Intelligent AI Connector
Allowing AI to participate in your Processes.
AI is useful when you use it as a tool, but it becomes transformative when you integrate it into your workflows. By making AI an active participant in your processes, you can stop managing the task and let the system handle it entirely. In Frends, AI can be included in the Processes natively using Intelligent AI Connector, automating tasks that have so far required human input.
Intelligent AI Connector
Intelligent AI Connector is a new building block for Processes in Frends. It allows the AI to participate in your Processes by performing actions you have specified as prompts, in the location of the Process you need to.
Multiple prompts can be provided for the single shape, making sure the provided answer follows the required format and is repeatable. By default, Frends system prompt is given to the shape in order to make the AI work with Frends ideology and Processes specifically, instead of providing the answers in generic text or other unsuitable format. From there, it's up to the developer to specify the action taken by AI.
Why Use AI Connector
BPMN 2.0 with AI Connector enables AI Orchestrations. Practically, you can implement the thought process of an employee as steps and decisions. Instead of requiring a human decision for categorizing an incoming message into tech or sales support cases, AI Connector with LLM models can read and categorize the message for you automatically. The Frends Process can then continue handling the case into correct destination. The whole process for multiple different ticket cases can be written as a series of prompts, decisions, and actions that combine the basic operations from Tasks and programming with input from AI.
Choose your AI Model
The AI Connector in Frends is not tied to a single available model. While Frends provides its own AI services with a curated set of LLM Models to use for various use cases that work out of the box and no configuration, you can also provide your own Azure Inference service, or take a step towards owning your own data even further and use an on-premise Ollama installation to perform the AI actions.
Agentic AI
The Intelligent AI Connector now supports agentic mode, enabling it to perform MCP Tool invocations directly within the AI step rather than requiring each action to be modeled as a separate Task. In agentic mode, the AI Connector enters a reasoning loop. It evaluates context, calls an available MCP Tool, receives the result, and determines whether the goal has been met or whether a further action is needed. This continues until the objective is satisfied or the configured reasoning depth limit is reached.
Agentic behavior is activated by configuring tools on the AI Connector — either by pointing it at internal MCP Tools published within Frends, or by connecting it to an external MCP server. Without Tools configured, the Connector operates as a standard single-response step.
Reference Documentation
In order to learn the specifics of how to use the AI Connector in your Processes, head over to our reference documentation for AI Connector, which will give details of the shape.
Last updated
Was this helpful?