By Adam Mills
AI Editor
0
Share
Newsweek is a Trust Project member
See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.
Enterprise leaders have spent the past few years testing what generative AI can do. Balkrishan “BK” Kalra believes that experimental phase is ending.
“Time of proof of concepts is gone. Time of experiments is gone,” the Genpact president and CEO said during Newsweek’s Sept. 17 “AI Impact Forum” webinar, “The Inflection Point: How Agentic Operations Compound Enterprise Value,” hosted by Dr. Ranjit Tinaikar. “It is now scale use cases.”
Read More on AI
Moving from experiments to scale forces companies to confront problems that a proof of concept can leave unresolved. Data has to be usable, processes need to hold up across business units, employees need enough fluency to work with the tools and someone must remain accountable when an agent acts.
Kalra said scale use cases should ultimately be judged against business performance. He framed the test in three terms: whether companies can “grow faster, operate leaner or convert more to cash.”
Those gains become harder to reach when AI leaves a contained test and enters the systems, data and rules that govern everyday operations.
Tinaikar cited research from Genpact and HFS Research, based on a survey of 2,002 enterprise executives across 16 industries and 14 functions, that found that only 6 percent of organizations qualified as proven debt re-mediators under the study’s framework—meaning they had established initiatives to address those problems, executed them and measured the results. He asked Kalra what senior management teams should address before pushing further into agentic operations.
Kalra described technology debt as the accumulation of old systems, infrastructure and layers of patched-together technology.
“Technology debt is only [the] tip of the iceberg,” Kalra said.
Beneath it, he argued, are data debt, process debt and talent debt, each of which can constrain what companies are able to do with AI.
Agents make those deficiencies harder to overlook because they depend on usable data and company-specific context to act. Business processes introduce another layer of complexity: What looks like a single process from the top may contain different rules, exceptions and practices across countries, states or business units. Kalra’s point is that a general-purpose model cannot reliably navigate those variations without the specific data and process knowledge of the enterprise.
Kalra used global operations to illustrate how requirements can vary by jurisdiction. Much of the knowledge needed to navigate those differences sits inside the enterprise, sometimes undocumented, in the way employees handle exceptions and move transactions from one function to another.
“There is no artificial intelligence, no gains from artificial intelligence, if it is not coupled with process intelligence,” Kalra said.
IT and governance need to enter those conversations early. An audience member asked how companies could build confidence among IT departments when deploying agentic systems. Kalra’s answer began with who should be involved from the start.
“Bring in upfront your CIO or a CDO partner,” he said. “Because if they are not in the tent, they will not buy into the solution early on.”
...
Security concerns also need to be understood and addressed, Kalra said, alongside a responsible AI framework. The issue becomes more consequential as agents take on execution inside finance, supply chain and other operations where mistakes can carry financial or regulatory consequences.
In Kalra’s formulation, agentic operations move from work that is human-processed and human-validated toward processes that are increasingly machine-processed and human-validated. People retain responsibility for exceptions and accountability even as agents perform more of the underlying execution.
Workforce readiness is equally important. Kalra said workers need enough exposure to AI tools to understand what is possible before they can participate meaningfully in redesigning how work gets done. At Genpact, he noted, thousands of employees have been given access to those tools.
Tinaikar connected the point to the time companies give workers to develop new skills. Training has traditionally competed with urgent day-to-day responsibilities, he said.
“It is not a privilege, it is an imperative,” Tinaikar said.
Kalra uses two broad categories for the skills Genpact is developing. AI builders combine technical expertise with knowledge of a relevant business domain. AI practitioners begin with deeper expertise in areas such as finance, supply chain, procurement, banking or insurance but are expected to develop enough command of AI and data to work directly with the technology.
Some tasks and roles will change significantly as machines take on more execution, Kalra said. Tinaikar raised the smartphone era as an example of how predictions of displacement can miss what a technology creates around it.
“When the mobile phones came, they thought the laptops and the PCs would become obsolete,” Tinaikar said. “Instead, it created a whole new ecosystem of apps on mobile phones.” He noted that new businesses and jobs grew around that ecosystem.
Kalra agreed with the broader point.
“When new technology comes in, new business models come in, new roles come in,” he said.
Kalra also invoked Jevons paradox—the idea that greater efficiency can lead to more overall use, creating additional demand and activity.
His more immediate warning was for workers.
“Your job will not be taken by AI, but your job can be taken by somebody who knows AI better,” Kalra said.
More capable models can expand what companies are able to automate, but scaling them still requires usable data, workable processes, security controls and employees prepared to operate differently.
Kalra distilled the problem as the conversation closed.
“Aspirations are really high,” he said. “Readiness is low.”
We would love for you to attend our next edition of our "AI Impact Forum" webinar series.
On Thursday, October 22, Dr. Ranjit Tinaikar will sit down with Firdaus Bhathena, executive vice president and chief technology and transformation officer at S&P Global, for a conversation about enterprise technology, AI and the organizational changes required to turn new capabilities into business results.
Register for the webinar, for free, today.
View 0 Comments
Submit Correction
View Editorial & AI Guidelines