How to Generate Content with AI

The **Content Generator** is an AI drafting tool that turns a short prompt into message or document text you can copy into your own templates, so your agency never starts from a blank page.

### Overview

The **Content Generator** page sits under **Admin** > **Messaging** in eCourtDate. A configuration form runs down the left side of the page (**Template**, **Your Prompt**, **LLM Provider**, **Model**, and an **Advanced Settings** panel), and the conversation area on the right shows each prompt you send labeled **You** and each reply labeled **AI Assistant**.

The generator drafts text and nothing else. It does not save templates, schedule anything, or send a message to a client. The only action available on a reply is the copy button, so you copy the text and paste it wherever you need it.

The **Template** select starts on **Custom Prompt** and groups its starter prompts into four categories:

| Group | Templates |
|-------|-----------|
| **Case Communications** | **Court Date Reminder**, **Case Status Update**, **Document Request** |
| **Event Notifications** | **Hearing Notification**, **Event Cancellation Notice**, **Filing Deadline Reminder** |
| **Document Generation** | **Client Welcome Letter**, **Process Instructions**, **FAQ Response** |
| **General** | **Professional Communication**, **Simplify Legal Language** |

Selecting a template writes that template's starter prompt into **Your Prompt** for you to edit.

### Prerequisites

- An eCourtDate account that can open the **Admin** menu in the top navigation bar (the menu is hidden for accounts whose role hides it)
- A clear idea of the message or document you want drafted, including the audience and the key details to include
- A destination for the finished text, such as a message template, an auto message, or a document, because the generator does not save its output
- Time to review the draft, since AI output needs a human check before it reaches clients

### How-To Steps

1. Click **Admin** in the top navigation bar.
2. Select **Messaging** from the dropdown menu.
3. Click the **Content Generator** card.
4. Select a starter prompt in the **Template** select, or leave it on **Custom Prompt** to write your own. You can also click one of the template cards shown on the empty page, which fills the prompt and scrolls back to the form.
5. Edit the text in **Your Prompt** so it describes your agency, the situation, and every detail the draft must include.
6. Select an option in **LLM Provider**: **Auto (Best Available)**, **OpenAI**, **Anthropic (Claude)**, or **Google (Gemini)**.
7. Select a **Model** if you picked a named provider. Choosing a provider selects that provider's first model for you, and **Auto (Best Available)** leaves the list empty so the platform picks for you.
8. Click **Advanced Settings** to show **Temperature** and **Max Tokens** when you want to tune the output.
9. Drag the **Temperature** slider between 0 and 1 (the default is 0.7). Higher is more creative, lower is more focused.
10. Enter a value in **Max Tokens** to set the maximum response length (the default is 500).
11. Click **Generate**.
12. Read the reply that appears under **AI Assistant** in the conversation area and check the wording, the tone, and every fact it states.
13. Click the copy button on the reply to copy the text to your clipboard.
14. Paste the text into the message template, auto message, or document where you want to use it.
15. Click **Clear History** and confirm the prompt when you want to empty the conversation area.

### What to Expect

While the request runs, the **Generate** button shows **Generating...** with a spinner. Your prompt appears in the conversation area under **You**, and the reply follows under **AI Assistant** with a relative timestamp such as "Just now" or "5 mins ago". Replies are rendered as markdown, so headings, lists, and bold text in the draft display as formatted text.

The form resets after every generate. **Your Prompt** empties, **LLM Provider** returns to **Auto (Best Available)**, **Temperature** returns to 0.7, and **Max Tokens** returns to 500, so set anything you changed again before your next request.

The conversation lives in the page only. Reloading the page, navigating away, or clicking **Clear History** removes it, so copy anything worth keeping before you leave.

### Best Practices

- **Give the prompt real detail.** Name your agency type, the scenario, the audience, and the facts the draft must contain. A vague prompt returns generic text.
- **Start from a template, then edit.** The starter prompts already set a professional government tone, so adjusting one is usually faster than writing a prompt from scratch.
- **Tune Temperature to the job.** Keep **Temperature** low for formal, repeatable notices and raise it when you want more varied phrasing.
- **Raise Max Tokens for longer pieces.** Letters and instruction guides often run past the 500 default and get cut short otherwise.
- **Swap the bracketed placeholders for real merge tags.** Starter prompts use placeholders such as [Case Number], [Date], and [Location], which are examples rather than working eCourtDate merge tags.
- **Review every draft before you use it.** AI output can be fluent and still wrong, so verify legal and procedural details against your agency's standards.

### Frequently Asked Questions

**Q: Does the Content Generator send anything to your clients?**

A: No. It only writes text into the conversation area on the page. Nothing is sent until you copy the text into a message template or auto message and send it through your normal messaging setup.

**Q: Can you save a reply as a message template from this page?**

A: No. The copy button is the only action on a reply. Copy the text and paste it into the template, auto message, or document where you want it.

**Q: Why is the Model list empty?**

A: The **Model** list is filled from the provider you select, and **Auto (Best Available)** has no models to list. Select **OpenAI**, **Anthropic (Claude)**, or **Google (Gemini)** to see model options, or leave **Auto (Best Available)** selected and let the platform choose.

**Q: Where did your prompt and settings go after you clicked Generate?**

A: The form resets after each generate. Your text moves into the conversation area under **You**, and **Your Prompt**, **LLM Provider**, **Temperature**, and **Max Tokens** return to their defaults.

**Q: Is the conversation still there tomorrow?**

A: No. The conversation is held in the open page only, so a reload or a page change clears it. There is no saved history to reopen.

### Troubleshooting

**Issue:** The **Generate** button will not click.

**Symptoms:** The button is greyed out and clicking it does nothing.

**Solution:**

1. Enter text in **Your Prompt**, because an empty prompt disables the button.
2. Select an option in **LLM Provider** if the field is blank.
3. Wait for a request that is still running to finish, since the button stays disabled while it shows **Generating...**.

**Issue:** Text you typed disappears when you change the **Template** select.

**Symptoms:** **Your Prompt** is replaced by a starter prompt, or empties when you select **Custom Prompt**.

**Solution:**

1. Choose the template first, then edit the prompt, because each change to **Template** overwrites the field.
2. Copy your own wording out of **Your Prompt** before switching templates.

**Issue:** The reply stops in the middle of a sentence.

**Symptoms:** The draft ends abruptly, usually on longer letters or instruction guides.

**Solution:**

1. Click **Advanced Settings** to open the panel.
2. Enter a larger value in **Max Tokens** than the 500 default.
3. Click **Generate** again with the same prompt.

**Issue:** No reply appears and an error message is shown.

**Symptoms:** The spinner stops and the conversation area still shows only your prompt.

**Solution:**

1. Check your internet connection, since the generator needs one to reach the AI provider.
2. Click **Generate** again, and wait a moment first if the error mentioned too many requests.
3. Select a different **LLM Provider** and try once more.
4. Create a ticket by clicking the ticket button in the bottom right corner of the screen if the failure continues.

### Related Articles

- [How to use ai generator](/how-to-use-ai-generator)
- [How to create message templates](/how-to-create-message-templates)
- [How to Create an Auto Message](/how-to-create-an-auto-message)
- [How to Use Merge Tags](/how-to-use-merge-tags)
- [How to generate documents automatically](/how-to-generate-documents-automatically)