Synthesize Examples

In this section, we'll use Entry Point's Synthesis feature to create new examples for our training dataset. This can be useful when you only have a limited amount of data and want to use AI to generate more.

The Synthesis feature requires you have at least 3 examples already, since it works by using an internal prompt with few-shot learning.

Create a Template

Navigate to the templates section on the left-hand sidebar and press + in the top-right.

Leave the Type set to "Chat". For the System Prompt, we'll enter the same message we used during the previous Transform Data step in this guide.

You are a friendly and helpful support agent. Your job is to do the initial intake of support events. You will gather additional information from the customer before we connect them with a human agent. 

Please start the conversation with the customer by: 
- Greet them warmly - thank them for reaching out, apologize for the inconvenience they are experiencing, assure them that we'll get them taken care
- Tell them you're an AI agent and that you'll get them connected with the support team after they answer a few very brief questions

Then, ask the customer the questions below: 

- How long have they been experiencing the issue? 
- Have they taken any steps to try to resolve it already? If so, ask them to explain briefly. 
- Are they able to replicate the issue? Does it happen all the time, or only intermittently? 

Format the list of questions as bullet points to make it easier to read. Also, if the initial client outreach already contains an answer to any of the three questions, please do not ask them to provide that information again.

Stick to the basics and keep things concise. Use a casual tone; there's no need to be super formal. Do not use emojis.

Set the user message to the "Inbound Request" Field and the assistant message to the "AI Agent Response" Field. Then, press "Save".

Your template will be named automatically. Press the "..." menu and select "Rename" to give it a different name.

Synthesize Examples

Navigate to the synthesis page using the sidebar menu and press "Settings".

In this example, we'll use the following settings. For the template dropdown, be sure to select the template that we created in the previous step.

• Platform: OpenAI

• Model: GPT-4 Omni

• Template: Select the Template we created in the previous step

• Alignment Text: Leave blank

• Total # of examples to create: 12

• Batch size: 3

• Validation examples: 20%

• Automatically save examples: Leave unchecked

Press "Save & close" and then "Start" in the top-right.

Add Synthesized Examples to Dataset

Entry Point will synthesize new examples for your dataset. You can review the examples and add them manually to your dataset.

Once you are comfortable with the quality of the results, you can choose to run a new batch with the box checked next to "Automatically save examples".

This allows you to easily synthesize as many training examples as you'd like and automatically add them to your training dataset!

Note that Entry Point provides an estimate allowing you to see how your model provider costs scale as you increase the size of your batch.