Quick summary
A custom research assistant GPT lets you configure a purpose-built AI model in ChatGPT with tailored instructions, knowledge uploads, and conversation starters — no coding required. Follow the steps below to create, configure, and publish your own GPT directly from the ChatGPT interface.
Steps
- Navigate to ChatGPT's main interface, open the left-side panel, and click 'Explore GPTs'.
- Click the 'Create' button at the top-right corner to begin building a new GPT.
- Switch the tab from Create to 'Configure' to access the full setup options.
- Click the 'Add' icon to upload a display photo that represents your research assistant GPT.
- Enter a unique GPT Name and Description into their respective input fields.
- Write a comprehensive set of guidelines covering purpose, behavior, and capabilities into the 'Instructions' field.
- Add sample prompts one by one into the 'Conversation Starters' field to guide users on their first interaction.
- Click 'Upload Files' under the Knowledge section to import any reference documents you want the GPT to use.
- Review the Capabilities section and tick the checkboxes for the GPT functions you want to enable; configure any Actions for third-party integrations.
- Go to the Preview Area, type a test prompt, and click 'Submit' to evaluate your GPT's initial performance.
- If satisfied with the results, click the 'Create' button at the top-right corner to finalize your GPT.
- Select your preferred access configuration and tap 'Save' to proceed.
- Your custom research assistant GPT is now created and published and ready for use.
