The next major battleground in artificial intelligence is expected to be agent management software—tools designed to help businesses manage a growing number of AI agents from various providers. In this arena, established enterprise software companies believe they hold an advantage over startups like OpenAI and Anthropic.

This advantage stems from the traditional vendors' deep understanding of their clients' business processes, which industry executives consider difficult for AI startups to match. The gap is particularly evident in software used for routine workflows such as invoice processing and new employee onboarding.

This week, interviews with executives from ServiceNow and Atlassian highlighted their extensive experience in the field, emphasizing that such industry knowledge is crucial for AI agents to function effectively.

The focus on agent management software intensified following OpenAI's preview release last week of a product called Frontier. OpenAI's announcement suggested that Frontier could become a primary way for enterprises to use software: essentially directing AI agents through a ChatGPT interface to utilize various third-party applications for tasks in HR, IT, sales, and more.

Several software industry sales personnel indicated that many established enterprise software vendors view Frontier as a signal of OpenAI's intent to seize their direct relationships with customers.

Currently, Frontier is available only to a select group of large clients, so it remains unclear how aggressively OpenAI will promote the product to customers using its models and other offerings. However, cloud data platform Snowflake has expressed the view that customers should have the freedom to choose their management software. Snowflake, which recently signed a $200 million agreement to use OpenAI's models while also selling a competing product, stated its position.

"OpenAI is an important partner and innovator, but in the enterprise market, we believe no single company should dominate the agent control platform," said Baris Gurtakin, Snowflake's Vice President of AI, in an emailed statement.

As OpenAI attempts to leverage its model expertise to enter the agent management space, traditional giants are mounting a counteroffensive. As reported last week, Judson Althoff, CEO of Microsoft's Commercial Business division, stated in an email to the sales team that OpenAI has not yet established the "platform company" status that Microsoft holds, which allows it to offer a diverse product portfolio.

Althoff also pointed to other advantages held by Microsoft, such as its proprietary infrastructure. Microsoft's competing product, Agent 365, is a direct competitor to OpenAI's Frontier.

The controversy surrounding OpenAI's new product partly stems from its implied positioning: that ChatGPT will become the sole, authoritative console for employees to manage all AI agents. This conflicts with the approach advocated by vendors like Microsoft, Salesforce, and ServiceNow, who argue that customers should manage agents through their existing consoles. It is anticipated that only one winner will emerge from this battle for console dominance.

Nevertheless, existing major players remain confident in their advantages. They already control the core applications where enterprises store and process backend data for HR, IT, and other functions—a foundational element that OpenAI lacks.

"We offer much more than just a console. Users don't care where they view information, but for us, the key is being able to manage the full lifecycle of an agent here: automation, budgeting, development permissions, deployment methods, operational responsibility. These are capabilities that a large language model alone cannot provide," said Dorit Zilbershot, Vice President of AI Product Management at ServiceNow.

Although no one is willing to state it publicly, it is almost certain that sales leaders at other enterprise software firms are expressing similar views internally, pointing to the significant challenges OpenAI and Anthropic will face in persuading enterprises to adopt their agent management software.

**Is Databricks Preparing for a Market Crash?**

Enterprise AI software provider Databricks recently completed a new funding round totaling $70 billion, which included $20 billion in debt financing. Last quarter, the company reached an annualized revenue run rate of $54 billion, representing growth of over 65% year-over-year. The 13-year-old company is now valued at $1.34 trillion.

However, Databricks co-founder and CEO Ali Ghodsi told CNBC earlier this week that recent discussions with some investors have raised concerns that the tech industry might be heading towards a crash similar to the dot-com bubble. He mentioned that some investors are worried about circular financing patterns in the AI market, such as Nvidia investing in cloud provider CoreWeave, which then uses those funds to purchase more Nvidia GPUs to expand its data centers.

Ghodsi also told CNBC that current turbulence in the enterprise software market adds to his concerns. Many investors believe that AI agents will replace numerous functions for which customers currently pay long-term fees, potentially even disrupting entire application systems like CRM.

Although usage of Databricks' products continues to rise as more knowledge workers adopt AI, Ghodsi stated that investors are equally concerned about the potential negative impact on traditional software vendors and database companies.