The future of enterprise software in the age of AI

There is a central tension shaping the future of enterprise software: AI capabilities are advancing faster than many organizations, both vendors and clients, can operationalize them. Put another way, AI is simultaneously becoming extraordinarily capable yet difficult to turn into enterprise value. 

As AI moves from enhancing software to doing the work itself, vendors face fundamental questions: 

• What product architecture best capitalizes on the unique abilities of AI?
• Where do humans fit in the equation?
• Is our business model structurally sound amid these changes?

The challenge is making the right decisions before the outcome is clear. Business leaders need more than a forecast of where AI could go. They need data on their specific markets so they can make better business decisions. They need evidence that promising use cases can progress from pilot to measurable value for their clients. Otherwise, today’s AI investment can become tomorrow’s shelfware.

This is where an AI Market Research Agency becomes strategically valuable: not by predicting the future, but by replacing assumptions about it with evidence and insights.

Enterprise software is dead, long live enterprise software

Helping people perform work has been software’s fundamental value proposition since its creation. Clearly, AI is disrupting and redefining that paradigm, moving from optional to imperative. Complicating matters, turning AI into value across the enterprise may require organizations to redesign not only the software, but also the economics and organizational structures around it. 

The companies most at risk may not be those that fail to adopt AI. They may be those that successfully add AI to a product whose underlying assumptions are becoming obsolete. 

Two companies illustrate the strategic divide. Their experiences also demonstrate why technological capability alone cannot determine the right response. AI market research can reveal what technology forecasts cannot: which products customers will continue to value, how competitors are repositioning, and where sustainable market opportunities are emerging.

Sacrifice the interface, save the platform

AI agents are software systems that can make decisions and take actions with limited human direction. Instead of simply presenting information or recommending a next action, software can analyze data, make decisions, execute tasks, and coordinate activities across multiple systems with progressively less human intervention.

ServiceNow is betting that the underlying value of its platform survives the transition to AI agents—even if the interface customers use to interact with it does not.

Rather than protecting a pre-AI version of its product, ServiceNow is embedding AI throughout its portfolio and positioning its workflow platform as a controlled environment through which both humans and AI agents can act. That distinction matters. If employees increasingly delegate work to AI rather than navigating enterprise applications themselves, the traditional user interface becomes less strategically important. The workflows, permissions, integrations, business rules, and enterprise data underneath it may become more important. 

ServiceNow is also adapting the economics of its business to this change. Traditional software-as-a-service (SaaS) monetization is heavily tied to human users: more employees using an application means more seats and more subscription revenue. But autonomous software complicates that equation. If an AI agent can do work that previously required multiple paid users, or “seats”, value can increase even as the number of human interactions declines. Bain & Company sees a corresponding shift in software economics: “price for outcomes, not log-ons.” ServiceNow’s emerging consumption model reflects that same economic logic. 

This is defensive cannibalization in its most useful form. ServiceNow is not protecting the old economic model at all costs. It is attempting to move the unit of value from the interface to the platform, shifting from the number of people using the software toward the amount of valuable work the platform enables.

From Rattle to Von: The product is the problem

Rattle reached a radically different conclusion. The sales-software company initially tried to maintain its existing SaaS business while developing an AI “superagent.” As AI models improved, CEO Sahil Aggarwal concluded that the legacy product was constraining what the company could build. 

Rattle stopped selling its original product, reduced its workforce, and shifted its forward strategy to Von, an AI-native system designed to perform work across the existing technology stack.

The contrast with ServiceNow is important. ServiceNow is willing to cannibalize aspects of the traditional user experience to preserve the platform underneath it. Rattle concluded that the product itself was no longer the right foundation for an AI-native future. For software executives, the implication is uncomfortable: protecting a legacy product only creates value when the legacy product remains worth protecting.

Rattle sold a better way for people to use software. Von sells an artificial worker that uses the software for them.

Software on the payroll

The same transition reshaping software products will reshape the organizations that use them. McKinsey & Company reports that the share of employees using AI at work rose from 30% in 2023 to 76% in 2025. We’re already seeing work shift from people to software, with routine decisions increasingly delegated to AI agents. 

Deloitte calls this an emerging “silicon-based workforce.” Organizations may soon manage AI agents much as they manage employees: onboarding them, assigning permissions, monitoring performance, and deciding when human judgment must enter the workflow.

That changes more than headcount. Managers may oversee work performed by both people and AI. Technology, operations, and workforce planning are coming together around a new question: not how many people does this process require, but what combination of human and digital labor should perform it?

This changes things for enterprise software vendors. Tomorrow’s applications won’t merely serve the workforce. They will increasingly become part of it.

Beyond the application

AI agents could also change the architecture of enterprise software itself.

Today, employees often connect applications manually, moving information among ERP, customer relationship management (CRM), productivity, analytics, and other systems. Agents can reverse that relationship. Instead of choosing an application, users may increasingly specify an outcome. Agents can determine which systems, data, and tools are required to achieve it.

That doesn’t mean current applications are obsolete. Their data, permissions, business rules, and transaction histories may become more valuable as autonomous systems depend on them. But value could migrate from the interface toward the infrastructure agents need to act. 

McKinsey also notes a change in software economics as AI moves from enabling work to performing and orchestrating it. As human interaction declines, value becomes less tied to the number of users logging in, which increases pressure on traditional seat-based pricing. 

The next AI decision belongs to the market

ServiceNow and Rattle made different bets because there is no universal playbook for AI disruption. The same decisions are now confronting companies across the enterprise technology market. What should you protect? What should you cannibalize? What should you build? And what are your customers already telling you about where your current strategy may be wrong? Technology forecasts alone can’t answer those questions. They require evidence from the market.

How can an AI market research agency support strategic decisions?

An AI technology market research firm tests the assumptions behind high-stakes AI decisions. The objective is not to predict precisely what AI will look like five years from now. It is to make today’s decisions with better evidence about where customers, competitors, and markets are moving.

Our research capabilities can help business leaders answer questions such as:

• Customer research: Which AI capabilities address problems customers value enough to pay for?
• Product and usability research: How do customers use, evaluate, and adopt AI-enabled products and features?
• Industry tracking: How are technologies, competitors, customer expectations, and market conditions changing over time?
• Market opportunity, feasibility, entry, and sizing: Where are viable AI opportunities emerging, how large are they, and which warrant investment?

How can SIS International help enterprise leaders navigate AI disruption?

SIS International’s AI market research services combine market intelligence, customer research, competitive analysis, and strategy research to help executives evaluate emerging opportunities and reduce uncertainty around AI investment.

The future of enterprise software will not be determined by which companies add the most AI. It will be determined by which companies correctly identify where value is moving and move with it.

Endnotes

1. ServiceNow. ServiceNow Moves Beyond the Sidecar AI Era, Giving Customers a Complete AI-Native Experience Across All Products and Packages. ServiceNow Newsroom, April 9, 2026.
2. Bain & Company. Technology Report 2025. Bain & Company, 2025.
3. ServiceNow. ServiceNow — Opens Its Full System of Action to Every AI Agent. ServiceNow Newsroom, May 5, 2026.
4. The Wall Street Journal. Facing AI ‘Apocalypse,’ Once-Hot Software Companies Race to Reinvent Themselves. Clark, Kate, updated August 8, 2026. 
5. McKinsey & Company. How AI is—and isn’t—changing the future of work. April 6, 2026.
6. Deloitte Insights. Tech Trends 2026. Deloitte, 2026.
7. McKinsey & Company. Upgrading Software Business Models to Thrive in the AI Era. September 22, 2025.