Best practices for BMC HelixGPT

We recommend that you review best practices for the BMC HelixGPT product prior to implementing the product or a feature.

incl

Best practices for generating prompts

Use the following best practices for generating prompts in BMC HelixGPT:

Limit the topics for your domain. For example, create prompts specific to the IT or HR domain.
Use a friendly, helpful, and formal tone of response.
Prompt HelixGPT to ask for a confirmation when information is asked for entities, such as name, date, and location in long prompts.
Display the knowledge article summaries as paragraphs, bullet points, or numbered lists.
Use headlines or sections for long answers.
Do not create jokes.

Use the following best practices to use the automatically generated prompts in BMC HelixGPT:

A generated prompt works best with the supported models.
For more information about the supported models and the default model configuration values, see Models-in-BMC-HelixGPT.
Create a new skill in HelixGPT to import the automatically generated prompts.
Important
The OOTB sample skills are read-only and the new prompts for publishing the service catalog data can't be added to them.
Copy the instructions of an existing router prompt. Create a new router prompt for the new skill, and reuse the copied instructions for the new router prompt.
Important
If you want to reuse an existing out-of-the-box DWP Router Prompt, you must create your own copy of this prompt with the Read and Write permissions. Also, you must give a unique name to your router prompt. For more information about the available OOTB sample skills and their associated prompts, see Out-of-the-box-skills-in-BMC-Helix-Digital-Workplace.
A generated prompt works best if you use the model that supports JSON response.

Best practices

When importing a service through the Publish Chat-enabled Service wizard, make sure that you use clear, concise, and unambiguous labels.
To achieve the best response from the LLM, use a meaningful service name and description.