Salesforce is leading the software industry’s attempt to collect fees from businesses that use Claude and other external AI tools to automate work in enterprise apps.

As soon as next month, Salesforce plans to require companies to register outside AI agents, a step toward charging them each time those agents update or analyze data stored in Salesforce. The fees let Salesforce earn money from customers who use such AI instead of paying for Salesforce’s built-in AI products. 

The Salesforce tollgate has implications for the rest of the industry, which is coming to terms with the idea that so-called AI agents may be able to use software firms’ apps more efficiently than humans do, meaning their seat-based or per-user pricing model might not have staying power. The question is whether the new AI tollgates—which Atlassian, Workday, ServiceNow and Datadog are also erecting to various extents—and other pricing shifts will more than make up for an eventual decline in per-seat subscription revenues. 

The vast majority of workers haven’t yet outsourced such work to AI, of course, and companies like Salesforce are trying to get more revenue from them as well. They charge fees based on how much they use the apps’ built-in AI tools and are also experimenting with outcome-based pricing, meaning customers only pay when such tools actually complete a task successfully or help the customers make more money.

For now, Salesforce and some other software firms developing tollgates for customers’ AI agents haven’t started charging the tolls. They’re letting customers experiment for free with the technology a little while longer. (The customers still have to pay providers such as Anthropic to use the AI agents in the first place.)

Starting around November, according to the company’s website, Salesforce will require customers to register any externally-made AI agents if they plan to use them to take actions on their behalf in Salesforce apps. Each action an agent takes in Salesforce counts as a billable “interaction,” according to the company. The move has lifted investor confidence in Salesforce recently.

Salesforce describes the registration requirement as a security measure because the company gives each agent its own identity and narrower data-access permissions, compared to humans. It isn’t clear how Salesforce will clamp down on customers using their own AI agents, such as OpenClaw-powered ones that run on a person’s computer and can access any app that runs on it. (Salesforce did not respond to requests for comment.)

Salesforce hasn’t announced prices tied to such agent usage, but customers could eventually end up paying a lot more money than they do now, given how many of them are already using their own AI on top of Salesforce apps and bypassing Salesforce’s own built-in AI tools, according to software consultants and chief information officers. 

Nick Hamm, a software consultant, said that most of the 150 enterprises he works with are currently using AI from Anthropic and OpenAI to automate work inside Salesforce apps, for free. But they’re reconsidering that usage in anticipation of the upcoming fees. If the fees are too high, the customers may move their business data outside of Salesforce’s apps so they can use AI to analyze the data without paying an extra cost, he said.

New customers will have a three month grace period after the new pricing is announced to register their AI agents, and the pricing won’t affect existing customers until their contracts are renewed. 

In addition to the billable “interactions,” customers have the option to buy unlimited agent usage by purchasing a “headless license,” a Salesforce account executive told Hamm in a LinkedIn message viewed by The Information. 

Some customers such as Lumen, a fiber-optic network provider to enterprises and data centers, say they are waiting to test Salesforce’s new connectors until pricing is disclosed.

So far, Lumen has leaned on Salesforce’s built-in AI. A longtime customer of Salesforce’s customer relationship management software, Lumen made a “major” investment in using Salesforce’s AI sales agents in the last 12 months, according to CIO Jim Fowler. The sales agents help with tasks such as collecting customer information and generating sales quotes for the customers to buy fiber connections between data centers, saving Lumen’s human salespeople considerable time, he said.

So far, Fowler said, Salesforce has been “super flexible” about letting Lumen use existing subscription fees to also cover its use of built-in AI features such as the sales agents.

Early customer reactions to the pricing announcement will likely determine how flexible Salesforce is with customers that start renewing their contracts after November. If its past AI pricing efforts are any guide, Salesforce could easily switch things up depending on the feedback, or any blowback, it receives.

New From Our Reporters

Exclusive

Datadog Offered $4 Billion to Buy Cloud Security Startup ‘Upwind’

By Aaron Holmes and Valida Pau

Exclusive

Andreessen Horowitz to Drop ‘Crypto’ From Name of Flagship Blockchain Fund

By Leo Schwartz

Exclusive

Hackers Used Claude and Grok to Automate Attacks on Banks

By Aaron Holmes and Cory Weinberg

As companies create more applications using AI, corporate officers say demand is rising for a new type of software that stress-tests those apps, simulating how they handle network problems and other disasters that humans may not think to test for—including cyberattacks by rogue AIs.

