Create generative actions in cloud flows
Important
Some of the functionality described in this release plan has not been released. Delivery timelines may change and projected functionality may not be released (see Microsoft policy). Learn more: What's new and planned
| Enabled for | Public preview | Early access | General availability |
|---|---|---|---|
| Admins, makers, marketers, or analysts, automatically | Dec 2024 | - | Apr 2025 |
Business value
Instead of building out complex rules-based processes from beginning to end, you work with AI to inject intelligent activity orchestration from your cloud flow. You create intent-based generative actions where the large language model (LLM) considers all the resources available to it and dynamically selects the most appropriate actions and sequence of these actions to achieve the desired outcome. This allows for more dynamic automation and saves makers time.
Feature details
Generative actions use generative AI reason over unstructured content and determine what steps and sequences are needed to achieve a process outcome based on instructions and parameters provided by you in natural language.
Building a generative action within cloud flows involves the following steps:
- Describe the objective of your process in natural language, and let AI develop an automation plan to achieve it.
- Refine the automation plan generated by the large language model (LLM) and adjust inputs, outputs, and variables as needed.
- Add reference sources and guidelines in natural language that influence the LLM as it executes the flow.
- Validate the intelligent action by reviewing how it is achieving the business objective prior to moving the flow into production.
- Review production run history, analytics, and individual flow run details to ensure your process is performing to expectations.
The ability of generative actions to easily incorporate both structured and unstructured data makes it flexible as information resources are updated. Its reasoning abilities make generative actions uniquely beneficial in the building and maintenance of complex process flows to handle conditions, exception management, and process loops.