NEW YORK - Autonomous "agents" and profitability are likely to dominate the artificial intelligence agenda next year, business executives and researchers predicted this week in interviews at the Reuters NEXT conference in New York.

Agents, or systems that can perform actions like making purchases and scheduling meetings on people's behalf without their direct involvement, have long been an elusive goal for AI researchers.

However, such capabilities were likely to be enabled by the emergence this year of step-by-step reasoning approaches like those used in OpenAI's o1 model, the executives told Reuters.

"I think we are going to see a lot of motion next year around agents, and I think people are going to be surprised at how fast this technology comes at us," said OpenAI CFO Sarah Friar.

"We think that's just the beginning of what 2025 will be about: agents who are really there to help you with day to day tasks," she said.

Friar, who joined Microsoft ( MSFT )-backed OpenAI six months ago, also forecast that artificial general intelligence (AGI) - a threshold where autonomous systems surpass humans in most economically valuable tasks - was likely to be achieved in the coming few years. 

Asked if that would be closer to two years or a decade, she predicted it would be "in the shorter term."

"I don't believe it's a decade away," said Friar.

George Mathew, managing director at venture capital firm Insight Partners, said that the reasoning capabilities discovered in the last year were already starting to pay dividends.

"We're starting to see a full reimagination of entire back-office functions, as well as front office functions," said Mathew. 

He cited as an example a company he invested in called Relevance AI, which offers digital sales teams that can replace human workers at a quarter of the labor cost.

With those kinds of productivity gains in mind, another venture capitalist, Molly Alter at Northzone, predicted that 2025 would be "the year of profitability for AI."

"Last year it was more focusing on revenue growth," Alter said. "This year, it's going to be, instead of all about grow, grow, grow, it's going to be, 'wow, if I use AI, my margins are going up.'"

Outside the world of tech startups, executives from major banks and other businesses said they had moved beyond experimentation and integrated the technology into their workflows, too.

Thousands of employees at bank BNY were now enabled to build and commission LLM-fueled agents to help them with their daily tasks, said CEO Robin Vince. LLM refers to large language models, a type of AI.

"We're investing in and we have tools live that create more leverage for our people, that create solutions for clients, and insights and data for clients that otherwise would have been almost impossible to derive through traditional means," said Vince.

To view the Reuters NEXT news page, go to: https://www.reuters.com/world/reuters-next/