrial AI more practical for manufacturers.

The company’s Bengaluru hub will play a role in testing that proposition. Augury says it is already receiving interest from Indian manufacturers, although it did not name the prospective customers discussed during the interview.