Can I share context for Katalon AI agent across my team

AI Testing
1

Is it possible to create and share Agent context across my team?

What I want to know is if I setup the context for Katalon AI to think and reply in a specific manner and it should reply the same way to my teammates within same org.

It would be easier to use as a team and get more consistent replies from AI assistant.

2 Likes
2

Problem Analysis

You’re asking whether Katalon supports creating and sharing custom AI context (system prompts/behavior instructions) at the organization level so your entire team gets consistent AI responses. Based on my research of current Katalon documentation and features, here’s what’s actually available:

What IS Available for Team-Level AI Configuration

Organization-Level AI Services :white_check_mark:

Katalon does support organization-wide AI configuration, but it’s limited to AI provider and model selection, not custom context/prompts:

At the organization level, you can configure:

  • Enable/disable AI features for all team members
  • Select which AI provider your organization uses (Katalon Default, OpenAI, Azure OpenAI, AWS Bedrock, Gemini)
  • Set organization-specific API keys that all members use

How to set this up:

  1. Go to Account Settings > System > Configurations > AI Services
  2. Toggle Enable AI Features
  3. Select Organization-Specific AI Key and configure your preferred provider
  4. All team members in that organization will use these settings

Individual-Level Prompt Customization (Partial Solution)

Katalon does offer prompt customization, but it’s per-user, not organization-wide:

In Katalon Studio, each user can customize:

  • Prompt Library (v10.2.3+) - Customize engineering prompts for:
    • Ask mode (general questions)
    • Agent mode (agentic testing behavior)
    • Generate code
    • Explain code
    • Analyze test failure

How to set this up:

  1. Go to Katalon Studio > Preferences > Katalon > StudioAssist > Prompt Library
  2. Edit prompts to include your specific context and instructions
  3. These customizations apply only to that user’s instance

What’s NOT Currently Available

:cross_mark: No built-in feature to:

  • Create organization-level custom system prompts
  • Share individual prompt customizations across team members
  • Define how AI should “think and reply” at the organization level
  • Enforce consistent AI behavior across all team members

Recommended Workarounds

Since organization-level prompt sharing isn’t available, here are practical alternatives:

Option 1: Document & Share Prompt Templates

  • Create a shared document (Confluence, GitHub Wiki, etc.) with your custom prompts
  • Have team members manually copy these into their individual Prompt Library settings
  • Include guidelines on how to structure queries for consistent responses

Option 2: Use Organization AI Keys

  • Configure a shared organization AI key with your preferred provider
  • This ensures all team members use the same AI model and provider
  • Combine with documented prompt guidelines for consistency

Option 3: Standardize via Project Context

  • Enable Auto-include project context in StudioAssist preferences
  • Include Object Repository and Custom Keywords automatically
  • This provides consistent context without custom prompts

References

3

I will loop @huy.tieu, our AI expert to this convo ^^

4

You can set up Katalon to reply in a certain way and share this setup with your team. This way, everyone will get similar responses, making it easier for the whole team to work consistently.

1 Like
5

this achievable in copilot. not sure on katalon

1 Like
6

afaik, the short answer is not possible yet

from my experience:

  • Org settings only control AI provider/model
  • Prompt/context is per user, can’t be shared

what to do instead:

  • Create a standard prompt template
  • Share it (Confluence/Git)
  • Team manually copies into Prompt Library

also enable auto project context → helps keep outputs consistent

7

Hi! We’re rolling out context and knowledge management soon, but I think from a logical model point of view, sharing (setting up standards) within a project is the first priority since a project encapsulates an atomic unit/team/product that may have distinct standards. But I am sure @huy.tieu will take above-project context into consideration.

1 Like