Hyperflow FAQ's

Answers to commonly asked questions about Hyperflows

General

1. What exactly is a Hyperflow in the Aisera ecosystem?

Hyperflow is an intelligent, autonomous workflow agent that leverages Aisera's domain-specific fine-tuned Large Action Models to understand intentless user inquiries and sequence appropriate function calls in real-time. It can execute actions, generate analytics, and provide personalized responses without requiring traditional workflow configuration.

1. How does Hyperflow differ from traditional workflow automation?

Unlike traditional workflows that require manual configuration and predefined paths, Hyperflow eliminates the need for a traditional workflow builder. It uses Aisera's Agentic AI system to dynamically understand user inquiries and execute appropriate actions in real-time.

1. What are the main benefits of using Hyperflows?

Hyperflows provide several key benefits:

• Autonomous execution of complex workflows without manual configuration

• Continuous improvement through feedback learning mechanisms

• Ability to handle even complex, intentless user inquiries

• Personalized responses based on user context and needs

1. How do we trigger a Hyperflow?

Hyperflows are triggered in the same way as Workflows - you assign a Hyperflow as a Fulfillment to an Intent.

1. Can Hyperflows execute only a single function at a time?

• Hyperflows can plan and execute multiple functions sequentially to fulfill a user's request. It can also pass outputs from one function to the next.

• For example, for a Hyperflow with functions of “check_order_by_id”, and “check_supplier_info”, when a user asks, “What is the address of the supplier of order 12345?”, the Hyperflow will determine to call the first function “check_order_by_id”, the execution result will then contain the supplier name, therefore, the Hyperflow will determine the next step to call is “check_supplier_info”. Then the “check_supplier_info” returns the address of the supplier, and a final answer is provided to the requester.

Implementation

1. What prerequisites are needed before implementing Hyperflows?

• Integrations and data sources required for use as functions should be built, tested, and deployed

• Workflows required for use as functions should be built, tested, and deployed

1. What kind of permissions or access controls are available for Hyperflows?

• Hyperflows can only be used by authenticated and authorized users of the Aisera application (depends on customer's setup)

• When reading/writing data or executing actions in external systems, the Hyperflows will act on the user's behalf. This requires users to explicitly authorize the Hyperflows to act on their behalf (eg. OAuth)

Troubleshooting

1. Why isn't my Hyperflow executing as expected?

There could be a few different reasons that can be diagnosed from the debug information available in AI Lens (offline mode) or from the debug information in the Requests details page (online mode). A few common issues are:

• Ambiguous requests that prevent Aisera applications from triggering the correct flow

  
• Function execution failure due to incorrect invocation (eg. incorrect parameters) or issues with the underlying function

  

  

  
• Aisera or third-party service outage or maintenance

  

1. How can I monitor the performance of my Hyperflows?

Hyperflows execution analytics are available on the Requests page - total, resolved requests, unresolved requests, etc. Additionally, you can drill into the debug information of flows that failed.