For years, robotic process automation (RPA) was the buzzword that defined enterprise efficiency. It promised a world where repetitive work vanished into code. Yet, those early automations still needed human judgment.
That equation is changing fast. Dominic Pereira, VP – Product Management, Automation Anywhere, describes a transition from rule-based bots to reasoning-driven agentic automation.
The company’s Agentic Process Automation (APA) system now integrates context, reasoning, and decision-making, qualities once considered too “human” for software. “With traditional automation, you would get things like speed, accuracy, reliability, right?” he quipped during an interaction with AIM.
“If I bring that along with reasoning…recommendations, suggestions, which were a big missing block, from 20, 30% automations or processes, we’re able to automate up to 40 to 80%.”
From RPA to Reasoning: The Agentic Shift
At the heart of this transformation lies what Pereira calls thoughtful automation, the ability to let AI agents work in tandem with humans rather than replace them.
“Every company is talking about agentic automation or agentic AI in one way or another,” he said.
Agentic automation introduces what Automation Anywhere calls a process reasoning engine, a system that understands enterprise context, evaluates data across workflows, and makes adaptive decisions. This is not a step-by-step automation of tasks but an orchestration of outcomes. The new paradigm focuses on long-running, outcome-driven processes rather than siloed task automation.
In use cases such as banking or insurance, Pereira explained that agents can now process massive amounts of documentation, analyse risk profiles, and suggest recommendations, while humans validate the final decision.
“But what the human is [trying to avoid] is the busy work, right? … and that’s where you see the digital assistant helping the user,” he said.
To help enterprises move beyond small proof-of-concept projects, Automation Anywhere has begun introducing pre-built agentic solutions.
“You can rinse and repeat, you can build a solution with a few tweaks, you can make it work,” Pereira explained.
“We started introducing several solutions, and you’ll see our announcements almost every month; we release a new solution based on the demand we get from our customers.”
India: The Nerve Centre of Global Automation
While Pereira’s global role keeps him at the centre of product evolution, his base in India isn’t incidental. Half of Automation Anywhere’s workforce and nearly all its process orchestration design and product management operations are based in the country.
India, he explained, not only builds but runs Automation Anywhere’s global automation stack.
“I think globally we have close to 290, I would say, automations that run, 95% are run out of India.”
Much of the innovation in process reasoning engines, document automation, and orchestration tools originates from Indian teams.
“The entire design, product management, and the core team are based out of India.”
Their document automation product, one of the fastest-growing offerings, experienced nearly 5x growth in recent years. The key product management teams for this entire portfolio are based in India.
“It’s like building in India, building for the globe,” he said.
This synergy of development, experimentation, and deployment gives India a unique position, not just as a delivery hub but as a laboratory for the next leap in automation.
Reimagining Work in the Age of AI Agents
If RPA once took away manual work, APA is now redefining how humans work. Pereira views the next phase of automation as human+AI cohabitation, where every employee will have a digital co-pilot or assistant.
He explained, “A lot of that busy work can now be done by the agents, and it will be reimagining work.”
He continued, “Our belief is that in the next few years, it’s already happening, every employee will have a co-pilot or a digital worker as an assistant.”
“[AI will do] the busy work, do the crunching, do the analysis, whereas [humans] will focus more on things like customer experience, customer engagement, and creative work.”
The company has also invested heavily in developer experience, from low-code interfaces and drag-and-drop AI agent builders to self-healing automation that adapts to changes in application interfaces.
To ensure safety, Automation Anywhere embeds four layers of governance: design-time testing, human-in-the-loop validation, LLM evaluation, and real-time monitoring.
“We definitely recommend bringing in a human in the loop,” he said. “There are 10 steps where, you know, eight can be done by AI agents; we still recommend having a human.”
Pereira shared ideas on where it is headed. “When we started our journey, we built what we call task-specialised AI agents. It did a specific task,” he said. “Now, with goal-based AI agents, it understands an intent or a goal, it breaks it down into specific steps.”
The Path to an Autonomous Enterprise
Automation Anywhere’s journey mirrors the broader evolution of automation itself, from recording macros to designing thinking systems.
Pereira explained that the company’s newer approach is focused on outcomes, rather than just speed.
India, in this story, is not just a location but a force multiplier. It’s where agentic automation is being built, tested, and scaled, proving that the centre of enterprise AI innovation may already have shifted eastward.
The post India Is Quietly Powering the Brains Behind Global Enterprise Automation appeared first on Analytics India Magazine.
