LATEST / 22 OCT 2026
JAPAN · Hasegawa discusses AI for workflow design, human judgment and RPA executionTHE LONG VIEW · Banking operations → UiPath Japan → business orchestration

People / The unfinished office

Koichi Hasegawa and the work computers forgot

He left a banking career to build UiPath in Japan with a team of three. His enduring question is why people still spend their days finishing the work their software leaves behind.

Advertisement

Three people in a rented Tokyo office is an unusual destination for someone who has been chief operating officer of a major bank’s Japanese business. In early 2017, that was where Koichi Hasegawa had arrived. He had left Barclays and joined UiPath. The work ahead included hiring a team and persuading Japanese companies to let software robots take over some of their repetitive tasks. A very long corporate career had led to a very small beginning.

The attraction was a familiar irritation. Companies bought computer systems, yet people still had to finish the job. Between applications, departments and formats sat a quantity of manual work that resisted the grand promises of information technology. Hasegawa had spent years close enough to those systems to see the leftovers. He was ready to try selling a remedy.

A lawyer’s education, a technologist’s working life

Hasegawa studied law at Keio University, graduating in 1983. That year he joined Arthur Andersen, the business identified in his career accounts with today’s Accenture. In 1993 he moved to Goldman Sachs. His time there included work in New York, Hong Kong and London. Deutsche Bank followed in 2000, with responsibilities as Japan CIO and Asia-Pacific fixed-income CIO.

In 2005 he joined Barclays. His roles encompassed Asia-Pacific technology, global outsourcing strategy and Japanese operations. The sequence matters because it placed him on several sides of a corporate software decision. Technology leaders commission systems. Operations leaders live with what happens after installation. Outsourcing leaders see how work travels when a company sends it elsewhere.

His eventual move into automation drew on that experience. There is a particular education in watching an expensive system arrive and then discovering the small chores it has generously preserved for its users. The people making the purchase and the people doing those chores may have rather different reviews.

The last mile, done by hand

Hasegawa first heard about robotic process automation around 2005 or 2006, through offshore automation projects in India. His hands-on encounter came in 2016 at Barclays Securities Japan. He saw a way to tackle what he called the last mile: the remaining manual tasks that larger systems had failed to absorb.

He could have pursued that idea inside the bank. Instead, he wanted to try it across Japanese business. He also had a view about why foreign software sometimes struggled to gain acceptance in Japan, having watched products arrive over more than two decades in international finance.

“I thought it would be best to try my own hand at the software game.”

Koichi Hasegawa, 2018

UiPath appealed to him through its engineering culture, product architecture and usability. Its co-founder Daniel Dines had an engineering background, another attraction. Hasegawa was choosing both a technology and an organization through which to put his accumulated opinions to work.

A career changes direction
  1. 1983Arthur Andersen
  2. 2005Barclays
  3. 2017UiPath Japan begins
  4. 2025Returns as chairman & CEO

Japan was part of the product

He joined UiPath in January 2017; the Japanese subsidiary was established in February. By September, he was explaining the opportunity to prospective recruits. He wanted people who could listen carefully to customers, understand what needed doing and carry it through. He also wanted an entrepreneurial willingness to help build the business.

One of his examples was almost comically recognizable: a new employee joins, and the same information must enter the building-access system, personnel records, payroll, career-management tools and IT access arrangements. The employee has arrived once. The paperwork enjoys several encores.

Hasegawa argued that Japan’s requirements could improve UiPath’s software more broadly. Customers expected detailed service and accuracy. Meeting those expectations would help prepare a product for other markets. Local demands were therefore part of his case for investing in Japan, with consequences for engineering and customer support as well as sales.

Koichi Hasegawa gesturing during an interview, wearing a jacket over a UiPath T-shirt
The office robot has a human spokesman. Hasegawa in conversation with Workflow Research Institute. Photograph: ATLED.

The small print of a working day

He developed that argument through the character of Japanese office work. In 2019, he described many processes as small in volume, complicated and varied, with repetitions that included branches and exceptions. That was a different challenge from running a vast quantity of identical tasks through an offshore operation.

The distinction helps explain his recurring interest in people at the front line. They know the exceptions. They know which customer needs a particular format, which procedure contains an extra approval and which screen requires an apparently unnecessary step. A software design that overlooks those details leaves them to repair the process manually.

