Everyone's Talking About AI Agents. Barely Anyone Knows What They Are.
'When I hear some of the conversations around agentic, I sometimes wonder whether it's like that old elephant thing? Everybody's touching a different part of the elephant,' said Prem Natarajan, chief scientist and head of enterprise AI at Capital One. 'Their description of it is different.'
AI agents are broadly understood to be systems that can take some action on behalf of humans, like buying groceries or making restaurant reservations.
But in some cases, the question of what constitutes an 'action' is blurry. (Is querying enterprise data and delivering an answer based on it an 'action'? In some cases it might be and in other cases it might not).
Further, not all software actions are considered agentic.
For example, if AI is simply taking an action based on specific details provided by a human user, it isn't agentic, said Tom Coshow, senior director analyst with Gartner's Technical Service Providers division. Software needs to reason itself and make decisions based on contextual knowledge to be a true agent, he said.
Gartner held an AI agents webinar earlier this year to explain the technology and discuss use cases, Coshow said. Afterward, participants were polled on whether they had ever deployed agents. Only 6% said yes.
A lot of what companies are calling AI agents today are really just chatbots and AI assistants, he said.
'It's a conversation that comes up even internally, as you can imagine,' said Capital One's Natarajan.
Capital One recently deployed its first AI agent use case, a concierge chatbot for its car-dealership clients to help consumers get information on what cars are available and schedule test drives.
On the surface, a user interacting with the concierge might feel like they are having a similar experience to interacting with a chatbot like OpenAI's ChatGPT or Anthropic's Claude. They ask questions and get responses. What truly makes Capital One's use case agentic, according to Coshow, is that there is a large language model making decisions about what content to serve the customer and the bot can also take action to schedule a test drive.
'Keeping the definition of AI agents simple would be: Does the AI make a decision and does the AI agent take action?' Coshow said.
Still, Robert Blumofe, chief technology officer at Akamai Technologies, said many of the use cases he is seeing in the wild resemble 'assistive agents,' rather than 'autonomous agents, requiring direction from a human user before taking action and narrowly focused on individual use cases.'
'You could argue that 'assistive agent' is a bit of an oxymoron,' Blumofe said.
Ori Goshen, co-founder & co-CEO of AI21 Labs said he shies away from the term. 'It's become very, very overloaded,' he said.
'There are many different things that are folded into this term,' he said. And when it comes to helping enterprises understand what they are spending money on and truly deliver on the value of AI, 'it just needs to be discussed more precisely,' he said.
Write to Isabelle Bousquette at isabelle.bousquette@wsj.com
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
New York Post
7 minutes ago
- New York Post
Ex-Google exec's shocking warning: AI will create 15 years of ‘hell' — starting sooner than we think
A former Google executive warned that artificial intelligence will plunge society into more than a decade of severe disruption and hardship as it eliminates many white-collar jobs — and the 'hell' will begin as early as 2027. Mo Gawdat, who left Google X as its chief business officer in 2018 and has become a popular author and public speaker, painted a grim picture of widespread job losses, economic inequality and social chaos from the AI revolution. 'The next 15 years will be hell before we get to heaven,' Gawdat told British entrepreneur Steven Bartlett on his 'Diary of a CEO' podcast on Monday. 4 Mo Gawdat, a former Google executive, warns that AI could trigger over a decade of upheaval, wiping out white-collar jobs and fueling social unrest. YouTube / The Diary Of A CEO Gawdat, 58, pointed to his own startup, which builds emotional and relationship-focused artificial intelligence. It is run by three people. 'That startup would have been 350 developers in the past,' he told Bartlett in the interview, first reported by Business Insider. 'As a matter of fact, podcaster is going to be replaced.' Gawdat specifically warned that 'the end of white-collar work' will begin by the late 2020s, representing a fundamental shift in how society operates. Unlike previous technological revolutions that primarily affected manual labor, he argues this wave of automation will target educated professionals and middle-class workers who form the backbone of modern economies. The Egyptian-born tech whiz, who was a millionaire by age 29, believes this massive displacement will create dangerous levels of economic inequality. Without proper government oversight, AI technology will channel unprecedented wealth and influence to those who own or control these systems, while leaving millions of workers struggling to find their place in the new economy, according to Gawdat. Beyond economic concerns, Gawdat anticipates serious social consequences from this rapid transformation. 