What is this?
This approach uses AI to generate the front-end structure of an Infocapture form based on a natural language description.
It focuses on:
- Form layout (sections and fields)
- Field types, labels and hints
- Required vs optional inputs
- Default values and select options
It does not include:
- Workflows
- Notifications
- Field conditions or infocapture logic
- Permissions
Why use AI for this?
Creating forms manually, especially large ones, can be time-consuming and repetitive. The use of AI offers a more time-efficient way to produce a form by:
- Converting plain English requirements into structured forms
- Reduce manual effort
- Allow less technical users to configure forms more easily
How does this work?
The process uses AI as a translator between your Infocapture form requirements and Infocapture’s CSV import format.
The process:
- Describe the form to the AI model in plain English
- The AI model analyses your request and clarifies your requirements by asking questions
- The AI produces a structured definition of the form in JSON format
- The JSON output is converted into a CSV file, which can be downloaded and imported into Infocapture
What you will need
Before starting, you must have the following resources available:
1. KB articles that the AI can reference
- Infocapture field types: https://discover.claromentis.com/knowledgebase/articles/432/infocapture-creating-a-new-form
- Using calculations/expressions: https://discover.claromentis.com/knowledgebase/articles/59/expressions-calculating-fields-in-infocapture
2. Infocapture CSV template
This will be used as a reference, so the AI knows what to generate and how it should be formatted. This link will download the Infocapture CSV template.
3. An AI tool capable of generating a structured output
There are various AI tools available to try this out with. Choose one based on your preference and or business policy.
Step 1: Set the AI rules
Paste the following into your AI tool:
You are generating an InfoCapture form.
Rules:
- Only generate the front-end form structure
- Do NOT include workflows, notifications, conditions, or permissions
- Use only valid InfoCapture field types (based on provided reference)
- Do NOT assume missing requirements — ask questions instead
- Always restate the form requirements in plain English before generating anything
- Output must be structured and suitable for conversion into CSV
- Choose field types based on character length:
- Short string = maximum 10 characters only
- Medium string = up to 27 characters
- Anything longer must use a long text / textarea field
- Do not assign short string fields to inputs that may exceed 10 characters
- Infer expected input length from the field description (e.g. names, addresses, descriptions)
Step 2: Provide Infocapture KB articles, which will be used as a reference:
Here are the valid InfoCapture field types and their behaviour:
https://discover.claromentis.com/knowledgebase/articles/432/infocapture-creating-a-new-form
https://discover.claromentis.com/knowledgebase/articles/59/expressions-calculating-fields-in-infocapture
Step 3: Provide your form requirements, and describe the form in plain English
For example:
I need a ‘New Starter’ form to collect information about a new employee who is joining the company.
The form should start by capturing the employee’s basic personal details, such as their full name, personal email address, phone number, home address, and date of birth.
Next, I need a section for their job information. This should include their job title, which department they will be working in, who their manager is, whether they are full-time, part-time or a contractor, their start date, and whether they will be working in the office, remotely, or a mix of both.
There should also be a section for IT and equipment requirements. This should ask whether they need a laptop and what type (Mac or Windows), as well as any additional equipment like monitors or accessories. It should also allow us to capture what systems they should have access to (Basic, Enhanced, Full).
Finally, I need a section for payroll and HR information. This should include their national insurance number, their bank details, and an emergency contact name and phone number.
The form should be clear and structured so that each section groups related information together, and it should make it easy for the person filling it out to provide all the necessary details in one go.
Step 4: AI clarification (Important step)
After providing the requirements, the AI should automatically ask clarifying questions, as provided in the first prompt. If the AI fails to do so, it may make assumptions. Use the following prompt if required:
Before generating anything, ask me any clarification questions needed to fully define the form.
Do not make assumptions.
The AI should ask about:
- Which fields are mandatory vs optional
- Field types (e.g. string, text, radio, dropdown)
- Options for dropdown/radio fields
- Formatting structure
Answer these questions before continuing.
Step 5: Generate structured output (Plain English and JSON)
Once the output has been generated in plain English using the below prompt, carefully review and correct anything before proceeding.
The JSON output is just a structured format that translates your plain English output into something that the AI can work with more accurately.
Produce both of the following:
1. A structured breakdown of the form in plain English, including:
- Form purpose
- Sections
- Fields per section
- Field types
- Required vs optional fields
2. A JSON specification of the form including:
- Sections
- Fields
- Field types
- Labels
- Required flags
- Options (for dropdowns/radio fields)
- Help text if applicable
- Ensure field types reflect expected input length based on the defined character limits
Do not proceed further until I confirm the structure is correct.
Step 6: Provide the Infocapture CSV template
Provide the CSV template to the AI with the following prompt:
Here is an InfoCapture CSV template.
Using the previously generated JSON, create a valid Infocapture CSV file.
Requirements:
- Match the exact column structure from the template
- Do not change column names or order
- Use correct field type values (e.g. RADIO, SELECT)
- Use "Section" correctly for grouping
- Format option values correctly (Label,#value)
- Ensure required fields are marked correctly
Step 7: Import into Infocapture
- Download the CSV the AI provided
- Navigate to the required Infocapture Project on your intranet
- Check out the form
- Import the CSV file
Refer to this guide for more information about how to import the CSV: https://discover.claromentis.com/knowledgebase/articles/428/
Key rules to follow
- Always provide the AI with the CSV template so it can generate a useful copy
- Never skip the clarification step
- Do not allow the AI to guess field types
- Check the requirements output before generating JSON
Video Demonstration
