Every enterprise AI product team faces the same apparent choice. You can build narrow and automation-focused, price on tasks replaced, and run a cost-savings narrative through procurement. Or you can build a platform, price on usage and outcomes, and pitch revenue growth to the CEO. The conventional wisdom treats this as a positioning decision: cost-cutters want point solutions, growth-chasers want platforms, so pick your archetype and design backward.

But the framing is wrong. Buyers approve AI purchases on cost grounds and expand them on growth grounds, and these are two different decisions, made by different people, at different points in the relationship. Building for only one produces a product that either closes deals but churns, or builds retention but can't get through procurement.

What buyers say they want

McKinsey's 2025 State of AI survey found that 80% of enterprise respondents set efficiency as an objective of their AI initiatives. Budget justification in large organizations requires a countable return, and headcount or FTE equivalents are the easiest number to put in a business case. CFOs and procurement committees speak that language. The business case for AI, written for finance approval, almost always leads with labor hours saved, headcount avoided, or process steps eliminated. The buying side, taken at face value, looks like a cost story: fewer people needed to do the same work.

What actually drives value

The problem is that realizing the value from cost-focused AI doesn't require good AI products. It requires good process discipline. Gartner found no correlation between AI-driven workforce reductions and AI ROI. Companies that cut headcount from AI implementations reported similar ROI to companies that didn't.

Only 17% of enterprises use productivity gains for headcount reduction; the majority reinvest in capabilities, R&D, and upskilling. The firms seeing the most EBIT impact from AI, McKinsey found, set growth and innovation as objectives alongside cost, not instead of it. Some industry analyses put the ROI from AI applied to customer experience and revenue generation at three to four times the return from pure task automation.

Two decisions, two actors

The approval decision and the expansion decision are not made by the same person. A CFO approves a tool that reduces invoice processing time. A VP of Sales renews and expands a tool that has become part of how the team generates pipeline. These two decisions have different logic. Approval requires a defensible ROI number, preferably tied to something the finance team already measures. Expansion requires that someone inside the organization cannot imagine doing their job without the product.

The first decision looks backward at savings to be captured. The second looks forward at what the product makes possible. Products built purely for the approval decision often survive procurement and then churn when the renewal lands on someone who cares about what the tool actually does, not what the business case said.

What it means for product

The product implication is that entry point and expansion surface have different requirements. The entry point should be specific enough to produce a clean ROI number for the approval decision: one workflow, one measurable outcome, one defensible claim. The expansion surface, meaning what you build toward and what your usage-based pricing actually prices, should be oriented around what makes users more capable over time, not just more efficient.

Messaging can do both jobs if it's structured around the buying sequence rather than a single archetype: cost language in procurement materials, growth language in product marketing and user onboarding.

The pricing model that fits this structure is usage-based, not seat-based or task-based, because usage-based pricing grows with the growth use case and remains defensible on cost grounds during approval. Within that, hourly consumption pricing anchors to labor cost -- the approval metric. As salaries rise while token prices fall, model efficiency gains become margin rather than pass-throughs.

Sources

McKinsey & Company. "The State of AI: How Organizations Are Rewiring to Capture Value." March 2025. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai

Andreessen Horowitz. "How 100 Enterprise CIOs Are Building and Buying Gen AI in 2025." 2025. https://a16z.com/ai-enterprise-2025/

Confessore, N. "AI-driven layoffs aren't making business sense." CIO / Gartner research. 2025. https://www.cio.com/article/4171054/ai-driven-layoffs-arent-making-business-sense.html

Menlo Ventures. "2025: The State of Generative AI in the Enterprise." 2025. https://menlovc.com/perspective/2025-the-state-of-generative-ai-in-the-enterprise/