AI pricing models shift from software access to work performed

The economics of AI-enabled work

As generative AI moves from pilot projects to production deployments in customer service, a fundamental question is emerging that extends beyond technology capabilities: When AI systems perform work traditionally done by people, who captures the economic value created?

The question matters because AI in customer service is no longer just augmenting human workers—it's increasingly performing complete tasks autonomously. Systems now resolve billing inquiries, process returns, schedule appointments, and update accounts without human intervention. This shift is forcing enterprises, vendors, and investors to reconsider how AI-enabled work should be priced and valued.

From features to outcomes

Historically, enterprise software was sold through licenses, subscriptions, and per-seat pricing. Customers paid for access to technology. AI introduces a different dynamic, according to analysis first reported by No Jitter.

IDC analyst Mickey North Rizza argued in April 2026 that enterprise software competition is shifting "from features to outcomes." Organizations are beginning to evaluate AI investments based on productivity improvements, workflow completion rates, automation outcomes, and measurable business impact rather than feature lists or user counts.

That shift is becoming visible in how investors value AI companies. When AI-native customer service vendor Sierra reached a reported valuation of approximately $15 billion despite being only a few years old, it intensified debate about what investors actually value: better software, labor replacement, or an entirely new economic model.

Victor Manzanera, founder of Scale Edge Group, suggested that investors aren't simply valuing superior AI technology. They're valuing the opportunity to capture labor budgets that historically flowed to service organizations and outsourcing providers. If AI can complete customer service tasks without human intervention, the addressable market extends beyond software budgets into operating budgets.

Infrastructure economics remain uncertain

Industry analyst Josh Bersin has raised concerns that many AI business cases are built on pricing models that may not hold indefinitely. Pointing to enormous investments by cloud providers, chip manufacturers, and AI companies, he notes that investors will eventually expect returns consistent with the capital being deployed. If AI providers must generate sustainable profits, the economics underpinning today's business cases could change significantly.

Yet Sheila McGee-Smith, founder and principal analyst at McGee-Smith Analytics, noted that dramatic cost shifts in customer experience technology aren't new. Telecom costs raced toward zero in the 1990s. Cloud contact center licenses that once commanded over $200 monthly now deliver more functionality at a fraction of that cost. "The dynamics in the cost of AI are far less predictable, causing uncertainty on the part of CX technology buyers," McGee-Smith said. "That is why the ability to solve a specific business problem, with little or no professional services investment, and consumption-based pricing, are increasingly attractive."

Organizational readiness lags technology

While economists and investors debate value capture, enterprise buyers face a more immediate challenge: technology is advancing faster than organizational readiness.

Gartner predicted earlier this year that by 2027, half of companies that reduced customer service headcount because of AI would rehire employees into new roles. The prediction challenged the narrative that AI would simply replace human workers. The underlying message had more to do with operating models than staffing levels—expertise, judgment, governance, exception handling, and oversight remain critical even as AI becomes more capable.

IDC research continues to highlight governance, workforce readiness, data quality, integration, and operational redesign as major barriers to successful AI adoption. Organizations are discovering that adding AI to existing processes rarely delivers transformational results. Meaningful gains often require redesigning workflows, redefining roles, strengthening governance, and improving supporting systems.

Why it matters

The pricing model question matters because it will determine how economic value created by AI is distributed among infrastructure providers, model developers, software vendors, enterprises, and customers. Many organizations are building business cases around today's AI economics while simultaneously reducing investments in human expertise. If model pricing changes or usage expands faster than expected, recreating that expertise may prove more difficult than eliminating it. The long-term economics of AI may depend as much on organizational choices as technology costs.

These details were first reported by No Jitter, with analysis from Rob Hilsen examining how the shift from software access to work performance is reshaping enterprise AI economics.