Down arrow button icon

Hello. This week I’m in Davos, Switzerland for the World Economic Forum. Tomorrow’s visit by US President Donald Trump is dominating the conversation here. But when people aren’t talking about President Trump and his imposing tariffs on European allies who oppose his attempt to wrest control of Greenland from Denmark, they’re often talking about AI.

The ski town’s promenade is transformed into a technology fairground during WEF, with the logos of prominent software and consulting firms plastered on storefronts and billboards promoting various AI products. But while last year’s Davos was dominated by hype around AI agents and the excessive hand-wringing that the debut of DeepSeek’s R1 model during the 2025 WEF could put the US AI company’s capital-intensive plans to waste, this year’s AI debate appears to be more sober and grounded.

The business leaders I spoke to here in Davos are more focused than ever on how to drive business returns from AI spending. The days of pilots and experiments seem to be coming to an end. The same goes for the age of imagining what AI can do. Many CEOs now realize that deploying AI at scale is neither easy nor cheap. There is now more focus on practical advice for using AI to impact the entire enterprise. (But as you can see, there’s still a tinge of idealism here.) Here’s some of what I’ve heard in our conversations so far.

CEO leads AI implementation

There is a consensus that the bottom-up approaches that were popular in many companies two years ago at the beginning of the generative AI boom (for example, giving all employees access to ChatGPT or Microsoft Copilot) are a thing of the past. At the time, CEOs believed that front-line employees, those closest to the business processes, knew the best way to implement AI to increase efficiency. This turned out to be wrong in the end. Or, perhaps more accurately, the benefits from this tended to be difficult to quantify and, when added up, rarely produced a significant change in sales or profits.

Instead, top-down, CEO-led efforts aimed at transforming core business processes are considered essential to unlocking ROI from AI. Jim Hageman Snabe, chairman of Siemens and former co-CEO of SAP, told a group of his fellow executives during a breakfast discussion I moderated here in Davos today that CEOs need to identify where their businesses deploy AI and become the “dictators” driving those efforts. Similarly, both Christina Kosmowski, CEO of IT and business data analytics company LogicMonitor, and Bastian Nominacher, co-founder and co-CEO of process mining software company Celonis, told me that board and CEO sponsorship is essential to enterprise AI success.

Nominacher learned some other interesting lessons. Among them was research commissioned by Celonis that showed how establishing a center of excellence to find ways to use AI to optimize work processes resulted in eight times more profits than companies that failed to establish such a center. He also said that having data in the right place is essential for successful process optimization.

Race to become the orchestration layer for enterprise AI agents

It is clear that there is competition among SaaS companies for a new interface layer for AI agents working within the enterprise. Workday CEO Karl Eschenbach said he believes the company is well-positioned to become the “gateway to work” because it not only controls critical human resources and financial data, but also already handles employee onboarding, data access and permissions, and performance management. Now you can do the same for AI agents.

But some people are eyeing this award. Srini Tallapragada, chief engineering and customer success officer at Salesforce, spoke about how the company is leveraging its “forward deployment engineers” at 120 of Salesforce’s largest customers to bridge the gap between customer pain points and product development, creating agents tailored to specific verticals and functions, and learning how best to serve Salesforce’s broad customer base. Microsoft Commercial CEO Judson Althoff said the company’s Data Fabric and Agent 365 products are gaining traction among large enterprises that need an orchestration layer for AI agents and a unified way to access data stored in disparate systems and silos without moving to a single platform. Meanwhile, Snowflake CEO Sridhar Ramaswamy believes his company’s deep expertise in maintaining cloud-based data pools and controlling access to that data, combined with its newfound expertise in creating its own AI-coding agents, makes his company ideally suited to compete as an AI agent orchestrator. Ramaswamy told me that his biggest concern is whether Snowflake can continue to move fast enough to realize this vision before OpenAI and Anthropic move lower down the stack, from AI agents to data storage, potentially replacing Snowflake.

Here are some more insights from Davos so far. There are still many concerns about AI leading to widespread job losses, but they are not yet showing up in economic data. In fact, Svenja Gudell, chief economist at job site Indeed, said the tech sector has seen a significant decline in hiring since 2022, but the trend predates the generative AI boom and is likely to be due to companies “rightsizing” after the massive pandemic-era hiring boom, not AI. Many industries are also not hiring much at the moment, which Goodell says is due to global macroeconomic and geopolitical uncertainty, not AI.

