In the AI era, enterprises reject the fragmented, best-of-breed SaaS model. They prefer a single AI platform that handles entire workflows across departments. This avoids data silos and streamlines compliance, making end-to-end automation the key value proposition.

Many companies initially build their own AI gateway, viewing it as a simple, thin proxy layer. However, upon moving agents to production, they quickly discover that real-world complexity around governance, observability, and security requires a far more robust, specialized control plane platform.

Enterprise software companies report huge AI revenue growth, but this is often a sales tactic. Systems like Workday's 'flex credits' are packaging innovations designed to capture AI budget from CIOs, not fundamentally new, agentic experiences that transform how work gets done.

AI companies are selling large, seat-based contracts based on hype and experimental budgets, inflating current ARR. Investors are skeptical because, like early SaaS, customers will eventually demand usage-based or outcome-based pricing, challenging the long-term revenue stability of these startups.

Startups face a "governance tax" where monitoring platforms charge significantly more than the underlying AI API usage. An example cited is a team nearly paying $36,000 annually for a tool to manage a $14,400 AI spend, representing a 250% markup just for monitoring.

Beyond upfront pricing, sophisticated enterprise customers now demand cost certainty for consumption-based AI. They require vendors to provide transparent cost structures and protections for when usage inevitably scales, asking, 'What does the world look like when the flywheel actually spins?'

Contrary to traditional software evaluation, Andreessen Horowitz now questions AI companies that present high, SaaS-like gross margins. This often indicates a critical flaw: customers are not engaging with the costly, core AI features. Low margins, in this context, can be a positive signal of genuine product usage and value delivery.

In enterprise AI, competitive advantage comes less from the underlying model and more from the surrounding software. Features like versioning, analytics, integrations, and orchestration systems are critical for enterprise adoption and create stickiness that models alone cannot.

AI startups often use traditional per-seat pricing to simplify purchasing for enterprise buyers. The CEO of Legora admits this is suboptimal for the vendor, as high LLM costs from power users can destroy margins. The shift to a more logical consumption-based model is currently blocked by the buyer's operational readiness, not the vendor's preference.