How AI Is Changing Enterprise

Levie argues that enterprise AI products create value through the software around a model: business logic, customer data, integrations, and workflows. Businesses buy outcomes such as resolving support requests or processing contracts, so providers should swap in better models as they improve. He expects model differences to matter less for most business uses over time.

Levie expects intelligence costs to approach the underlying cost of computing, making software that applies AI to real business problems more important. He points to opportunities in both industry-specific products and tools that connect AI systems. As tokens become a smaller share of a product's value, the surrounding software can make it more useful and harder to replace.

Levie says major companies are testing AI across many areas, though privacy, compliance, and regulation keep adoption early. He distinguishes context, such as HR and ERP systems that most businesses need, from core capabilities that shape a company's unique value. Businesses may buy common software externally while building proprietary AI for areas such as drug development or personalized services.

Levie compares AI with SaaS, which expanded software markets by making products accessible to many more businesses. He argues AI can add spending by enabling work companies previously left undone, such as translating more advertising or extracting contract details. He estimates software including AI could become five times larger over the next decade, and says companies may reinvest efficiency gains to grow and serve customers.

Levie argues that faster work can lead companies to build better products, grow revenue, and hire in other areas, rather than simply keeping all the savings as profit. He says competition pushes businesses to reinvest efficiency gains, while consumers benefit from better products and lower costs. His closing vision links AI automation to more accessible education and healthcare.