Finally, in a comment related to one of the bigger AI news stories of the week, OpenAI is introducing advertising to ChatGPT, Siemens Chairman Snub gave an interesting answer to the question of how AI should be regulated. He said that rather than trying to regulate AI use cases, as in the EU AI law, governments should mandate that AI adhere to human values ​​more broadly. And one of the best regulations to ensure this is to ban advertising-based AI business models, he said. Ad-based AI models will allow companies to optimize user engagement, but all the negative effects on mental health and democratic consent we have seen with social media will be even more severe.

So, here’s more AI news for you.

jeremy kern
jeremy.kahn@fortune.com
@jeremyakahn

Beatice Nolan wrote the news and subsections for Eye on AI.

The fate of AI

Former OpenAI CTO Mira Murati’s departure from $12 billion startup Thinking Machines continues to signal fierce competition for AI talent – ​​by Jeremy Kahn and Sharon Goldman

ChatGPT tests ads as new era of AI begins—Written by Sharon Goldman

Filmmakers deepfaked Sam Altman for a movie about AI. Then things got personal—Written by Beatrice Nolan

PwC’s global chairman says most leaders are forgetting the ‘basics’ as 56% still get ‘nothing’ from AI implementation – Diane Brady and Nick Lichtenberg

AI in news

Focus on AI research

Researchers say ChatGPT has a “silicon gaze” that amplifies global inequality. A new study from the Oxford Internet Institute and the University of Kentucky analyzed more than 20 million ChatGPT queries and found that AI systematically favors wealthy Western regions, rating them as “smarter” and “more innovative” than poorer countries in the Global South. Researchers coined the term “Silicon Gaze” to describe how AI systems view the world through the lens of biased training data, reflecting historical power imbalances rather than providing objective answers. They argue that these biases are not errors to be corrected, but rather structural features of AI systems that learn from data shaped by centuries of uneven information production, privileging places with widespread English coverage and high digital visibility. The team created a website – inequalities.ai –. There, you can explore how ChatGPT ranks your neighborhood, city, or country based on various lifestyle factors.

AI calendar

January 19th-23rd: World Economic Forum, Davos, Switzerland.

January 20th to 27th: AAAI Artificial Intelligence Conference, Singapore.

February 10th-11th: AI Action Summit, New Delhi, India.

March 2nd-5th: Mobile World Congress, Barcelona, ​​Spain.

March 16th-19th: Nvidia GTC, San Jose, California

brain food

Claude Code Anthropic’s ChatGPT Moments? Anthropic started the year with a viral moment that most labs could only dream of. Despite Claude Code’s technical interface, the product has gained traction beyond the developer community, allowing users to build personal websites, analyze health data, manage email, and even monitor tomato seedlings without writing a single line of actual code. The company launched Cowork after several users pointed out that the product was a much more general-purpose agent than the marketing and name suggested. This is a more user-friendly version with a graphical interface built for non-developers.

Claude Code and Cowork’s ability to autonomously access, manipulate, and analyze files on a user’s computer gave many people their first experience with an AI agent that could actually take action on a user’s behalf, rather than simply offering advice. As a result, Anthropic also saw an increase in traffic. Claude’s web viewership has more than doubled since December 2024, and daily unique visitors on desktop are up 12% year-to-date globally compared to last month, according to data from market intelligence firms Similarweb and Sensor Tower. Wall Street Journal. But while some are hailing the product as a step toward having a true AI personal assistant, the announcement also raises concerns about job losses and appears to be putting pressure on dozens of startups that have developed similar file management and automation tools.

FORTUNE AIQ: A year in AI and what lies ahead

From hiring chief AI officers to experimenting with AI agents, companies are taking big steps in their AI journeys in 2025. The lessons learned, both good and bad, combined with the latest innovations in technology will make 2026 another defining year. Explore all about Fortune AIQread our latest playbook below.

–Three trends that have become mainstream Enterprise AI deployment is scheduled for 2025.

–2025 was the year of agent AI. How was it?

–AI coding tools exploded in 2025. The first security exploit shows what can go wrong.

–AI in Business for 2026 Big New Year’s Resolutions: ROI.

–Enterprises are faced with a confusing patchwork of AI policies and rules. Is transparency on the horizon?