4 Gawdat says rapid advances in AI technology will soon threaten even highly skilled professions once thought immune from automation. Nina Lawrenson/ – Gawdat said AI will trigger significant 'social unrest' as people grapple with losing their livelihoods and sense of purpose — resulting in rising rates of mental health problems, increased loneliness and deepening social divisions. 'Unless you're in the top 0.1%, you're a peasant,' Gawdat said. 'There is no middle class.' Despite his gloomy predictions, Gawdat said that the period of 'hell' will be followed by a 'utopian' era that would begin after 2040, when workers will be free from doing repetitive and mundane tasks. 4 The rapid advancements in AI have been demonstrated in products such as OpenAI's ChatGPT. Ascannio – Instead of being 'focused on consumerism and greed,' humanity could instead be guided by 'love, community, and spiritual development,' according to Gawdat. Gawdat said that it is incumbent on governments, individuals and businesses to take proactive measures such as the adoption of universal basic income to help people navigate the transition. 'We are headed into a short-term dystopia, but we can still decide what comes after that,' Gawdat told the podcast, emphasizing that the future remains malleable based on choices society makes today. He argued that outcomes will depend heavily on decisions regarding regulation, equitable access to technology, and what he calls the 'moral programming' of AI algorithms. 'Our last hurrah as a species could be how we adapt, re-imagine, and humanize this new world,' Gawdat said. Gawdat's predictions about mass AI-driven disruption are increasingly backed by mainstream economic data and analysis. 4 Dario Amodei, CEO of Anthropic, has warned of a 'white-collar bloodbath.' AP Anthropic CEO Dario Amodei has warned of a 'white-collar bloodbath,' predicting that up to half of all entry-level office jobs could vanish within five years. The World Economic Forum says 40% of global employers expect to reduce staff due to AI, and Harvard researchers estimate that 35% of white-collar tasks are now automatable. Meanwhile, Challenger, Gray & Christmas reports that over 27,000 job cuts since 2023 have been directly attributed to AI, with tens of thousands more expected. Goldman Sachs and McKinsey project a multi-trillion-dollar boost to global GDP from AI, but the IMF cautions that these gains may worsen inequality without targeted policy responses. Analysts from MIT and PwC echo Gawdat's fears of wage collapse, wealth concentration, and social unrest — unless governments act swiftly to manage the transition.
CNBC
8 minutes ago
- CNBC
How GE Vernova found itself in the middle of the frantic race to build out AI
The AI boom has presented energy stalwarts like GE Vernova with a monumental task — and opportunity: Make enough gas turbines to support an entire technological revolution. As tech giants invest billions of dollars to build AI data centers, a shortage of power has emerged as a new and unexpected bottleneck. In the initial wake of ChatGPT's launch in late 2022, the main supply crunch was for Nvidia's graphics processing units, which has moderated. Now, the biggest demand is for energy, putting GE Vernova, one of the world's largest producers of gas-fired turbines used to create electricity, at the center of a frantic race to build out generative AI infrastructure. Gas turbine orders from utilities, independent power producers, and industrial players are on a record-breaking pace for 2025, according to a S & P Global Commodity Insights. Think of these turbines as huge modified engines located at power plants that burn natural gas to make a lot of electricity very quickly. Manufacturing them is a complex process, and right now supplies are tight. "You may have trouble getting GPUs," Jim Cramer said. "Try getting power from GE Vernova." The fervent demand for GE Vernova's turbines is leading to a financial windfall, as its recent blowout quarterly earnings release made clear. So strong was the report that, in response, Jim described the industrial stock as "maybe the best story in the entire market." Shares of GE Vernova have doubled year to date and are on pace for a fresh record close Monday around $662 a share. Following earnings, the Club raised its price target by $150 on GE Vernova to $700 a share. Orders for its power segment rose 44% organically last quarter as gas power equipment orders, in particular, increased nearly threefold. Meanwhile, a