"It's a very disruptive force and will continue to be for a while," observed Gil Luria, Head of Technology Research at D.A. Davidson, during a recent discussion on CNBC's 'Closing Bell Overtime'. However, Luria quickly tempered this widely accepted notion, asserting that despite the pervasive narrative, artificial intelligence has not yet profoundly reshaped the application software market. This nuanced perspective offers critical insights for founders, venture capitalists, and AI professionals grappling with the true impact and timeline of AI integration across enterprise technology.
Gil Luria, speaking with CNBC’s Jon Fortt and Sarah Eisen, delved into the complex interplay between nascent AI capabilities and established business software ecosystems. His central argument posits that while AI is undeniably transformative, its immediate disruptive power is concentrated in specific layers of the tech stack, rather than broadly upending the precise and integrated world of application software. The prevailing market excitement often overstates AI's current readiness for direct, wide-scale application within core business processes.
One core insight from Luria is that business software demands an exacting level of accuracy and integration that current AI models are not yet equipped to consistently deliver. "Business software is very precise, it's very integrated, there's workflows, permissions, connections," he explained. Unlike consumer-facing AI, where a degree of imprecision might be tolerated, enterprise applications require results that are "exactly right." This fundamental requirement creates a significant barrier to entry for AI-native startups hoping to immediately displace incumbent application providers. The intricate web of existing systems, regulatory compliance, and deeply embedded workflows means that ripping and replacing core software with unproven AI solutions carries immense risk and cost for large organizations.
Consequently, the most tangible benefits of AI are currently accruing to companies operating lower in the technology stack, particularly those focused on data infrastructure and observability. Luria highlighted Snowflake and Datadog as prime examples. Snowflake, a cloud data warehousing company, thrives because AI applications necessitate consolidated and well-structured data, which it provides. Similarly, Datadog benefits from the need to "observe all the applications" as AI drives increased code development and operational complexity. These companies provide the essential plumbing and monitoring that facilitate AI adoption, rather than directly offering AI-powered business applications themselves.
Another critical insight emerged from Luria's comparison of Salesforce and ServiceNow. He argued that the current differentiation in the application software space lies not in who is embracing AI, but in the underlying health and strategic focus of a company's core business. Salesforce, according to Luria, "bet the farm on AgentForce before it was ready and they neglected the core business." This premature and perhaps overzealous pivot towards AI, before the technology or the market was fully mature, led to underperformance. In contrast, ServiceNow, which continued to "execute very, very well in its core business," has seen its stock perform significantly better. This underscores that a strong foundational business, capable of steady execution and measured innovation, is paramount.
Related Reading
- Agentic AI Will Transform Enterprise Software, Eating Services
- Databricks' Billion-Dollar Journey: Lessons in Leadership and Growth
- Agentic Commerce: The Unstoppable Force Reshaping Tech
The reality, Luria suggests, is that the widespread disruption of application software by AI will be a "slow roll." Companies are not yet ready for a wholesale shift to AI-driven business software because the technology itself is still evolving, and the integration challenges are formidable. He noted, "Companies aren't ready because it's really complicated to change your business software." The inherent inertia and complexity of enterprise IT environments mean that even when AI solutions mature, adoption will be gradual. Furthermore, the high expectations for flawless operation in business-critical systems mean that AI must achieve a level of reliability and accuracy far beyond what is currently acceptable in many consumer applications.
This implies a longer runway for established software providers to adapt and integrate AI into their existing offerings. While AI startups will eventually emerge to challenge incumbents, the current landscape favors those with robust core businesses, deep customer relationships, and the infrastructure to responsibly experiment with and deploy AI. The lesson is clear: strategic patience and a focus on core value delivery, rather than chasing every AI trend, will likely define success in the evolving software market.
