Build custom agents for ClearCo app

Follow

Building custom agents in the ClearCompany application helps your employees get fast, accurate answers to their policy questions and reduces repetitive requests for your HR team.

Setting up ClearCo app custom agents requires IT Admin or User Admin permissions and the custom agents add-on.

NOTE

The ClearCo app and the LMS do not share workspaces. Agents built in the ClearCo app only work in the ClearCo app and cannot be used in the LMS.

Open the Agent Studio workspace

Go to the main setup area where you create and manage custom agents.

Click Tools > Setup > Agent Studio > Add Agents.

agentstudioaddagents 


⚠️ Don’t use your browser’s back button

Navigating away from this page may result in lost work. Use the Cancel or Save buttons on the page instead.

Define the name and purpose details for your agent

Setting up basic details identifies your assistant in the chat window and keeps it focused on the right topics.

  • Add a clear title in the Agent Name field that users will see in the chat window, such as Employee Handbook Assistant or Workplace Safety Coach.
  • Enter a short summary in the Agent Description field so users understand how the agent can help. For example: Answers common questions about company policies and workplace rules.
  • Describe the agent’s main role in the Agent Purpose field to guide how it responds. For example: Use the employee handbook and policy documents to answer employee questions about time off, conduct, and remote work.

NOTE

The Agent UID field is reserved exclusively for internal troubleshooting support; external API connectivity is not yet supported.

Select a personality tone for your custom agent

Choosing a personality tone ensures the agent's voice matches your company's style and the topic of discussion.

  • Friendly & Empathetic tone creates a warm, encouraging voice that works perfectly for new hire onboarding programs or student training.
  • Professional & Knowledgeable tone provides clear, confident answers that fit formal company policies and employee handbooks.
  • Insightful & Analytical tone delivers detailed answers and thoughtful explanations for tracking company data or review reports.
  • Strict & Precise tone ensures firm, exact messaging for legal rules, compliance policies, and workplace safety guidelines.
  • Custom Personality allows you to build unique response guidelines that match your specific company voice.

Control response formatting 

Choose how the agent should organize its responses so the format fits the type of help you want to provide.

  • For short paragraphs and a casual chat style, select Conversational. Great for coaching conversations.
  • For bulleted lists, clear sections, and step-by-step answers, choose Structured. Good for a policy or process agent.
  • For workplace examples and easy-to-follow explanations, use Storytelling. Great for manager support or coaching agents.
  • For short, direct responses, pick Technical. Perfect for a compliance or workplace safety agent.
  • Or, write your own response rules with Custom Formatting. Example: Use short sections, bullet points for steps, and propose recommended next steps.

Set the opening icebreaker

The icebreaker is the welcome message users see when they open the chat. Use it to explain what the agent can help with and give users a clear way to get started.

Go to the Icebreaker Message field and enter a short welcome message that explains what the agent does. Add a sample question to show users what they can ask.

For example, I can answer questions about the employee handbook. Try asking: What is our remote work policy?

Select the language model

Different AI models are optimized for different types of tasks, so the one you choose dictates how fast the agent replies, how well it handles complex logic, and how it formats its text.

Depending on the specific job you want your agent to perform, you can choose from several options:

  • For fast, conversational answers and basic tasks (like a standard employee FAQ or brainstorming partner), choose GPT 5 Mini, GPT 4.1, or GPT 5.1.
  • For highly complex or sensitive topics that require advanced step-by-step reasoning (like navigating nuanced compliance scenarios or coaching managers through interventions), choose GPT 5.1 Thinking or GPT 5.2 Thinking.
  • For reading multi-page documents and generating structured, professional writing (like interpreting extensive company handbooks), choose GPT 5.2.
  • For clear summaries and intricate enterprise tasks (like summarizing broad leadership training feedback), choose Claude Sonnet 4.6 or Claude Opus 4.6.
  • For processing massive data files and running broad audits (like identifying trends in organizational skills matrices), choose Gemini 2.5 Pro.

Attach knowledge libraries

Attach the files your agent should use when answering questions. These knowledge libraries give the agent company-specific information instead of general internet data.

Select the Contexts tab to toggle on the knowledge libraries you want the agent to use.

Set behavioral boundaries for custom agents

Choosing how your assistant talks and acts keeps every chat safe, accurate, and professional. These behavior rules prevent your agent from sharing incorrect facts or discussing off-topic issues. 

Select the Policy tab to set up your safety limits, language standards, and human routing options.

Validate behavior in the preview environment

Use the Preview tab to test your agent before you share it with employees. This helps you confirm that the agent responds in the right tone, follows your safety rules, and stays within scope. 

Next steps

0 out of 0 found this helpful

Comments

0 comments

Please sign in to leave a comment.