Create an Assistant in Gen AI Builder

Suggest edits

Who is this for

Platform users building AI-driven Assistants in Gen AI Builder in Hybrid Manager AI Factory. Typical users include developers, AI architects, conversational designers, and content owners.

What you will accomplish

You will create an Assistant — an AI-powered agent that can:

  • Interact with users through natural language
  • Retrieve knowledge from Knowledge Bases
  • Follow behavioral guidelines using Rulesets
  • Optionally perform actions via Tools

You will configure the core components of an Assistant and prepare it for use in production or testing.

Why create an Assistant

Assistants provide the interactive layer for your AI Factory content and capabilities — enabling trusted, conversational access to your organization’s knowledge and services.

They unify:

  • Large Language Models (LLMs)
  • Knowledge Bases
  • Rulesets
  • Tools
  • Memory (for context-aware interactions)

Assistants allow end users to interact with your AI in a guided, governed, and useful way — powering applications such as internal copilots, support agents, and knowledge assistants.

For background, see:

Complexity and time to complete

  • Complexity: Moderate
  • Estimated time: 10–15 minutes

When to create an Assistant

  • When you want to expose Knowledge Bases through a conversational interface.
  • When you want to control AI behavior and tone through Rulesets.
  • When you need the AI to perform external tasks via Tools.
  • As the first step in building an AI-powered experience (copilot, chatbot, API agent, etc.).

Pre-requisites

  • At least one Knowledge Base (optional but strongly recommended).
  • One or more Rulesets (optional but strongly recommended).
  • (Optional) Tools configured and available.

How to create an Assistant

1. Navigate to Assistants

  • In Gen AI Builder UI, go to Assistants.
  • This is typically found under APPS or Agent Studio.

2. Create a new Assistant

  • Click Create Assistant.

3. Configure Assistant fields

Name (required)

  • Provide a unique, descriptive name.
  • Example: PG Financial Client Support AI

Description (optional)

  • Describe the Assistant’s purpose and target audience.
  • Example: Virtual support agent for PG Financial clients.

LLM Model Selection (required)

  • Select the Large Language Model (LLM) to power the Assistant.
  • Example options:
  • GPT-4
  • Claude 3 Sonnet
  • Gemini Pro
  • Custom models deployed via Model Serving

System Prompt / Instructions (optional)

  • Define the Assistant’s persona or high-level behavior.
  • Example: You are a helpful and friendly customer support agent for PG Financial.

Knowledge Bases (optional)

  • Select one or more existing Knowledge Bases to enable Retrieval-Augmented Generation (RAG).
  • Example: PG Financial Product FAQs KB, PG Financial Account Services KB

Rulesets (optional)

  • Select one or more existing Rulesets to define behavioral guidelines.
  • Example: Polite Support Tone, Financial Advice Disclaimer

Tools (optional)

  • Select Tools the Assistant can invoke (if Tools are configured).
  • Example Tools:
  • Calculator
  • Calendar lookup
  • Custom API connector

Memory Configuration (optional, advanced)

  • Configure how the Assistant remembers conversation context.
  • Typical options:
  • Number of turns to remember
  • Summarization or summarization + truncation strategy

Other Advanced Settings (optional)

  • Temperature (controls response creativity)
  • Top_p (controls nucleus sampling)
  • Tool usage settings
  • RAG behavior settings (advanced tuning)

4. Finalize creation

  • Click Create to save the Assistant.
  • The new Assistant will appear in the Assistants list.

Example: Create a Support Assistant

  1. Navigate to Assistants.
  2. Click Create Assistant.
  3. Configure fields:
  • Name: PG Financial Client Support AI
  • Description: Virtual agent providing customer support for PG Financial.
  • LLM Model: GPT-4
  • System Prompt: You are a polite and helpful customer support agent for PG Financial.
  • Knowledge Bases:
  • PG Financial Product FAQs KB
  • PG Financial Account Services KB
  • Rulesets:
  • Polite Support Tone
  • Financial Advice Disclaimer
  • Tools:
  • Account Balance Lookup API (example)

  1. Click Create.

  2. Next steps:

  • Test the Assistant using the built-in test chat UI.
  • Refine Rulesets, Tools, and Memory settings based on results.

Troubleshooting

Assistant not responding as expected

  • Verify that:
  • Knowledge Bases contain relevant content.
  • Appropriate Rulesets are assigned.
  • System Prompt aligns with desired behavior.

RAG behavior inconsistent

  • Check Retriever configuration and Knowledge Base indexing.
  • Verify that Tools are configured correctly if used.

Memory not working as expected

  • Review Memory settings.
  • Adjust summarization parameters if applicable.

Tools not functioning

  • Verify Tool configuration and API availability.
  • Confirm that the Assistant has the Tool assigned.

Related topics

← Prev

Configure the Data Lake in Gen AI Builder

↑ Up

Gen AI How-To Guides

Next →

Create a Knowledge Base in Gen AI Builder

Could this page be better? Report a problem or suggest an addition!