"AI-first" has become a common label in company positioning and investor materials, but it covers two genuinely different situations that carry very different risk profiles, and the label alone doesn't tell you which one you're looking at.
Built AI-first versus rebranded AI-first
Some companies were architected from the beginning around an AI capability that is genuinely core to what the product does - remove the AI and there's no product left. Others had an existing, functioning business and have layered AI features onto it, then adopted "AI-first" language in marketing and investor communications without the underlying product architecture changing as fundamentally as the label implies.
Why this distinction matters for evaluating a business
A genuinely AI-native company's fortunes are tightly coupled to AI capability and cost trends in ways that are core to its investment thesis. A company that has rebranded around AI without a comparably deep architectural shift is making a narrower, lower-risk bet - enhancing an existing business rather than staking its existence on one. Treating both as equally exposed to AI-specific risk and opportunity misprices both.
How to tell them apart
A useful diligence question is what the company's core product or revenue would look like with the AI component removed entirely. If the answer is "basically the same business, minus one feature," the AI-first label is more marketing than architecture. If the answer is "there's no business left," the label is accurate and the associated risk profile follows.
Why the label persists despite the ambiguity
"AI-first" carries genuine appeal to investors and customers in the current environment, which gives companies in the second category real incentive to adopt language associated with the first. That incentive isn't going away, which means the burden is on evaluators to look past the label rather than expecting companies to self-classify accurately.