Earn production confidence: why passing tests don’t mean production-ready APIs

Most API failures don’t look like failures. The service responds. The status code is green. But somewhere between the last time an API was tested and the moment it’s called in production, the contract quietly changed, and nobody was watching. New data shows that 63% of APIs drift from their tested behavior within 90 days¹, and when things do break, most of the failures trace back to authentication and schema mismatches, not server crashes.  

Meanwhile, API estates keep growing faster than most teams’ ability to govern or verify them, and that gap is now showing up on the balance sheet: the average API-related incident costs $700,000,² and leadership consistently overestimates how solid their testing actually is compared to the engineers doing the work. This session connects those threads: why growth outpaces testing by default, what that gap looks like in practice, and what it takes to close it before it becomes an incident, a compliance finding, or a customer-facing outage. 

What we’ll talk about 

1. Why the gap opens in the first place. APIs multiply faster than ownership, governance, or testing discipline can keep up, and testing coverage usually ends up treated as a minor metric instead of the one that determines whether everything else is trustworthy. 
2. What the gap costs. The financial and reputational toll of untested APIs, and the disconnect between how confident leadership feels and how confident the engineers running tests are. 
3. The mechanism behind the gap. Schema drift, contract mismatches, and the specific, well-documented failure patterns (auth/authz, schema/validation) that account for most real-world API failures, and why most teams only catch them after something breaks. 
4. The shift already underway. How testing itself is changing, from a one-time pre-release task to a continuous, AI-assisted layer that catches drift as it happens instead of after the fact. 
5. Closing the gap in practice. What it takes for a team to move from testing APIs occasionally to testing them continuously, without slowing releases down. 
6. The takeaway. A single reframe we want people to leave with: a step-by-step guide to API confidence.