Breaking profile: Mihir Shukla on work after busywork Automation Anywhere enters its third decade The Five-Year Century became a national bestseller in 2026

Person / Founder / The future of work

Mihir Shukla Is Rewriting the Office’s Oldest Habit: Busywork

He left a PhD path for the early internet, watched office workers copy data between screens, and spent two decades turning that frustration into Automation Anywhere. Now his bet is larger: AI should change the design of work, not merely add another tool to it.

The founding insight arrived in a room full of expensive software and bored humans. Mihir Shukla was traveling through offices where educated workers spent long stretches copying information from one screen and pasting it into another. The computers were everywhere. The work was still mechanical. “We weren’t put on the planet to do this,” he would later say. “This is not a human thing to do. It is very robotic.”

The observation sounds obvious now, when software agents can summarize documents, answer questions and take actions across applications. In 2003 it was a stubborn premise on the edge of enterprise computing: teach software to operate the interfaces companies already had, and let it carry the repetitive load. Shukla and co-founders Neeti Mehta Shukla, Ankur Kothari and Rushabh Parmani started a company called Tethys Solutions in San Jose. It became Automation Anywhere, and the nuisance they picked at became an industry.

Shukla’s story is less about a sudden flash than a sequence of technologies becoming ordinary. He studied computer software at the Maharaja Sayajirao University of Baroda from 1990 to 1994. He came to the United States intending to pursue a PhD just as the commercial internet began opening. The new medium proved more compelling than the degree. He has said he caught “the disruption bug.”

A front-row seat to things becoming normal

His early career reads like a tour of the internet’s unfinished rooms. He co-founded RCube Technologies and worked with Kiva Software, whose technology landed inside Netscape. At Infoseek, he led engineering for a platform group during the era when searching the web was still a problem to be defined. He advised OmniSky, which put internet access on Palm devices. Shukla likes one detail from that period: for a moment, he was among a tiny group of people who could stand anywhere and use a mobile device to locate the nearest restaurant. Billions can do it now without noticing.

The expanding surface area of Shukla’s career: first computers reached information, then they began to execute work across systems. Bar length is illustrative, not a market measure.

The useful lesson was not that technology moves quickly. It was that a startling capability eventually becomes plumbing. At E2open, where Shukla served as director of engineering, the plumbing was literal in the software sense: integrating a global supply chain across hundreds of applications and thousands of people. The brittle handoffs exposed a gap. Companies owned powerful systems, but the work between them still depended on people ferrying information.

“I never wanted to be a CEO. I always thought about what impact I wanted to create, what change I wanted to bring, and what values I wanted to represent.”Mihir Shukla

Automation Anywhere began by making those handoffs programmable. A software bot could read a screen, type into an application, follow rules and connect tools without a company rebuilding its entire technology stack. The proposition had a wonderfully practical quality. Keep the old systems. Give the repetitive path through them to software.

The patience inside a noisy category

There is an odd rhythm to building a company before its category has a familiar name. Automation Anywhere operated for roughly fifteen years before taking institutional venture money. In 2018, it raised $550 million across Series A financings. A $290 million Series B followed in 2019. By then, robotic process automation, or RPA, had moved from an operations specialty into executive vocabulary.

2003Company founded in San Jose
15 yrsBefore its first institutional round
$840mRaised across 2018 and 2019 financings

Shukla describes his leadership education in similarly cumulative terms. A software engineer by training, he became curious about the hundreds of other functions that make a company work. He learned to articulate a picture of change, then to build systems that could direct people’s energy toward it. The CEO title, in his telling, was incidental to the impact he wanted.

His favorite measure of that impact was not confined to a dashboard. In the company’s early years, he visited customer operations and asked employees who used automation whether they would ever go back. He remembers their eyes lighting up and the answer arriving: “Never, ever.” The anecdote contains a sharp product test. A tool becomes part of work when the person doing the work refuses to surrender it.

The Automation Anywhere thesis, compressed: begin with an actual process, then add intelligence without losing control of the result.

From robots that click to agents that reason

The word “robot” helped people picture the early product, but it also created confusion. These were not machines rolling through warehouses. They were software workers operating applications. In a 2019 demonstration, Shukla described a customer automating a process spanning 68 systems in three weeks. The drama lived in the connective tissue: old applications, inconsistent interfaces and a process that crossed all of them.

