Skip to main content

What Is an Agent?

An agent in Amplifi is a customized AI assistant that can interact with your data using natural language. Each agent can be tailored to a specific task or dataset by selecting the tools and LLM (large language model) it uses.

Agents make it easy to create purpose-built assistants for specific business needs — like answering support queries, interpreting financial documents, or navigating product manuals.

Creating an Agent​

Follow these steps to set up your own agent in Amplifi.

Step 1: Access the Agent Creation Screen​

  1. Navigate to the Agents tab from the left sidebar.
  2. Click the Create New Agent button in the top-right corner.

Step 2: Fill in Agent Details​

  1. Agent Name:
    Enter a name for your agent (max 30 characters).

  2. Agent Description (optional):
    Briefly describe what this agent does (max 150 characters).

  3. Select LLM Provider:
    Choose the language model that will power this agent.
    Available options may include:

    • GPT-3.5 – Fast, basic tasks only. Not for image/audio.
    • GPT-4.1 – General-purpose with stronger reasoning.
    • GPT-4o – Best for text, image, and audio tasks.
    • o3-mini Reasoning – Lightweight, optimized for deep reasoning.
    • OpenAI GPT-5 – Latest-generation model with advanced reasoning capabilities.

    Select the model that best fits your use case and data type.

  4. Select Tools to Associate:
    Pick one or more tools that the agent will use.
    These tools are created earlier in the Tools tab and define how the agent accesses and interacts with datasets.

  5. Agent Instructions:
    Add system-level instructions that guide the agent's behavior. Use 0-shot or 1-shot prompting for best results. Be concise, specific, and direct in your instructions (20-30 words).

    • Example: Always be polite, concise, and fact-based. Use 1-shot examples when needed. Respond in under 3 sentences unless more detail is requested.

Step 3: Create the Agent​

Once all fields are filled, click Submit to create the agent.
Your new agent will now appear in the Agents list and can immediately be used for search or chat.

Managing Agents​

  • Edit Agent: Update the name, model, tools, or instructions anytime.
  • Delete Agent: Remove agents that are no longer needed.
  • Search Agents: Use the search bar to quickly locate existing agents.

Tips for Creating Effective Agents​

Here are essential best practices for building powerful agents in Amplifi:

🎯 Define Clear Purpose​

  • Give agents specific, well-defined roles with clear boundaries
  • Choose appropriate LLM models based on complexity needs (GPT-3.5 for simple tasks, GPT-4o for multimodal, o3-mini for deep reasoning)

🔧 Strategic Tool Selection​

  • Attach only relevant tools that directly support the agent's purpose
  • Test tool combinations to optimize performance and accuracy
  • Ensure tools have access to appropriate datasets

📝 Craft Effective Instructions​

  • Write concise, comprehensive system prompts with behavioral guidelines
  • Include domain-specific context and terminology
  • Set clear response standards for format and detail level

🧪 Test and Refine​

  • Start with simple queries and progress to complex scenarios
  • Use realistic test cases that match actual user needs
  • Monitor performance and iterate based on feedback
  • Track accuracy, response quality, and user satisfaction

📊 Monitor and Maintain​

  • Regularly review agent performance and usage patterns
  • Update instructions and tool configurations as needed
  • Maintain documentation for team collaboration
  • Scale deployment gradually with user feedback

Your agent is now ready to assist! 🚀
You can start interacting with it from the search or chat interfaces depending on the tool configuration.