# Configurable AI Checklist Prompts
### Introduction
Configurable AI Checklist Prompts enable Admins to customize how AI interacts with checklists in Unifize. This feature allows organizations to define AI behavior for specific checklist items and ensure that AI-generated outputs align precisely with process requirements, language standards, and compliance expectations.
By configuring AI prompts, admins can automate how checklist data is extracted, generated, and structured from contextual information like descriptions, conversations, and uploaded documents within a record.
### How it works
Admins can create AI-powered checklist fields by defining:
* The input data the AI will read from Your input field can also be a File upload doc, which can be added in the input field
* The output fields the AI will populate (e.g., product details, root cause, corrective actions).
* The prompt or instruction that guides the AI on how to process the input and generate the output.
This configuration is applied at the checklist level within a process and is fully customizable per use case
### Setting up configurable AI Checklist prompts
#### Step 1: Add an AI field to the Checklist
1. Go to Manage View section and open the process where you want to add the AI functionality. You can either create a new process or use an existing process
2. Inside your process, navigate to the Checklist section and click on the “Add Fields” button
3. Select “AI” as the field type
4. Give your AI Checklist field a name to move forward with your process
#### Step 2: Configure AI field settings
Move to the settings section for your AI Checklist field. In the AI field settings tab, you will see:
* **Button Label:** Name the action button for your AI field. This will appear to the end user (e.g., “Generate AI suggestions”).
* **Add Prompt:** Write the instruction for the AI. This tells the AI what to do with the input. Example:\
\&#xNAN;*“Take the input from the input fields and generate relevant outputs for the output fields. Give the product name, the quantity of the product in numbers, select the reason from the picklist, give the product number in the Product Number field, and write the corrective actions in bold.”*
* **Select Input Fields:** Choose fields the AI will reference for context. You can choose from your list of existing checklist fields.
* **Select Output Fields:** Choose checklist fields the AI will populate
### What AI can help you with for different Checklist fields
| **Checklist field** | **What AI can do** | **What AI cannot do** |
| -------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Text / Rich Text** | Populate with generated summaries, descriptions, narratives like root cause or corrective actions. | — |
| **Number** | Fill in numeric values like quantities, counts, or measures. | — |
| **Date** | Populate specific dates (e.g., date of report, incident date). | — |
| **Picklist** | Select a specific picklist option **if explicitly mentioned in the AI prompt.** | Cannot dynamically interpret or choose from indexed options. Requires exact instruction (e.g., *“Select ‘Product Defect' for Picklist section”*). |
| **Linked Record** | Create a new linked record to a specified process if clearly stated in the prompt (e.g., *“Link to Customer Complaint”*). | Cannot read embedded fields within the linked record. Cannot suggest relationships based on existing linked data. |
| **File Upload** |
Can parse through documents added in the File Upload Checklist field and fill Checklist data based on the document. Can be chosen as an input field for AI
|
Cannot create or upload files Won't work if chosen as an output field for AI
|
#### Step 3: Save the AI Field
* Click **“Update Field Settings”** to save the AI configuration into the checklist.
### How end users experience it
1. In any record, users see a button labeled (e.g., “Generate AI Suggestions”) on the checklist.
2. When clicked, the AI reads the content from the selected input fields (such as “What happened?” or other descriptions).
3. The AI processes this input based on the prompt configured by the admin.
4. AI-generated suggestions appear in the designated output fields, filling details like Product Name, Quantity, Reason, Corrective Actions, and Root Cause.
5. Users can review, edit, accept, or modify these AI-suggested outputs before saving the record.
6. Input field can be upload a doc, then a doc can be uploaded
***
{% @arcade/embed flowId="FR0v9AeS8RugKtGUMwbT" url="" %}
### Best Practices
* Write clear, specific prompts that match the intent of the checklist output.
* Choose input fields that contain comprehensive context (e.g., problem descriptions, incident summaries).
* Use output fields that benefit from structured data like picklists, dates, quantities, or narrative text like root cause analysis.
* Always encourage users to review AI-generated content for accuracy before saving.
* Apply AI prompts to reduce repetitive data entry for processes like complaints, CAPAs, supplier audits, or investigations.
---
# 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.unifize.com/release-notes/june-2025/configurable-ai-checklist-prompts.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.