Generate with AI

Top  Previous  Next

Using artificial intelligence to structure job accountabilities and technical requirements rather than relying solely on manual, human-led assessments. This process converts descriptions into structured and standardized data.

 

Generate with AI

The 'Generate with AI (Sparkles Emoji)' icon can be found in the 'Add' job detail screen or the 'Edit' job detail screen of the text (WYSIWYG) editor. This can be accessed only with an Access Level 5 (administrative users) and 4.

 

addjob01a

 

Click the pencil icon to enter the WYSIWYG editor.

To the left of the screen is an AI icon to generate text in the selected field.

oIn this case, Nature of the Work

 

AIGenerateicon_editor

 

The AI will generate content for these fields using prompts that will pick up on the following:
1. Nature of the Work - Working conditions, physical demands, hazards
2. General Accountabilities - Overall function and purpose (2-3 sentences)
3. Specific Accountabilities - Primary functions and duties (8-10 statements) including how that goal is achieved (3-4 sub-bullets)
4. Technical Requirements - Minimum skills, certifications, qualifications

 

 

Steps for Generation 

 

Select a Job you would like to use the AI for.

Double click the job or click on the edit icon to open the detail screen. 

Click the pencil icon ,pencil2_zoom80_zoom80by the text field to open up the editor.

Within the WYSIWYG editor, the AI icon will be available in the editor toolbar. 

 

Generate Output

Click the AI Generate icon (Sparkles Emoji)

Wait 2-5 seconds for AI to process (pulsating progress indicatorPulsatingAIIndicator)

Generated content appears in the text field section

Iterate and Refine
1. Review the generated output that appends to what you currently have in the text field if an existing job. 

Note: AI will not overwrite your accountability statements or technical requirements. You will want to refine the statements generated to be specific to your organization.

 

GeneralaccountabilitygeneratedwithAI

 

2. If not satisfied, edit the prompt text in the AI Generation Settings panel under custom icon.
3. Click "Generate with AI" to try again with the updated prompt.
4. Repeat of required.

 

Click Save save to save AI Settings for all Users or cancel cancel

Click OK,ok,  to ensure your text field has been save.

 

 

AI Generation Settings

Use the AI Generation Settings, AIGenerationSettings (only available to those with an Access Level 5) to:

Choose the Default, defaultprompticonor Custom, customprompticon Prompt 

Select how you want the AI to be instructed:

oUse Default Prompt - Uses the system's built-in prompt (recommended for organizations)

oUse Custom Prompt -  Write your own instructions for how AI should generate content

 

When using a custom prompt:
1. Click Use Custom Prompt icon (customprompticon)
2. The text editor will appear, pre-filled with the default prompt as a starting point
3. Edit the prompt text to match your organization's needs
 

AIGenerationSettings2

 

Shorter prompts (200-500 characters) tend to produce better results

Focus on 'what' you want, not 'how' the AI should think.

Be specific about format requirements (bullet points, paragraph, etc.)

 

Note:

Always review generated content before using it in production job descriptions:

Check for accuracy

Ensure appropriate tone and professionalism

Verify it matches your organization's voice

 

Understanding the Messaging
 "AI service is currently unavailable"
The AI service may be down or experiencing issues. Try:
1. Wait a few minutes and try again
2. Check with your system administrator
3. Verify your organization's AI service configuration

 

 

AIMessage

 
"Rate limit exceeded"
You've made too many generation requests. Wait for the countdown to complete (shown in the error message), then try again.

 
"Generation failed"
The AI encountered an error processing your prompt. Try:
1. Simplifying your prompt
2. Checking for unusual characters or formatting
3. Using the **Regenerate** button (results may vary)

 
Generated output is poor quality
If AI generates low-quality content:
1. Refine your prompt to be more specific
2. Add more context about what you want
3. Test on different jobs to see if it's job-specific or prompt-specific
4. Consider reverting to the default prompt as a starting point

 

Canceling AI Generation

If you decide you want to cancel the generation of the AI text, click the Sparkles Emoji icon and the following message will generate.

 

AIMessage2

 
Save Changes
Changes are not saved automatically. 
 

Single Prompt Architecture

The AI generation uses a single prompt approach.

One Prompt: All context, job data, and instructions are combined into a single system prompt

User Instructions Embedded: The user's custom instructions (if any) are injected directly into the prompt in the "Instructions for Writing the '{Field}' Field" section

Minimal User Message: The actual user message sent to the AI is simply "Generate now."

 

"View Prompt" button, infobutton1: Shows the complete prompt that will be sent to the AI, including job context and instructions

 

ViewPrompt

How it Works

1.The system builds a comprehensive prompt containing:

oRole and context (Encompassing Visions software)

oCurrent job data (title and existing fields)

oFormatting rules

oInstructions section - either default field-specific guidelines OR user's custom instructions

2.When user clicks "Generate with AI":

oIf "Default" mode: Uses built-in field-specific writing guidelines

oIf "Custom" mode: User's custom instructions replace the default guidelines in the prompt

3.The user can click "View Prompt" to see exactly what will be sent to the AI, including their custom instructions

The default instructions for each of the job fields only pertains to the Encompassing Visions default terms. So if the default term is changed, for example -  Nature of the Work to something else, the user will need to be sure to switch the Instructions from Default to Custom and enter something there that's appropriate for the new term.

Note: One thing, that is set up for the AI: name of the new field. You can see this when viewing the prompt.

When you save a custom prompt, it saves it for that field (such as "Reports To:" or whatever the Nature of the work field is), and it saves it for all jobs. Same with the model: it saves it on a per-field basis, for all Jobs. If there are certain fields that have explicit per-field model settings, they do not pick up on the system default-they will use there own configured model.

To fix: If you want General Accountabilities and Nature of the Work to use your system default (Gemini 3 Pro), you need to clear their per-field model settings:

1.Go to Jobs > Edit a job > AI Settings (the WYSIWYG editor settings)

2.For General Accountabilities, set the model dropdown to "Default"

3.For Nature of the Work, set the model dropdown to "Default"

Once those are set to Default, they will inherit from the system default like the other fields.

Note: You may need to refresh the Jobs.aspx page after changing a default for the change to take effect

 

When reviewing the AI prompt, you may choose to try different models. This can be done through the AI Evaluator Settings.

Click on the AISettingsEvaluatoricon icon found on the AI Generation Settings screen. Located at the bottom left. This is the workbench for testing and comparing AI outputs.

AIGenerationSettings4

 

AI Policy Settings

If the organization chooses not to use AI generation, for their Job descriptions, there are policies set up under Job Settings>Job Description to enable/disable them.