Announcing New AI-Powered Capabilities in Asgardeo

We’re excited to announce major updates to Asgardeo that bring cutting-edge artificial intelligence (AI) capabilities to strengthen access control for AI agents and enhance developer productivity. These new features simplify complex tasks, accelerate development, and improve security posture, making Asgardeo more powerful and easier to use than ever.

These new features are a first step into the overall WSO2 strategy to bring together the AI-native capabilities enterprises need to easily code, integrate, manage and secure intelligent digital products and services.

Managing access control for AI agents

AI agents are rapidly becoming integral to business processes, automating tasks, analyzing data, and even making decisions. As they grow more capable and autonomous, it’s essential to control their access to sensitive systems and data with the same rigor as human users.

Why AI agent access control matters

Unlike human users, AI agents operate 24/7, process vast amounts of data, and often perform actions on behalf of users or systems. Without proper controls, AI agents could introduce significant security and compliance risks:

• Over-permissioning risks: If AI agents are given excessive access, they could expose sensitive data or execute unauthorized actions.
• Lack of accountability: If agents operate under blanket system credentials, it becomes difficult to trace specific actions back to an agent.
• Potential for compromise: If an AI agent is compromised or manipulated, it could be used to bypass security controls or exfiltrate data.

AI agents as first-class entities

Asgardeo treats AI agents as distinct entities with their own identities, separate from human users. This enables fine-grained control over what AI agents can access and how they operate:

Figure 1- identity assignment – Each AI agent is assigned a unique identity that defines its permissions and access scope

Figure 2 - Credential issuance – AI agents can be issued security credentials, including certificates and application keys, to authenticate their actions securely

Figure 3 - Role-based access – AI agents can be assigned roles on behalf of users or systems, ensuring they only have the permissions necessary for their task

Adaptive access management

AI agents' access can be adjusted dynamically based on context, such as time of day, location, or the sensitivity of the data involved.

Auditing and monitoring of AI agent access, actions, & anomalies

Visibility and control are critical when AI agents operate autonomously. Asgardeo enables full auditing of AI agent activity related to IAM, including access and audit logs.

Figure 4 - AI agent auditing and monitoring - Record when and how AI agents access systems and data such as getting tokens, accessing IAM APIs (user APIs), and state changing IAM operations in IAM

Delegation and authorization

AI agents often need to act on behalf of human users or other systems. Asgardeo allows agents to inherit permissions and roles under strict delegation policies:

• Scoped access – AI agents are granted only the minimum permissions needed for a specific task.
• Expiration and rotation – Credentials for AI agents can be time-limited and rotated to reduce the risk of misuse.

Revocation – If an AI agent is compromised or no longer needed, its credentials and access rights can be revoked instantly.

Boosting developer productivity with AI

We’ve integrated AI into key parts of the identity and access management (IAM) process to make life easier for developers:

AI-generated registration and login flows

Forget manual flow building—now you can use natural language to describe the ideal registration or login experience. Asgardeo’s AI will automatically generate a fully functional flow based on your input, reducing development time and minimizing errors.

Figure 5 - AI-Powered registration flow builder

Figure 6 - AI-powered login flow builder

AI-powered branding

Consistent branding across platforms has never been easier. Just provide your website’s URL, and Asgardeo will automatically apply your brand’s fonts and colors to all your UX interfaces such as:

• Web and mobile login pages
• TOTP (Time-based One-Time Password) pages
• Email templates
• And more

This ensures a polished, on-brand user experience without manual effort.

Figure 7 - AI-powered branding

AI-powered adaptive authentication

We are in the R&D research phase of making adaptive authentication smarter. Asgardeo will leverage AI to determine the appropriate authentication flow based on pre-defined security policies. This will mean more secure, context-aware authentication—without adding friction for legitimate users.

Availability

Most of these capabilities are available immediately. AI-generated registration is available in Asgardeo in preview mode. AI agent access control will be available in Asgardeo by June, 2025. AI-powered adaptive authentication is currently in research and we will announce availability in preview mode when complete.

As AI agents continue to play a larger role in business and security ecosystems, treating them as secure, identifiable entities is no longer optional—it’s essential. With WSO2’s new AI-powered access control and identity management capabilities, you can secure AI agents as effectively as human users, ensuring that your systems remain secure and compliant in an increasingly AI-driven world.

Try Asgardeo for free now, and don’t miss out on Geethika Cooray's (Vice President/General Manager - IAM at WSO2) opening keynote from WSO2Con 2025 for exclusive insights into the vision behind these major updates.