Wobby Docs
HomepageSign inCreate account
  • 🎊Welcome to Wobby
  • quick start
    • Access Wobby
    • Recruiting your first AI analyst
  • Agent
    • Creating an Agent
      • Link data sources to an Agent
      • Agent Instructions
      • Choose an LLM
    • Asking the Agent questions
      • How to ask data questions
  • data sources
    • Connect a data source
      • CSV files
      • PostgreSQL
      • MySQL
      • Snowflake
      • BigQuery
      • Microsoft SQL Server (MS SQL)
      • MariaDB
    • Metadata
  • Query templates
    • Quick Overview
    • Creation & Approval
    • How Agents Use Query Templates
  • Integrations
    • Slack
    • Teams
  • SECURITY & privacy
    • How Wobby protects your data
Powered by GitBook
On this page

Was this helpful?

  1. Agent
  2. Creating an Agent

Choose an LLM

When building an agent in Wobby, you’ll need to choose which large language model (LLM) it will use to process queries. Each LLM has different strengths—selecting the right one depends on your agent's typical workload and complexity of tasks.

Tips for Selecting

  • Start simple. If you're unsure, begin with GPT-4o—it offers a strong balance for most use cases.

  • Speed vs. depth. Need fast responses? Try GPT-4o Mini. Need deeper analysis? Consider Claude 3.5 or 3.7.

  • You can switch anytime. Your agent’s LLM can be updated later from the Agent Settings tab.

💡 Agents using GPT-4o will generally provide the most reliable results across a wide range of business intelligence tasks.

PreviousAgent InstructionsNextAsking the Agent questions

Last updated 2 months ago

Was this helpful?