> For the complete documentation index, see [llms.txt](https://docs.veza.com/4yItIzMvkpAvMVFAamTf/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.veza.com/4yItIzMvkpAvMVFAamTf/features/lifecycle-management/how-to/create-with-ai.md). # Create with AI for Lifecycle Management Use Access AI to draft and refine condition strings and formatters from inside the Lifecycle Management workflow editor. {% hint style="info" %} **Early Access:** Create with AI for Lifecycle Management is available to opt-in tenants. Contact Veza Support to enable it in your environment. See [Before you start](#before-you-start). {% endhint %} {% hint style="info" %} **AI-generated content:** Create with AI uses generative AI. The LCM Policy Agent drafts condition strings and formatters from your description, and its suggestions can be inaccurate or incomplete. Review every suggestion and its explanation before you apply it, and validate the result with a [Simulation Dry Run](/4yItIzMvkpAvMVFAamTf/features/lifecycle-management/policies-workflows/policies.md#simulation-dry-run-on-policy) before you save the workflow. {% endhint %} ## Overview The **Create with AI** button in the Lifecycle Management workflow editor opens an Access AI chat thread with the **LCM Policy Agent**. Describe the condition or the transformed value you want in plain language, and the agent returns the matching syntax with a short explanation. The suggestion appears as an action card in the Access AI chat panel, and nothing changes until you apply it. Applying a suggestion updates the form field only; the change is saved when you save the workflow. The button appears next to these fields: * **Condition strings** * The **Identity matches this condition** clause in a workflow's trigger panel. * A condition node, when its condition type is not **Always**. * **Formatters** * The **Formatter** of an attribute on a **Sync Identities** action, once a destination attribute is set. Structured attributes and the fallback formatter have no button. * The unique identifier attribute of a **Sync Identities** action. * The **Formatter** of a common transformer. Other actions, such as **Create Email** and **Deprovision Identity**, do not offer the button. ## Before you start Create with AI requires all of the following. If any is missing, the **Create with AI** button does not appear. * **Tenant enablement**: Veza Support enables Access AI, Access AI chat, and the LCM Policy Agent for your tenant. * **Early access option**: each user turns on the **Access AI Chat** early access option, which is off by default. * **The Administrator role.** Only administrators can edit Lifecycle Management workflows. See [Roles](/4yItIzMvkpAvMVFAamTf/administration/administration/users/roles.md). The LCM Policy Agent reads the policy's configuration with your own permissions, not an elevated service identity. It does not read identity or run data, and it cannot save a workflow. A suggestion takes effect only when you apply it and then save the workflow yourself. ## Create a condition string with AI 1. Open **Lifecycle Management**, select a policy, and open the workflow in the workflow editor. 2. Open the field to draft: * For the trigger condition, click the **Trigger** box. In the trigger panel, find the **Identity matches this condition** clause. * For a condition node, select the node on the canvas. 3. Click **Create with AI**. The Access AI chat panel opens. If the field is empty, the agent drafts a new condition string. If it has a value, the agent refines it. 4. Describe the condition in plain language, for example: "Active Workday workers in the Engineering department whose start date is within the next 7 days". 5. Review the **Condition String** action card, which contains the suggested condition and an explanation. 6. Click **Apply Condition** to write the suggestion into the field, replacing its current value. The card then reads **Condition Applied**. To decline the suggestion, click **Reject**. The card then reads **Rejected**, and you can keep chatting to ask for another suggestion. 7. Review the value, and save the workflow. {% hint style="warning" %} **Apply Condition** writes to the condition field that is open in the workflow editor, not necessarily the one you opened the thread from. Before you apply a suggestion, confirm that the intended trigger panel or condition node is open. {% endhint %} **Apply Condition** is available only while a condition field is open on a workflow editor page. On any other page, the card shows **Go to Workflow** instead, which returns you to the workflow. ## Create a formatter with AI 1. Open **Lifecycle Management**, select a policy, and open the workflow in the workflow editor. 2. Select a **Sync Identities** action and find the attribute to format. For a common transformer, open it for editing instead. 3. Click **Create with AI** next to the **Formatter** field. The Access AI chat panel opens, and the agent receives the destination attribute, the entity type, and the current formatter if one is set. 4. Describe the value in plain language, for example: "Lowercase the first initial and full surname, joined with a period, then append `@example.com`". 5. Review the **Formatter for** *attribute* action card. Pipeline functions in the suggestion appear under **Then apply:**. 6. Click **Apply Formatter** to write the suggestion into the **Formatter** field, replacing its current value. The card then reads **Formatter Applied**. To decline the suggestion, click **Reject**. 7. Review the value, and save the workflow. When the suggestion includes pipeline functions, applying it also replaces the **Then Apply** value. When it has none, the existing **Then Apply** value is left unchanged, so clear it yourself if it no longer applies. **Apply Formatter** is available only while a formatter field for the same destination attribute is on screen. If you close the attribute or move to another action, return to it to apply the suggestion. The formatter card has no **Go to Workflow** button. ## Conversation threads Each click of **Create with AI** starts a new Access AI thread. The LCM Policy Agent keeps the context it was opened with, such as the policy and the field, across every turn in that thread, so follow-up requests do not need to restate it. To return to a past thread, open **Access AI** > **Threads** in the main navigation. To rename, move, or delete a thread, use **Rename Thread**, **Move Thread**, or **Delete Thread** in the thread's options menu. ## Related * [Policies](/4yItIzMvkpAvMVFAamTf/features/lifecycle-management/policies-workflows/policies.md): workflow triggers, conditions, and actions. * [Trigger Conditions Reference](/4yItIzMvkpAvMVFAamTf/features/lifecycle-management/policies-workflows/actions/trigger-conditions-reference.md): SCIM filter syntax for condition strings. * [Transformer Reference](/4yItIzMvkpAvMVFAamTf/features/lifecycle-management/transformers/transformer-reference.md): formatter functions and the Then Apply pipeline. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## 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, and the optional `goal` query parameter: ``` GET https://docs.veza.com/4yItIzMvkpAvMVFAamTf/features/lifecycle-management/how-to/create-with-ai.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. 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.