measure of multiyear demand that includes both gas-turbine orders and "slot reservations," which will be fulfilled further into the future, increased to 55 gigawatts, up from 50 in April. By year-end, GE Vernova expects that to be at least 60 gigawatts — for context, if the Hoover Dam were running at max capacity, it could produce roughly 2 gigawatts of power . This "era of accelerated electrification is driving unprecedented investment," CEO Scott Strazik said on the conference call. GEV YTD mountain GE Vernova's year-to-date stock performance. The unprecedented investment is also testing the limits of the power industry. Demand for GE Vernova's turbines — and those from peers like Siemens Energy and Mitsubishi Power — are so large that it's hard for them to keep pace. GE Vernova's gas turbines are "largely sold out" for 2026 and 2027, while the company is "approaching filling out '28 and starting to sign agreements for later years," Strazik said in an earnings call this spring. In other words, customers placing new orders for these heavy-duty gas turbines can face wait times that extend several years. With strong demand and limited supply, turbine prices have jumped in recent years as well. The reason there's so much need for these turbines is simple: the U.S. needs a lot more electricity to keep up with the demand of AI computing. Club holdings Amazon , Microsoft and other hyperscalers seemingly cannot spend enough money to construct more power-hungry data centers. Alongside earnings last week, Microsoft and Meta reported massive capital expenditures on the back of increased AI investments and signaled that more outlays are on the way. "This is the number one driver of new [turbine] demand growth, especially in the near term," said Sam Huntington, director of North American power research at S & P Global Commodity Insights. "For decades, we had this long period of flat [and] minimal load growth, right up to the past couple of years, and now, the situation has changed again," he told CNBC. Bank of America analysts in a July note forecasted that U.S. electricity demand will grow at a 2.5% compound annual growth rate through 2035, up from half a percent between 2014 and 2024. Many energy executives say more natural gas will be necessary to meet the electricity demand, given some current limitations around solar and wind. Even though projections point to more turbine demand, expanding manufacturing capacity isn't an easy fix. These turbines are massive and complex machinery that can take years to get to customers , so management teams want to make sure that this AI data center buildout is durable before they commit to dramatically scaling up their own production. "They have to balance not building enough to meet data center electricity demand in the future versus making irreversible investment decisions today that might end up costing companies a lot of money over the next 30 years if demand doesn't show up," said Ramteen Sioshansi, an engineering professor at Carnegie Mellon University who leads its Electricity Industry Center. The industry doesn't want a repeat of the early 2000s when there was a similar hype cycle around natural gas power plant construction that ended poorly. Lower natural gas prices — coupled with growing environmental concerns that led to more interest in cleaner alternatives to coal — meant a boom in demand for these turbines at the time. But when gas prices rose again and the collapse of Enron made it harder for power producers to finance new plant construction , demand for gas turbines dried up, leaving manufacturers including GE Vernova — then part of the General Electric conglomerate — with canceled orders and excess inventory. "This led to a huge boom and bust market and people lost a lot of money," S & P Global's Huntington added. "So, there's still scars of 20 years ago [that] are relatively fresh." Plus, in the years leading up to the AI boom, major investments into renewable energy sources like wind and solar around the world also disincentivized turbine manufacturers from expanding their production capacity. That context helps explain why companies like GE Vernova may not want to open up the check book and spend wildly to expand output. At the same time, GE Vernova must maintain a healthy production pace while keeping its customers happy. "I think there's a sense at [GE Vernova] that, 'We're winning the deals that we want to win and, and we don't see a need to potentially risk over-capacitizing the market chasing demand,'" RBC Capital Markets analyst Chris Dendrinos told CNBC. Still, Dendrinos said GE Vernova was "an early mover in terms of expanding capacity" in the industry. In January, GE Vernova announced a $600 million investment in U.S. factories over the next two years to address the growing need for electricity. Of that capital, $300 million is specifically going to support