Generative AI changed what could happen inside that connective tissue. A bot following explicit steps could now be paired with models that interpret unstructured information and agents that select actions. Automation Anywhere renamed its broader category agentic process automation. Shukla’s current destination is the “Autonomous Enterprise,” an organization where the company says up to 80 percent of work could run autonomously or with AI assistance.

The percentage is a declared vision, not a description of the average office. More revealing is the unit of analysis. Shukla is not talking about a clever assistant attached to each application. He argues for complete processes that cross applications, with orchestration, context, security and accountability traveling alongside the agent. In this view, the chatbot is a door. The business result is in the rooms beyond it.

The useful question

Do not begin with “Where can we add AI?” Begin with the process people dread, map the systems and decisions it touches, and decide what outcome would make the work meaningfully better.

That framing also explains his emphasis on governance. Years before agents became standard conference material, Shukla warned that digital workers had to be built with ethics and morality in mind. The concern is now operational. An agent that acts across financial, customer or employee systems needs boundaries, identity, observability and a clear path back to a responsible person.

Automation Anywhere’s 2026 product announcements followed that logic: a refreshed orchestration platform, a reasoning engine trained on enterprise process data, a fabric for observing agents, pre-built solutions for IT and finance, and partnerships with Cisco, Nvidia, Okta and OpenAI around a governed agent environment. The catalog is broad, but the through line is old. The software should carry work across the cracks.

Opportunity is the other half of automation

Shukla’s public case for automation always returns to human potential. He grew up in a small town in India and has said he did not have access to a computer until college. In 2023, he visited Moshi Technical School in Tanzania, where donated equipment supported a computer lab. Students showed him coding projects, an astronomy tool, maps and an English-Swahili dictionary. The visit came after colleagues traveling with him had summited Mount Kilimanjaro, but the memorable climb in his account was the one inside the classroom.

He told the students a line that has become central to his worldview: talent is evenly distributed, opportunity is not. Automation Anywhere has paired product access with training and nonprofit collaborations. Shukla also sponsors the Shukla Centre for Future Studies in India, which introduces middle-grade students to AI, automation and robotics. Since 2025, he has served on the advisory board of the Institute for Digital Innovation & AI at Melbourne Business School.

“Innovation and drive can come from any human, anywhere in the world.”Mihir Shukla

The social bargain is clear in his telling. If automation removes routine work, institutions have to help people reach what comes next. Reskilling cannot be a memo issued after deployment. It belongs inside the design of the transition. That makes education more than philanthropy around an automation company. It is part of the product’s moral argument.

A five-year century

In 2026, Shukla and Automation Anywhere’s chief people experience officer, Nancy Hauge, published The Five-Year Century. The book’s title contains its wager: demographic pressure, stalled productivity and accelerating AI could compress a century’s worth of organizational change into five years. It became a national bestseller in June.

The pairing matters. Shukla brings the technologist’s view of what systems can do. Hauge brings four decades of experience with the people inside those systems. Their argument pushes AI out of the IT department and into leadership’s basic questions: how decisions get made, how teams are structured, where trust lives and what outcomes an organization actually values.

Joined Melbourne Business School’s digital innovation and AI advisory board.

Automation Anywhere announced its acquisition of Aisera, adding service agents for IT, HR and customer operations.

Presented the company’s expanded autonomous-enterprise platform and new agentic solutions.

The Five-Year Century was recognized as a national bestseller.

This is the newest version of the question Shukla asked in those early office visits. What work deserves a person? The answer changes as machines acquire new abilities, but his filter has stayed surprisingly consistent. Repetition belongs to software. Judgment, invention, empathy and relationships deserve more room. Technology establishes the range of possible choices; leaders remain accountable for making them.

There is also a useful humility in remembering the restaurant trick. Once, finding one from a handheld device felt like privileged access to the future. Then it became a background feature. AI agents may follow the same path, passing from spectacle into habit. If Shukla is right, the important story will not be how impressive the agent looked in a demo. It will be the quiet disappearance of a task nobody misses.