# Agentic AI Frends Agentic AI enables AI to take action within your Processes rather than simply producing output for a predetermined BPMN flow to work on. You define a goal, configure which tools the AI is permitted to use, and the AI handles the rest. It selects actions, calls tools, and iterates until the objective is reached or a defined limit is hit.

## Balancing Control & Flexibility For many integrations, deterministic Process written as a BPMN flow is the right design. The logic, branching, and sequence of steps are fully defined by the Process designer, and the Process behaves predictably every time. Agentic AI is intended for a different class of problem: scenarios where the correct sequence of actions depends on context that cannot be fully anticipated at design time. When the decision logic is variable enough that encoding it as explicit BPMN branches becomes impractical, the Agentic AI approach allows the Process to define a goal and a set of permitted tools instead. The AI then determines which tools to call and in what order, adjusting based on what each call returns. The Process remains the outer structure, controlling when the agentic step runs, what data it receives, and what happens with the result.

AI operates as part of your Processes, not replacing them.

The scope of what the AI can do is governed entirely by the toolbox. The AI cannot call anything that has not been explicitly made available. Reasoning depth is capped at a configurable limit, and every step in the reasoning process is recorded in the execution log alongside standard Process execution data. This includes each decision, tool call, and returned observation.

All AI Connector executions include a reasoning log, alongside other logged metadata.

## 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.

Picture showing the Native AI shape as part of a Frends Process. — Intelligent AI Connector in action.

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. [Read more about our Intelligent AI Connector here](/frends-development/agentic-ai/intelligent-ai-connector.md). ## Model Context Protocol Frends supports the **Model Context Protocol (MCP)**, an open standard that defines how AI agents discover and call external tools. In Frends, MCP works in both directions: you can expose your own Processes as MCP Tools for AI to call, and the Intelligent AI Connector can consume MCP Tools from both Frends and external sources.

Learn more about [what MCP is and what it enables for you](/frends-development/agentic-ai/model-context-protocol.md). --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.frends.com/frends-development/agentic-ai.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.