AI Automation - Optimize Flow

You can add AI automation to your applications/bots with Workflows (pre-written automations) and Hyperflows (dynamic task planning and execution using Agentic - domain specific - workflows).

These flows can then be triggered conversationally based on bot user Requests, by Events published using the Aisera APIs (see Event Studio), or on a schedule (see Event Studio) within a flow. You can also create Campaigns to provide specific collateral, such as Email Templates, to use within your flow.

These flows can be triggered conversationally, based on user Requests, by Events published using the Aisera APIs (see

Types of Automation

  • Workflows: Provide integration to third-party systems or applications to execute a business logic. Workflows collect information from or interact with external systems using Aisera’s pre-built integration capabilities or direct API calls. These workflow interaction with external systems can alter the state of external systems, exhibiting transactional behaviors.

    • Knowledge Workflows: In addition to RAG, pre-defined knowledge fulfillments can be designed using the Workflows Studio.

    • Service Catalog Workflows: Similar to the workflows for knowledge, pre-defined Service Catalog fulfillments can be designed using the Workflows Studio. For a number of use cases in regulated industries, customers prefer to this option instead of Aisera's RAG to serve list of catalog items.

    • Custom Messages: Customers use these simple workflows to serve pre-written messages to a user's request. Presentation and styling of these messages are completely customizable.

  • Hyperflows: Provide automation using agentic workflows that dynamically plan and execute actions. The Actions include API calls to external systems, and Aisera Workflow Actions. Hyperflows can be generated in minutes using natural-language prompts and API collections.

Invoking Workflows

Workflows can be invoked in one the following ways:

  • Intents: When an end user writes or says something, referred to as an end-user phrase, the most relevant intent matching the phrase is identified. This process is known as Intent Classification. An Intent identifies an end user's goal. For each application, you define many Intents, and a group of Intents then handle a complete conversation. The flows associated with the Intents are then invoked. The associated flows are also called the Fulfillment of the Intent. While both types of flows can be associated with Intents, the Intents can also be fulfilled in other ways.

  • Events: The flows can be invoked by the Aisera Platform on predefined schedules or systems events (see Event Studio).

Workflow Lifecycle

Every flows goes through the following stages during its lifecycle,

  • Build: During this stage, business process SMEs collaborate with developers to build flows. Often, SMEs (as citizen developers) use Aisera's Flow Studios to build the flows themselves,

  • Test: Next, the flows undergo robust functional and reliability testing before they are deployed in production,

  • Publish: Tested flows are published to production. Additionally, these flows can be exported and imported for use across applications, and select applications are also published to the flow library,

  • Configure: The published flows are attached to Intents or Event Handlers for invocation during bot conversations or from system/webhook events,

  • Monitor: Progression of requests through the flow funnel is continuously analyzed, enabling ongoing flow improvements.

Last updated