Agentic AI Disrupts SaaS With Consumption-Based Models

The AI Disruption: Beyond the Hype, Towards Strategic Advantage

This conversation reveals a critical, often overlooked, consequence of the current AI gold rush: the potential for significant disruption not just to technological landscapes, but to the very business models that have underpinned software and enterprise operations for years. The non-obvious implication is that the rapid advancement of AI, particularly agentic AI, is fundamentally challenging the recurring revenue models of Software-as-a-Service (SaaS) by shifting towards a consumption-based paradigm. This forces a difficult but necessary re-evaluation of how companies are valued and how competitive advantages are built. Investors, tech leaders, and strategists who can grasp this fundamental shift and adapt their models will gain a significant edge in navigating the evolving market. This analysis is crucial for anyone seeking to understand the deeper currents shaping the future of technology and business strategy, moving beyond the immediate excitement to the long-term structural changes.

The Agentic AI Shift: Undermining the SaaS Model

The current fervor around Artificial Intelligence, particularly the emergence of "agentic" AI, is not merely an incremental technological upgrade; it represents a potential paradigm shift that could fundamentally alter established business models, especially within the software sector. Ted Mortonson, Managing Director at Baird, highlights this seismic change, noting the exponential growth of agents and their impact on traditional software. This isn't just about coding assistance; it's about AI agents that can perform complex tasks, often initiated through simple voice or text commands, thereby bypassing the need for extensive human development or traditional software interfaces.

The immediate consequence for the software industry is a direct challenge to the Software-as-a-Service (SaaS) model, which relies on predictable, recurring revenue. Jensen Huang, CEO of Nvidia, is quoted as distinguishing between SaaS as "pre-recorded software" and agentic AI, powered by CUDA libraries, as "real-time." This distinction is critical. Traditional SaaS is built on subscriptions for access to software features. Agentic AI, however, operates on a consumption basis. As Mortonson explains, "Right now you introduce an agent which is consumption based and nobody can get to free cash flow. So how do you put a multiple on free cash flow? You can't. You can't model." This creates a significant modeling challenge for investors and companies alike, as the established metrics for valuing SaaS businesses become obsolete. The downstream effect is a potential devaluation of companies that cannot adapt their revenue models, even if their core technology remains robust.

"The large cap companies most likely will make it over the chasm on transitioning from a SaaS model. And what Jensen was pretty funny last night, he basically said software is changing and you know, Nvidia has a huge software component that nobody really talks about. And it's these CUDA libraries are very disruptive on traditional SaaS. So he called software as a service a prerecorded software where agents or agentic or what he's working on and CUDA libraries is real time. Wow, that's a big, big statement."

-- Ted Mortonson

This shift also has profound implications for the competitive landscape. While large companies like Salesforce might eventually adapt, the speed of this transition is a significant factor. The "Nvidia tax," as Mortonson terms it, refers not only to the direct cost of their hardware but also the lock-in effect of their integrated systems. Companies like Google and Amazon are actively developing their own silicon to avoid this dependency, signaling a broader industry push for alternatives. This creates an environment where companies that can offer more flexible, consumption-based AI solutions, or those that can successfully integrate agentic capabilities into their existing platforms without alienating their customer base, will likely emerge stronger. The conventional wisdom of building sticky SaaS products is being challenged by the inherent flexibility and power of AI agents, forcing a re-evaluation of what "stickiness" truly means in this new era.

The Productivity Paradox: AI as an Informant, Not an Educator

The conversation around AI's impact on productivity, particularly as discussed by John Bilton, Head of Global Multi-Asset Strategy at JPMorgan Asset Management, offers a nuanced perspective that moves beyond simplistic notions of job displacement. Bilton frames AI not as a magical solution that instantly makes us smarter, but as a powerful tool for parsing data and informing decisions. This distinction is crucial for understanding the long-term productivity gains.

The analogy of moving from "batch chemistry" to "flow chemistry" illustrates this point effectively. Just as flow chemistry enabled continuous, data-driven optimization of chemical processes, AI can provide real-time insights into complex systems, allowing for more efficient operations. Bilton states, "What it doesn't do is make the decisions around how to influence how that flow is going through. And that's what a human being does. But where AI, I think, has something which we can really lean into is that ability to help us quickly parse through data and understand the problem set. It doesn't mean that suddenly it educates us, but it informs us. And I think that's an important subtlety." This means AI enhances human capabilities by providing better information, faster, allowing individuals to make more informed decisions.