Luis Chávez, CTO of coding assistant provider Replit, says his company uses an AI-assisted software known as a deterministic simulator, a kind of flight simulator for software. The deterministic part refers to how the simulator lets companies run the same disaster test again and again so they can try out different fixes for bugs that crop up during network failures or hacking attempts.

Traditional stress-testing methods rely on developers guessing every potential problem that may arise. Merely using today’s AI models to evaluate code often produces false alarms or fake warnings, corporate executives say.

While the AI industry ties itself in knots over models it thinks will pose security threats to the internet and perhaps to humanity itself, researchers are warning about flaws in AI coding agents from Anthropic, OpenAI, Google and Microsoft that pose immediate risk to businesses.

For instance, a recently discovered (and now largely fixed) software flaw in Claude Code, Codex, Gemini CLI and GitHub Copilot would have let attackers hijack other people’s coding agents without them noticing, according to new research shared exclusively with The Information.

The research, published by Sequoia Capital-backed cybersecurity startup Air, discovered the same flaws in the way all four of those coding agents handle “skills” that direct agents to refer to a set of instructions and files to execute specific tasks.

Major Anthropic customers are still waiting to get “zero data retention” for the newest Fable AI models, meaning the AI firm would not store data related to their usage of its models. In the meantime, some are turning to OpenAI. 

Palantir and Booz Allen, for instance, are expanding their use of OpenAI’s Astra model, which they’ve been able to secure with zero data retention. Both have restricted their use of Fable models because Anthropic said in June it would store customer data for those using Fable models for 30 days, for security reasons.

Before that June announcement, Anthropic offered zero data retention policies in deals negotiated with individual customers. Starting later this fall, the AI firm plans to roll out a new program called Enterprise Frontier Safeguards, which effectively gives “eligible” customers zero data retention. 

Salesforce seems unafraid of a future in which people use an AI chatbot like Claude to access Salesforce enterprise apps and underlying databases to get work done, rather than visiting those apps themselves. Figma, an enterprise design app, naturally wants a different future.

Dylan Field, Figma’s CEO, said Wednesday that as the company’s in-house AI tools for designing apps and other user interfaces get better, “we'll see more and more of that work happening in Figma directly,” rather than using Claude or other prominent AI assistants to do such work.

Nvidia is doubling down on its open-source efforts, pouring investment into an ambitious new in-house AI model that it hopes will drive demand for its hardware—but that could also compete with customers and partners.

Nvidia is aiming for the new model, the biggest of a family that it is calling Nemotron 4, to be on par in its performance with the best open-source AI models in the world, according to several people who work on Nemotron. The research paper detailing Nvidia’s last big model had 570 authors, and employees said Nemotron 4 would involve even more people. “Everyone just wants to be involved at this point,” one former employee said.

Salesforce on Wednesday told state government agencies in Washington and California that it plans to permanently layoff 133 employees at the start of October

Some 74 employees will be cut in San Francisco, California, the company’s headquarters, while 59 will be laid off in Bellevue and Seattle, ...

Welcome back!

Software companies have been at pains to flaunt the success of their AI products, as they seek to dispel investor fears that AI apps are threatening their existence. Two reports this week show that developing popular AI products is only half the challenge. 

Canva, the design software company popular among students and business owners, warned investors this week that its annual revenue growth would slow to 20%. The reason: it was tapping the brakes on a planned AI rollout expected to drive paying subscriptions—after demand had proved higher and more expensive to serve than anticipated. 

Then on Thursday, Figma said it expects a slowdown in revenue growth for the September quarter, to 36% from 48% in June. In its case, the publicly traded design software company suggested it’s still figuring out how to generate sales from new AI tools that are still in beta, or testing phases, or early in their release. 

Atlassian shares soared as much as 38% in after-hours trading on Thursday after the enterprise software company reported its first quarterly operating profit in more than two years. Atlassian CEO and cofounder Mike Cannon-Brookes said he planned to start buying as much as $250 million in shares, ...