This is a step-by-step guide to create an Agent called General Knowledge Chat Agent using the New Agent Manually option.
This Agent is a high-performance, general-purpose Assistant built on different chat models, optimized for handling a wide range of tasks including reasoning, writing, factual answers, and open-ended ideation. Its design enables it to deliver precise, coherent, and innovative outputs, making it adaptable to almost any user scenario.
This Agent functions as a versatile conversational system capable of interpreting user intent, generating accurate information, and offering creative suggestions when needed. Leveraging the Chain of Thought reasoning strategy, it breaks down complex queries into logical steps, ensuring well-structured and contextually aware responses.
Functionally, the Chat Agent:
- Provides reliable answers across broad knowledge domains.
- Produces clear explanations, recommendations, and insights tailored to each request.
- Adapts seamlessly to both factual and creative tasks, maintaining clarity and usefulness.
- Ensures polite, safe, and context-sensitive interactions at all times.
- This configuration enables the Agent to serve as an intelligent, adaptable AI companion, capable of delivering high-quality outputs across reasoning, communication, and ideation tasks—fully powered by the capabilities of the selected chat model.
This chat is set up for working with four different chat models:
- GPT-5 is designed as a highly capable, reasoning-oriented model with strong multi-step problem-solving skills and advanced contextual understanding. It excels at producing nuanced, structured explanations and maintaining coherence over long conversations. In a general-knowledge chat, it delivers detailed, well-organized answers with strong analytical depth and high reliability across most domains.
- Gemini 2.5-Pro emphasizes multimodal intelligence, performing exceptionally well with text, images, and complex cross-media reasoning. It is optimized for speed, safety, and retrieval-augmented responses. In a general-knowledge chat, you can expect highly balanced, concise, and context-aware answers with strong integration of factual data.
- Claude Sonnet 4.5 is a refined version of Anthropic’s Claude line, known for its ethical alignment, thoughtful reasoning, and strong writing quality. It often delivers careful, thorough explanations with a calm and deliberate tone. In a general-knowledge chat, it would provide reflective, highly structured answers focused on clarity, safety, and nuance.
- Grok 4 builds on the characteristic openness and edgier personality of the Grok series, combining fast inference with strong access to up-to-date information from public sources. It typically provides more direct, informal, and occasionally humorous responses. In a general-knowledge chat, the experience would feel faster, more conversational, and more personality-driven compared to other models.
First, log in to the Console. In the Project Dynamic combo box, select the Project you want to work with. In this case, Default(DocumTeam) is used.
Next, on the left side of the screen, you will find the Backoffice menu. In this menu, click on The Lab.
A new window opens in the browser with The Lab. Once inside The Lab, the Agents Dashboard opens by default. From there, you must create an Agent by clicking on the New Agent Manually option. This opens the configuration interface where you must define the Agent’s core properties.
In the Configuration Tab of an Agent, fill in the required fields as shown below.
- Agent Name: General Knowledge Chat.
- Agent Purpose: Handles general-purpose tasks such as answering questions, giving recommendations, and generating insights.
- Agent Role: General Knowledge.
- Background Knowledge: You are a highly capable, general-purpose AI assistant, ready to help with any task or question.
- Guidelines:
Main goal: Serve as a versatile assistant for any general-purpose user request.
- Receive and understand user queries on any topic.
- Analyze the intent and context of the request.
- Provide clear, accurate, and helpful responses, explanations, or recommendations.
- If the query requires step-by-step reasoning, break down the answer logically.
- For creative or open-ended tasks, generate ideas or suggestions with clarity.
- Always maintain politeness, respect, and avoid disallowed or harmful content.
- Introduction: Hello! I’m your AI Agent, ready to answer questions and assist with diverse tasks.
- Description: A powerful, general-use Agent designed for reasoning, writing, and ideation, offering precise and innovative results for any use case.
- Conversation Starters:
- 1)
- Ask a factual question
- Who wrote 'To Kill a Mockingbird'?
- 2)
- Request a recommendation
- Can you suggest a good productivity app?
- Features
- 1)
- Feature Title: Universal Knowledge
- Feature Description: Can answer questions and provide information on a wide range of topics.
- 2)
- Feature Title: Flexible Assistance.
- Feature Description: Handles creative, factual, and open-ended requests with ease.
In the AI and Tools Tab of an Agent, complete the following fields to define how the Chat Agent will process and analyze information.
- AI Model: Select either GPT-5, Gemini 2.5-pro, Claude Sonnet 4.5 or Grok 4.
- Reasoning Strategy: Set Chain of Thought. This strategy enables complex reasoning capabilities through intermediate reasoning steps.
- Creativity Level: Set 0.1 to maintain predictable and consistent outputs during research tasks.
- Max Tokens: Define 44,813 to allow long and detailed reports without truncating the output.
- Max Runs: Keep the default value 5 to limit the number of reasoning iterations and avoid unnecessary loops.
After clicking on Create Agent, a confirmation message will appear, and new options will be displayed in the bottom-right corner of the screen. Click on Run Test to verify that the General Knowledge Chat Agent works as intended.
Once the Agent has been tested and its responses are correct, click on Publish to make it available in the Workspace as an Assistant.
How to create an Agent
Since version 2025-11.