Technology executives, industry experts and CNBC journalists gathered in New York on Wednesday for the annual CNBC Technology Executive Council Summit, an invitation-only event for council members. According to CNBC's reporting from the event, the day's discussions centered on how companies are reshaping their strategies as AI advances, how they are managing the costs of building it out, and where regulation, AI agents and workforce readiness fit into the picture.
A recurring theme was that deploying AI tools is only part of the challenge. Speakers argued that the technology delivers results only when the people running it understand the business around it.
Why it matters
The summit arrives as companies pour money into AI while struggling to measure what they are getting back. CNBC reported that AI is producing fundamentally new operating models for businesses, but that AI engineers often lack knowledge of the business they are working in, which complicates efforts to track return on investment and define success.
That gap was illustrated by Nigel Vaz, CEO of Publicis Sapient. Speaking with CNBC's Jon Fortt, Vaz described a breakfast before the summit where he met with a group of the firm's clients. One client, the CIO of an investment bank, recounted bringing in engineers from Anthropic.
According to Vaz, the client said the bank employs 50,000 developers and received six highly capable AI engineers from Anthropic, but those engineers did not understand investment banking or the bank's custom coding language.
Vaz said the anecdote reflects what he hears repeatedly: companies need to consider AI beyond the technology itself, because a tool handed to an enterprise addresses only one piece of the problem. He said organizations need people who grasp the business context — not just how it runs today, but how it will need to run in the future — because someone has to chart the path from the current state to the next one.
Confidence without accuracy
Campbell Brown, founder and CEO of Forum AI, raised a separate concern about how AI output is received. In a June op-ed for The Wall Street Journal, she argued that as reliance on AI grows, its answers need to be scrutinized more closely, writing that the technology has become good enough that its errors no longer resemble errors.
Brown, formerly head of global media partnerships at Meta, said she has been frustrated that the AI labs do not engage enough with genuine subject-matter experts. She said deep human expertise in specific fields is necessary to get this right.
What workers say about AI
Workforce questions were also on the agenda. CNBC reported that its Workforce Executive Council held its own annual summit on Tuesday, where helping employees adjust to AI topped the discussion. Walmart, the world's largest private employer, put that question to its 2.1 million associates in June for the first time, seeking to understand how a highly varied workforce is experiencing the technology.
Maren Waggoner, Walmart's chief people officer for associate experience and digital, told CNBC's Leslie Picker at the WEC Summit that the company received more than 1.7 million responses. The results showed varying levels of familiarity and comfort with AI, and indicated that training and hands-on experience make a difference.
Waggoner said 77% of Walmart's office-based associates reported enthusiasm about AI's potential to improve their work. Among U.S. frontline associates, 84% said they are confident using digital tools, which she attributed to the AI-powered tools many now use on the job. She said the company has provided certificates, training programs and tools, and that associates are applying them in their daily work.
Jill Bruning, chief technology officer of Amentum and a member of the CNBC Technology Executive Council, told CNBC's Susan Caminiti at the WEC Summit that strategy should evolve so that humans remain in charge. She described an AI adoption curve that is climbing quickly, faster than previous technology shifts, and said the distance between where the technology is and where people are keeps widening.
Bruning said she is unsure whether leadership is adapting quickly enough, which adds difficulty to upskilling and to deciding which tasks should be handed to AI. Delegation itself changes, she said, raising questions about how to instruct AI agents and define their roles. Her position: people stay above the loop rather than inside it, and remain in charge.
The TEC Summit's agenda also covered AI regulation, agents, how employees are using the technology, and where the accelerating adoption of AI may lead society. CNBC noted that the event is exclusive to members, with membership details available through its councils site.
What to watch: how quickly companies pair AI deployment with business expertise and worker training — and whether the labs respond to calls for deeper engagement with subject-matter specialists.