the buildout of its gas turbine business by expanding facilities. It is expected to create more than 850 new jobs. Dendrinos, who has covered the energy sector for over a decade, described the $600 million investment as "absolutely material" and "a testament to the demand" for its turbines. GE Vernova's peers are ramping up production, too. On a February earnings call, Siemens Energy CEO Christian Bruchs said the company is working to grow production capacity by roughly 30%. "However, we take a measured approach and expand our capacities in a responsible way, neither losing sight of cost nor the risk of adding too much capacity," the executive said on the call. Along with the committed capital, GE Vernova has announced deals to expedite production and promoted alternatives to its heavy-duty turbines to maintain its dominance in the industry and please customers. In March, for example, the firm announced an acquisition of Woodward's heavy-duty gas combustion parts business for an undisclosed amount. The move is intended to help the company "meet growing energy demands from load growth," according to a press release. By buying this business, it brings manufacturing of essential combustion parts for these gas turbines in house, which gives GE Vernova more control over its supply chain. Meanwhile, GE Vernova has also touted temporary alternatives to its heavy-duty turbines such as as its "aeroderivative" units. These are mostly preassembled, have faster start times and take up less space than their heavy-duty counterparts. On the other hand, aeroderivatives are less efficient at scale for data centers. GE Vernova announced in March that the company secured an order for three of these units for a municipal utility in Missouri. Overall, GE Vernova is a stellar position to capitalize on the data center buildout. However, Jim said the company could afford to further expand production in order to meet booming demand. "[Their] orders are so strong and ironclad, I wish [they'd] spend some more," he said. (Jim Cramer's Charitable Trust is long GEV, NVDA, AMZN, MSFT. See here for a full list of the stocks.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust's portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
Business Insider
8 minutes ago
- Business Insider
'No names': A brief moment in an interview with OpenAI engineers highlights the state of the AI talent wars
An OpenAI engineer said some of the company's "most-prized" talent is focused on debugging AI models — just don't expect him to mention their names. OpenAI chief data scientist Jakub Pachocki and Szymon Sidor, an OpenAI technical fellow, recently appeared on the "Before AGI" podcast and chatted about working at the company. A brief exchange in the episode stands out for what isn't said — and is an indicator of just how protective AI companies have become amid the AI talent wars, as Big Tech circles star employees. "We hired a bunch more people at OpenAI who are really great at debugging," Sidor said while speaking about the importance of debugging AI models. "And I think those are some of our most-prized employees, and I won't even…" Before Sidor could complete his thought (he mentions not going into "details"), another person on the podcast interjects by saying, "No names," before laughter can be heard. It's unclear who jumped in with "No names" — it was either OpenAI's Pachocki or podcast host Aleksander Mądry. While it would typically be easy to figure that out by watching the video version of the podcast, that part of the exchange is absent in both of the video versions published to YouTube and X. (You can hear that bit in the audio-only version on Spotify and Apple podcasts.) Business Insider reached out to Madry and OpenAI for comment. Mądry, the host of "Before AGI," is an MIT professor who is working at OpenAI while on leave from the university, where he is director of the MIT Center for Deployable Machine Learning. Regardless of why the exchange isn't found in the video version of the podcast (it's entirely possible the video version was simply edited down for length or flow or some other reason), Sidor's apparent reluctance to name-drop those employees he feels are some of OpenAI's "most-prized" is telling — if not particularly surprising. After all, Mark Zuckerberg reportedly created a list of names of AI stars to poach for Meta's Superintelligence Lab, and successfully hired away Shengjia Zhao, a co-creator of ChatGPT and former lead scientist at OpenAI, and Scale AI founder Alexandr Wang, among others. Sam Altman said earlier this year that Meta was offering his company's top researchers up to $100 million compensation packages, and Google recently hired away Windsurf's CEO and multiple employees who had been set to join OpenAI. It's an all-out fight over top AI talent — so it's understandable that an OpenAI engineer might want to be tight-lipped.
