Skip to Content

Prompts

A prompt is an explicit phrase or a clear instruction provided to BMC HelixGPT based on which the desired output or answers are returned. A prompt can be a single sentence or a paragraph. A user with the BMC HelixGPT Admin role can create prompts in the HelixGPT Agent Studio.

BMC HelixGPT provides different categories of prompts:

  • Global prompt
  • Skill-specific prompts

BMC HelixGPT provides out-of-the-box prompts for all applications that BMC HelixGPT supports. You can start using the out-of-the-box prompts as is. You cannot modify the out-of-the-box prompts. To use them for your requirements, create copies of the prompts and make changes to the copies that you made.

To learn more about the types of prompts, see Out-of-the-box prompts in BMC HelixGPT.

Related topics

Getting started

Setting up and going live

Creating and managing prompts

Agentic and chain-based prompts

BMC HelixGPT supports two prompting paradigms—chained and agentic:

  • Chained prompting uses a predefined, linear sequence of prompts. Each step provides specific instructions and expected outcomes, making this approach predictable and well-suited for scripted workflows or decision trees.
  • Agentic prompting follows the Reasoning and Action (ReAct) paradigm. A single prompt defines the goal and offers high-level guidance. The agent then autonomously determines how to achieve the goal—choosing tools, retrieving data, and asking clarifying questions as needed. This enables dynamic, context-aware interactions that adapt to user needs and environments.

Agentic prompting provides the following benefits: 

  • Autonomy—Agents can reason and act without predefined steps. 
  • Flexibility—Prompts can be reused across scenarios and user types. 
  • Scalability—Supports diverse employee roles and devices. 
  • Clarity—Encourages goal-oriented communication rather than procedural scripting. 

 

 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*

BMC HelixGPT 25.4