diff --git a/README.md b/README.md
index cc2a58b2..4b6c479a 100644
--- a/README.md
+++ b/README.md
@@ -71,15 +71,25 @@ To avoid waiting on backend calls when the Application starts, the suggested app
 `confidence.activate()` and then trigger a background refresh via `confidence.asyncFetch()` for future sessions.
 
 ### Setting the context
-The context is a key-value map used for sampling and for targeting, when flag are evaluated by the Confidence backend.
+The context is a key-value map used for sampling and for targeting, when flags are evaluated by the Confidence backend.
 It is also appended to the tracked events, making it a great way to create dimensions for metrics in Confidence.
 
 ```swift
 confidence.putContext(context: ["key": ConfidenceValue(string: "value")])
 ```
 
-Note that a `ConfidenceValue` is accepted a map values, which has a constructor for all the value types
-supported by Confidence.
+Another way to configure the context involves using the `track` API:
+```swift
+confidence.track(producer: contextProducerImplementation)
+```
+
+The "producer" conforms to `ConfidenceContextProducer`, which allows to dynamically push context changes
+to the Confidence object.
+
+In both cases above, any context change triggers a new asynchronous `fetchAndActivate` and the flags in the
+local cache are re-evaluated remotely according to the new context: until this background operation is complete,
+flag values are returned to the application according to the old context's evaluation, and with `resolveReason = .stale`.
+
 
 ### Resolving feature flags
 Once the Confidence instance is **activated**, you can access the flag values using the