So you have defined a bunch of intents for your agent. And you type in a query which you think almost certainly should match one of the intents, but for some reason it doesn't.
Tune the ML threshold
In your agent's settings, you can see a tab called ML settings. Under ML settings, you will see a ML threshold value. Reduce it to a smaller number.
As you reduce the threshold, API.AI becomes less strict in how it interprets matches between user's message and your intents' userSays text. Keep incrementally reducing the threshold, say, by steps of 0.05 until you get the match you expect.
Keep the ML threshold as high as possible but make sure it also satisfies your intent definitions. At the end of the day, you don't want to have the converse problem either!
Want to know how the converse problem manifests itself? Check out my eBook "10 features all API.AI developers MUST know".