Jadiroon ships with an AI generation feature for every major content task. The AI Prompts settings section (/{org}/app/settings/ai-prompts) lets HR administrators tune the voice, tone, and policy instructions baked into each of those generators — without touching the system or writing code.
What this section does
Each AI generator in the platform has two layers:
- The output shape — the fields, sections, and JSON structure the model must return. This is fixed by the system and cannot be changed. It ensures every generator's output slots cleanly into the rest of the product.
- The core prompt — the persona, tone, language, cultural framing, and policy instructions the model uses while filling in that fixed shape. This is fully editable.
The AI Prompts section surfaces the editable core for every generator. You can read the current default, edit it, save your version, and reset to the default at any time.
Editable generators
The following generators each have an editable core prompt:
Hiring
- Job description generator
- LinkedIn job post generator
- CV–job description match narrative
Training
- Training position description generator
- Trainer CV match narrative
Assessment and development
- Situational judgment tests (SJTs)
- Behavioral assessment narratives
- Cognitive assessment narratives
- Development report — employee profile section
- 70:20:10 plan recommendations
Talent framework
- Competency recommendations (for a role or employee)
- Competency proficiency ladders
How to edit a prompt
- Go to Settings → AI Prompts inside your workspace.
- Select the generator you want to adjust from the list.
- The current core prompt (or the system default if you have not edited it yet) appears in the editor.
- Edit the text directly — set the persona, tone, language formality, cultural notes, compliance rules, or any standing instructions you want the model to follow.
- Click Save to apply your version workspace-wide. All subsequent generations for that feature use your prompt.
Per-prompt chat assistant
Each prompt editor includes a built-in chat assistant. Use it to draft or iterate on the core prompt without writing from scratch:
- Describe what you want in plain language ("make the JD generator sound more formal, always write in Modern Standard Arabic, and avoid any gender-specific pronouns").
- The assistant drafts a revised core prompt based on your description.
- Review the draft, edit it further if needed, and click Save when it is ready.
The chat assistant does not change anything until you explicitly save.
Resetting to the default
Each prompt has a Reset to default action. Clicking it restores the system default for that generator. Your previous custom version is discarded. The reset takes effect immediately for subsequent generations.
What you cannot change
The output structure — the fields, sections, word-count budgets, and JSON schema each generator must produce — is fixed by the system. This ensures that every generated output integrates correctly with the rest of the product (pipeline cards, development reports, assessment summaries, and so on). No edit to the core prompt can override the output shape.
Access
Only HR administrators can view and edit AI prompts. Line managers and employees cannot access this section.
Changes apply workspace-wide and take effect immediately for all subsequent generations — there is no per-user or per-role scoping.
Next
Workspace settings — the catch-all for per-workspace configuration — is the next page in this section.