Acceptance Criteria
Specific, testable conditions that a user story must satisfy to be considered complete. Acceptance criteria define the boundaries of a feature by specifying expected behavior, edge cases, and quality requirements in unambiguous terms.
Well-written acceptance criteria eliminate ambiguity about what done means for a given user story. They are typically written in Given-When-Then format: Given a specific context, When the user takes an action, Then the system produces a defined result. Each criterion should be independently testable and verifiable by anyone on the team.
For AI features, acceptance criteria require careful thought because AI behavior is probabilistic rather than deterministic. Instead of specifying exact outputs, criteria might define acceptable accuracy ranges, maximum response times, and graceful degradation behavior. For example: Given a user query in English, When the AI generates a response, Then the response is factually accurate at least 95% of the time as measured by human evaluation. Growth teams should ensure acceptance criteria include analytics instrumentation requirements, so every shipped feature produces the data needed to measure its impact on growth metrics. Missing instrumentation is one of the most common and costly oversights in AI feature development.
Related Terms
Product-Market Fit
The degree to which a product satisfies strong market demand. Achieving product-market fit means customers are actively seeking, using, and recommending your product because it solves a real and pressing problem for them.
Jobs to Be Done
A framework that defines customer needs as functional, emotional, and social jobs people hire products to accomplish. It shifts focus from demographic segments to the underlying progress customers are trying to make in specific circumstances.
Minimum Viable Product
The simplest version of a product that can be released to test a core hypothesis with real users. An MVP delivers just enough functionality to gather validated learning while minimizing development time and cost.
Minimum Lovable Product
An evolution of the MVP concept that emphasizes delivering enough quality and delight that early users genuinely love the product. It balances speed-to-market with the emotional engagement needed to drive organic word-of-mouth growth.
Design Sprint
A five-day structured process for rapidly prototyping and testing ideas with real users. Developed at Google Ventures, it compresses months of debate into a focused week of mapping, sketching, deciding, prototyping, and testing.
Lean Startup
A methodology for developing businesses and products through validated learning, rapid experimentation, and iterative releases. It emphasizes reducing waste by testing assumptions before building fully-featured solutions.