Follow ZDNET: Add us as a preferred source on Google.

ZDNET’s key takeaways

• Agentic AI will disrupt enterprise software revenue models.
• The SaaS apocalypse is overrated, but disintermediation is real.
• Providers are honing their core capabilities to stay relevant.

An AI transformation that could disrupt the established hierarchy is playing out behind the scenes of the technology industry.

In this revolution, the enterprise applications that professionals rely on every day could be disintermediated, meaning AI agents would become the primary way businesses access those services, while new relationships with AI firms, their models, and their capabilities redefine how end users interact with IT.

Also: The new enterprise AI expert every company needs – and why

Technology analyst Gartner says agentic AI is set to disrupt enterprise software revenue models, with up to $234 billion in application spending exposed to agentic arbitrage — where agents complete tasks across multiple systems, reducing the need for users to interact with traditional interfaces — between now and 2030.

By the end of the decade, this interaction with AI will account for roughly 20% of enterprise application software-as-a-service (SaaS) spending.

Also: AI agents are your new colleagues – how to get the best results

It’s these kinds of numbers that have led to the rise of the term “SaaS apocalypse,” with some estimates suggesting markets have lost roughly $300 billion in SaaS valuations over the past 18 months due to a correction driven by fears that AI agents will replace traditional tools.

So, is your favorite software provider about to die? And how are vendors responding to the threat of disintermediation?

Prepare for metamorphosis

First, a word of caution — while the impact of AI has already affected the market values of some software firms, it’s important to recognize that disintermediation is a complex beast. Yes, some providers will struggle to retain their dominance in the future, but others will not.

Gartner suggests the shift to agentic AI is less an apocalypse and more a metamorphosis. SaaS will not be destroyed; it will emerge in a different form, with the winners defined by how they respond to the threat of disintermediation.

Also: 40% of enterprises will scrap AI agents – 3 ways to ensure yours don’t fail

A similar sentiment was expressed by Shannon Kalvar, research director at analyst firm IDC, when ZDNET asked about the future of the software industry in the AI era.

“The SaaS apocalypse is overrated, but your use of the word disintermediation is correct,” he said, reflecting on the potential for professionals to use AI-powered solutions that circumvent the traditional enterprise software interface.

“You can say, ‘I need these things,’ and Claude, ChatGPT, Perplexity, or whatever model you choose can go out and write a bunch of code and pull from capabilities,” he said.

“That ephemeral application becomes your new work surface, so you’re disintermediated from the application.”

Also: AI is causing cognitive fatigue. Here’s how to work with more haste and less speed

While this shift poses an existential threat to software providers, Gartner suggests it also creates a substantial revenue opportunity for vendors that are developing services and platforms to support cross-domain workflows.

Kalvar said the AI giants are moving quickly and could develop a much wider stack of services that encroach on incumbents’ territories. As a matter of urgency, providers must consider whether their unique selling points will still be original five years from now.

“Those durable capabilities are what you sell,” he said. “The idea of an application that is just the application’s logic and functions; that’s more problematic. Could you still have a work surface like that? For vendors, the question of ‘What durable capabilities are you providing? What unique thinking, what mathematical capability?’ becomes crucial.”

A question of trust

One executive who’s considering such questions is Clare Hickie, CTO for EMEA at Workday, who told ZDNET that AI brings uncertainty, and business leaders will look for providers that offer a degree of certainty, an area where she believes her firm excels.

“Digital leaders are not always ready to adopt AI because of concerns about trust,” she said. “What’s crucial to recognize is that privacy by design and security frameworks are built into absolutely everything that we do.”

Hickie spoke to ZDNET at Workday’s recent innovation media event at its EMEA HQ in Dublin, where the company’s executives discussed the firm’s product roadmap, including the development and integration of agents through its next-generation service Sana.

Also: The autonomous business is coming. Here’s why that shift is good news for professionals

They suggested the company is building the “front door to work,” where employees will log in to Workday and use agentic services to ask natural-language questions about key issues, such as payroll variations, and receive personalized answers from enterprise data sources.

As the risk of disintermediation rises, Hickie suggested Workday’s durable capability will be its ability to build a trusted platform for cross-business workflows.

“We have the most risk-averse organizations inside our environments, and they look at us because they have to trust what we’re constantly delivering,” she said.

Freshworks CTO Murali Swaminathan echoed similar sentiments at his firm’s recent Refresh 2026 event in New York City, where executives detailed the company’s aim to create an agile, open platform for customer support, with its Freddy AI agentic technology at its core.

Also: Forget productivity: Here are 5 strategic shifts that drive real AI value

Swaminathan told ZDNET that while AI is fundamentally changing all businesses, the data that run organizations’ operations will still be held in a system of record like Freshworks.

He suggested AI giants such as OpenAI and Anthropic are moving fast and are keen to gain market share, but it’s unlikely they’d target service management support.

“These AI specialists don’t want to build everything. They want to be that layer where you engage, but the underlying layer will still be systems like ours,” he said.

“The typical example I give is that if someone’s trying to build an app using Anthropic, they will have to code an app that does workflows, SLAs, business rules, all these things, right? That means they need to know about how the system really works, and that isn’t easy.”

Stick to your own field

The message from tech executives seems to be that while AI models are changing the game, the behind-the-scenes rules governing effective enterprise software, such as governance, security, and data management, mean there’s still room for IT firms with a unique selling point.

Snowflake co-founder and president Benoît Dageville, who spoke with ZDNET at his firm’s recent Summit 2026 event in San Francisco, recognizes that AI brings new specialists to the market, yet believes Snowflake will stay relevant by listening to customers and developing a trusted data platform: “We have the data, and there is no AI without data.”

Also: US workers are the world’s biggest AI skeptics – and it’s not just about job loss

Dageville said anyone trying to understand the risk of disintermediation should look to lessons from the past, including the rise of cloud computing giants.

He said people wondered whether Snowflake’s products would be usurped by cloud giants such as Amazon, especially when Amazon developed its data warehouse service Redshift.

However, today Snowflake has built a strong relationship with Amazon Web Services. Snowflake uses AWS infrastructure, and AWS customers can use Snowflake services.

Dageville said Amazon executives recognized that the priority was becoming the leading player in cloud, both by offering customers choices and by perfecting its services — and he suggested AI giants will draw similar conclusions.

Also: How to beat the AI algorithm and get the job of your dreams

“Yes, potentially Anthropic can be Snowflake, or OpenAI can be a Snowflake competitor. That’s possible, and then they can build a platform like Snowflake. But what is their magic advantage?” he said.

“They don’t fully understand this area — we are specialists in our technology, so I’m still not convinced they can build another Snowflake. And, what’s more, I don’t think they want to.”