Update approval-flows.mdx

docs/embeddable-uis/element/approval-flows.mdx

@@ -35,13 +35,15 @@ Permit.io Approval Flows uniquely feature:
 _![Audit Logs Element](/img/elements/approval-flows-full.png)_
 
 ### AI agents
+Today, most application users are humans - but this reality is bound to change drastically in the coming years, as other applications, such as AI and machine learning agents, will suppress human users. 
+These agents tend to perform actions much faster, much more frequently, and on an unprecedented scale. It is also often hard to know exactly which data these agents will require access to. 
+As completely blocking these agents defeats the purpose, we require a way to grant humans control over these agents in an organized way while maintaining the speed and scale they offer. 
 
-The integration of AI agents presents new challenges in managing authorization.
 Traditional models like Mandatory Access Control (MAC), where developers solely assigned permissions, have given way to Discretionary Access Control (DAC).
 This shift is crucial in the era of AI and cloud-native architectures, where users often need to grant access to third-party AI applications
 and define unknown data scopes.
 
-The **Approval Flows Element** addresses these complexities by enabling a dynamic, policy-based approval process.
+The **Approval Flows Element and API** addresses these complexities by enabling a dynamic, policy-based approval process.
 It allows developers to create flexible approval flows where permissions can be dynamically adjusted through policy changes,
 catering to the complex nature of AI agent interactions. It empowers users to control access to their data by AI agents,
 maintaining security and compliance.