"What it doesn't do is make the decisions around how to influence how that flow is going through. And that's what a human being does. But where AI, I think, has something which we can really lean into is that ability to help us quickly parse through data and understand the problem set. It doesn't mean that suddenly it educates us, but it informs us. And I think that's an important subtlety."

-- John Bilton

The implication for businesses is that AI's true value lies in augmenting human expertise, not replacing it entirely. This requires a strategic approach to AI integration, focusing on how it can improve decision-making processes and unlock new avenues of inquiry. Drew Matus, Chief Market Strategist at MetLife, echoes this sentiment, arguing that AI will not lead to an "AI jobs apocalypse." Instead, he suggests that AI will help individuals complete more tasks, but the constraint will remain human capacity for engagement and the generation of new questions stemming from the new knowledge. This creates a feedback loop: AI provides information, humans use that information to generate new questions and opportunities, which in turn drives further demand for AI-driven insights. Companies that foster this symbiotic relationship between AI and human intelligence will likely see the most significant productivity gains and competitive advantages. The conventional wisdom might focus on AI replacing jobs, but the deeper consequence is AI creating new, more complex human roles and demanding higher levels of critical thinking and strategic decision-making.

The Human Element: Navigating AI's Organizational Challenges

While the technological capabilities of AI are advancing at a breakneck pace, Drew Matus of MetLife points to a critical, often underestimated, bottleneck: organizational structure. The true challenge in leveraging AI, he argues, is not finding tasks for AI to perform, but adapting the internal frameworks of companies to effectively integrate and benefit from these new capabilities. This is where immediate discomfort can lead to long-term advantage.

Matus highlights that AI's potential to expand a company's knowledge base is immense, leading to increased competitiveness and economic growth. However, this expansion of knowledge inevitably generates more questions and requires more complex problem-solving. The constraint, therefore, becomes the human element -- the people within the organization who need to engage with, interpret, and act upon the AI-generated insights. He notes, "I think in terms of using AI, you know, one of the main constraints is going to be the organizational structures of companies rather than finding things for AI to do." This suggests that companies that proactively address their organizational inertia, fostering agility and a culture of continuous learning, will be best positioned to capitalize on AI.

"I think in terms of using AI, you know, one of the main constraints is going to be the organizational structures of companies rather than finding things for AI to do."

-- Drew Matus

The conventional approach might be to invest heavily in AI technology, assuming that the technology itself will drive transformation. However, Matus's perspective suggests that a more critical investment lies in reimagining workflows, training, and leadership to support a more dynamic, AI-augmented workforce. This requires a willingness to confront internal inefficiencies and resistance to change -- a process that can be uncomfortable in the short term. Companies that are willing to undertake this internal re-engineering, even if it means short-term disruption, will build a more robust and adaptable organization, creating a durable competitive moat. The "panic motif" he describes, where companies rush to adopt AI without strategic planning, is precisely what leads to wasted investment and missed opportunities. The advantage lies with those who patiently, but deliberately, align their organizational structures with the transformative potential of AI.

Key Action Items

• Re-evaluate SaaS Business Models: For software companies, begin modeling the shift from recurring revenue to consumption-based pricing for AI-driven services. This requires developing new metrics beyond traditional ARR and LTV. (Immediate to 6 months)
• Invest in AI Literacy and Training: Implement programs to educate employees across all departments on AI capabilities and how to effectively leverage AI tools for data analysis and decision support. Focus on developing critical thinking skills to interpret AI outputs. (Ongoing, with initial rollout over the next quarter)
• Develop Internal AI Integration Roadmaps: Identify key organizational bottlenecks and redesign workflows to accommodate AI augmentation. This includes rethinking team structures and decision-making processes. (Begin planning this quarter, with implementation over the next 12-18 months)
• Explore Alternative Silicon and Infrastructure: For companies heavily reliant on specific AI hardware vendors, investigate and pilot alternative silicon solutions to mitigate vendor lock-in and potential cost increases. (Begin R&D and pilot programs within 6 months)
• Foster a Culture of Experimentation: Encourage teams to experiment with agentic AI tools, understanding that initial use cases may be exploratory and not immediately profitable, but crucial for long-term learning. (Ongoing)
• Focus on Augmentation, Not Just Automation: Shift the narrative and strategy from purely automating tasks to augmenting human capabilities with AI, thereby increasing overall productivity and problem-solving capacity. (Immediate strategic reorientation)
• Build Flexible Financial Models: Develop financial models that can accommodate the volatility and unpredictability of consumption-based AI revenue streams, moving away from rigid annual forecasts. (Develop new modeling frameworks over the next 6-12 months)