Hasegawa wanted employees to become capable users of automation. He described work on accessible tools, shared know-how and education, with adoption reaching smaller businesses and regional organizations. A robot might be built by a specialist, but its usefulness would depend on the person trying to get through the working day.

A future with less washing

For all the technical language around his business, one of his explanations starts with laundry. In a conversation with workflow specialist Yasuhiro Okamoto, Hasegawa reached for the washing machine: a device that absorbed repetitive effort and became so familiar that people stopped treating it as a technological marvel.

He hoped office robots could become similarly ordinary. The point was to let people use them with the ease of a familiar appliance or smartphone. His phrase was “A Robot for Every Person.” It is a useful expression of ambition because its success would make the technology less remarkable in everyday conversation.

He also left the next decision with people. Automation can release time; employees and managers must choose how to spend it. That places a responsibility beside the sales proposition. Saving minutes is measurable. Turning those minutes into a more worthwhile job requires decisions about the job itself.

The lesson he wanted to travel

There was a social purpose behind his interest in those decisions. The 2011 Tohoku earthquake and tsunami prompted volunteer work with friends, and he looked for a way to make a wider contribution through his technology experience. His 2019 TEDxOtemachiED talk, Digital robots for everyone, connected automation with learning and the development of digital skills.

By the time of that talk’s published biography, the Japanese team numbered 300 and supported more than 1,200 clients. Those figures belong to 2019, a snapshot of the distance from the three-person start. They also gave his educational argument a practical audience: the people inside organizations beginning to use the tools.

At an education seminar that September, he addressed developing people who could work with robots. The event brought together education policy and practical examples from Waseda University and Ibaraki Prefecture. His attention extended to the administrative work surrounding education, where repetitive procedures consume time that professionals might spend elsewhere.

1,200+Japanese clients described in his 2019 TEDx speaker biographyA historical milestone, not a current customer count.

He preferred the second career

Five years after taking charge of UiPath Japan, Hasegawa was asked which work he found more interesting: his previous industry or the present one. He chose the present without hesitation. Solving problems through automation and seeing customers pleased with the result gave the role its appeal. So did the speed of change in technology.

By 2022, he was discussing how automation could reach across an entire company, and why partners mattered to that effort. He continued to emphasize understanding Japanese customers’ circumstances, including their use of both cloud and on-premises systems. The practical details of an organization’s existing environment still counted.

There is a continuity in that preference. His earlier roles had concerned the relationship between technology and operations. His new business let him work on that relationship for customers, with the satisfaction of seeing a problem resolved. The scale of his employer had changed dramatically in 2017; the subject holding his attention had remained close to home.

Back in charge, with a different toolkit

On September 1, 2025, Hasegawa returned to lead the Japanese business as chairman and CEO. UiPath also made Japan a region reporting directly to headquarters, alongside the United States and International. Dines welcomed his return and credited lessons from Japanese customers with contributing to the company’s growth globally.

Hasegawa’s stated task now included combining AI agents with RPA, bringing flexible task handling together with reliable execution. Trust was central to his explanation. The software vocabulary had expanded considerably since his early encounters with automation. The requirement to make a process work safely and consistently was familiar territory for a former banking operator.

In August 2026, he described orchestration as an enterprise-wide approach to optimizing processes. Departmental automation could leave processes fragmented, he argued. Connecting them required a broader view, together with developers who understood business problems and could turn new coding tools into useful working systems.

The handoff is still the story

His October 2026 discussion gave that broader view a specific division of work. AI could help understand systems and design workflows. People would judge the proposed process. RPA and other tools would execute it. He again pointed to Japan’s complicated, branching repetitions as a problem worth addressing, with AI now helping with the design.

That returns the story to the work between systems. A useful suggestion needs a route into execution. An automated task needs a connection to the next task. A process needs somebody to decide whether its result is acceptable. Buying a clever tool leaves those questions on the desk.

Hasegawa’s career offers a sustained interest in that desk: first from inside financial institutions, then from a small Tokyo office, and now with AI added to the toolkit. The promise he keeps returning to is time recovered for the person doing the work. The next worthwhile question is what that person gets to do with it.

Keep following the work