ZDNET's key takeaways

Agentic AI will reshape how enterprise software generates revenue.
The idea of a total SaaS collapse is overstated, though disintermediation poses real challenges.
Vendors are sharpening distinctive strengths to remain essential.

A quiet transformation is underway across the technology sector, one that could reorder how established enterprise tools fit into daily operations.

Under this change, familiar business applications risk being pushed aside as AI agents step in as the main gateway to services. Fresh ties to AI developers, their models, and their features would then set the tone for how teams engage with technology.

Analyst firm Gartner projects that agentic AI could affect up to $234 billion in application spending by 2030 through a process of agentic arbitrage. In this model, agents handle tasks across various platforms, cutting down direct user interaction with conventional interfaces.

By the close of the decade, such AI-driven interactions are expected to represent about 20 percent of enterprise SaaS outlays.

Figures like these have fueled talk of a SaaS apocalypse. Market observers note that concerns over AI agents displacing established tools have already trimmed roughly $300 billion from SaaS valuations in the past year and a half.

This raises questions about which software providers will endure and how they are adjusting to the risk of being sidelined.

Prepare for transformation

Caution is warranted. Although AI has already pressured valuations at some software companies, disintermediation remains a nuanced issue. Certain providers may lose ground, while others are positioned to hold steady.

Gartner frames the move toward agentic AI as a transformation rather than an end. SaaS will evolve rather than vanish, and success will hinge on how each vendor meets the challenge of disintermediation.

Shannon Kalvar, research director at IDC, offered a comparable view. When asked about the software sector's outlook in the AI age, he noted that the SaaS apocalypse narrative is overstated, yet the concept of disintermediation holds weight.

Professionals may increasingly turn to AI solutions that bypass traditional enterprise interfaces, he explained. Users could state their needs and let models such as Claude or ChatGPT generate code and draw on various capabilities. The resulting temporary application then becomes the primary workspace, creating distance from the original software.

Although this trend threatens some providers, Gartner sees a sizable opening for those building tools that support workflows spanning multiple domains.

Kalvar added that major AI players are advancing rapidly and may expand their offerings into areas once dominated by established vendors. Companies must therefore assess whether their core strengths will remain distinctive in five years.

Those lasting capabilities form the real offering, he said. Pure application logic alone may prove vulnerable. Vendors should focus on what unique expertise or computational methods they deliver.

Building trust

Clare Hickie, CTO for EMEA at Workday, is among the executives weighing these issues. She told ZDNET that AI introduces uncertainty, and leaders will favor providers that deliver reliability, an area where her company aims to lead.

Digital leaders often hesitate on AI adoption due to trust concerns, she noted. Privacy by design and robust security measures are embedded throughout Workday's approach.

Hickie spoke at the company's recent innovation event in Dublin, where leaders outlined plans for agent development through the Sana platform. The goal is to create a central entry point for work, allowing employees to pose natural-language questions on topics such as payroll and receive tailored responses drawn from company data.

As disintermediation risks grow, Hickie emphasized that Workday's strength lies in offering a trusted environment for cross-business processes. Its most cautious clients rely on this consistent delivery of security.

Freshworks CTO Murali Swaminathan voiced similar thoughts at his firm's recent event in New York. He described efforts to build an adaptable platform for customer support centered on the Freddy AI agent technology.

Swaminathan explained that while AI is reshaping operations, the core data driving business processes will continue to reside in established systems of record. AI specialists are unlikely to focus on service management, he added, because they prefer to serve as the engagement layer while underlying systems handle the detailed rules and workflows.

Focus on specialization

Executives across the sector stress that AI models are altering interfaces, yet the underlying requirements of governance, security, and data handling leave room for providers with clear expertise.

Snowflake co-founder and president Benoît Dageville, speaking at the company's recent Summit event, acknowledged the arrival of new AI specialists. He argued that Snowflake will remain relevant by maintaining a trusted data platform, noting that AI cannot function without data.

Dageville pointed to past shifts, such as the rise of cloud computing, as instructive. Early doubts about whether cloud giants would overtake specialized tools proved unfounded, he said, because each party recognized its strengths and formed partnerships.

AI leaders may similarly conclude that specialization matters, he suggested. While competitors could attempt to replicate Snowflake's capabilities, their advantage in other domains does not automatically translate to deep expertise in data platforms. He expressed doubt that they would pursue such a path.