You just landed a massive client with a complex product catalog. They have thousands of SKUs, specific pricing tiers, and custom inventory counts. You load everything into a massive spreadsheet. You feel incredibly organized.
Then, a hot lead messages your client at midnight asking for a specific part number. Your brand new AI bot is live, but it completely freezes. It only knows what it scraped from the website. It cannot read your beautifully organized spreadsheet. The bot tells the prospect to call back tomorrow, and that hot lead instantly goes to a competitor. You had the right data, but it was locked away in a static file.
Your AI needs direct access to your structured data. HighLevel just completely solved this with Table Search in the Knowledge Base. You can now upload your massive CSV files and let your AI query them using natural language. Your bots can instantly answer complex questions about customers, inventory, or any other tabular data. Let us give your AI a massive brain upgrade.
Part 1: Prepare Your Data File
Before you upload anything, you need to make sure your spreadsheet is clean and formatted perfectly.
- Save your file strictly as a .csv format. UTF-8 encoding is highly recommended.
- Ensure your file is under the maximum 50 MB size limit.
- Keep your data within the system limits of 50,000 rows and 500 columns per location. You will be able to select the 20 most relevant columns for indexing.
- Make absolutely sure your first row contains clear column names. The AI needs these headers to understand the context of the data.
- Clean your data completely. Remove any null values, hidden formulas, or merged cells. If the data is messy, the bot will get confused. If your CSV gets rejected, open it in a basic text editor to inspect for hidden formatting errors.
Part 2: Upload to the Knowledge Base
Ready to streamline this? Here is exactly how to do it in 60 seconds:
- Inside your HighLevel dashboard, navigate to AI Agents and click on Knowledge Base.
- Edit an existing Knowledge Base, or click Create Knowledge Base and give it a clear name and description.
- Click Add Source, then choose the Tables option.
- Upload your clean CSV file using the drag-and-drop tool or the file picker.
Part 3: Map, Index, and Test
Now it is time to connect the dots and test the output.
- Review the detected columns on your screen. The system uses automatic schema detection with a high confidence threshold, but you can adjust the data types manually if needed.
- Click Done to index the table. You will see a progress bar showing the chunking status.
- Attach this newly updated Knowledge Base to your specific AI Agent. This feature works perfectly for Chat, Voice, or Workflow AI.
- Test the bot with a sample natural language question, like asking which customers have overdue invoices, to confirm it pulls the exact right answer directly from your rows.
Stop letting your bots fail because they cannot read your spreadsheets, and start turning your static data into dynamic conversational assets.
Frequently Asked Questions
What should I know about prepare Your Data File?
Before you upload anything, you need to make sure your spreadsheet is clean and formatted perfectly. Save your file strictly as a .csv format. UTF-8 encoding is highly recommended. Ensure your file is under the maximum 50 MB size limit. Keep your data within the system limits of 50,000 rows and 500 columns per location. You will be able to select the 20 most relevant columns for indexing. Make absolutely sure your first row contains clear column names. The AI needs these headers to understand the context of the data. Clean your data completely. Remove any null values, hidden formulas, or merged cells. If the data is messy, the bot will get confused. If your CSV gets rejected, open it in a basic text editor to inspect for hidden formatting errors.
How do I upload to the Knowledge Base?
Ready to streamline this? Here is exactly how to do it in 60 seconds: Inside your HighLevel dashboard, navigate to AI Agents and click on Knowledge Base. Edit an existing Knowledge Base, or click Create Knowledge Base and give it a clear name and description. Click Add Source, then choose the Tables option. Upload your clean CSV file using the drag-and-drop tool or the file picker.
What should I know about map, Index, and Test?
Now it is time to connect the dots and test the output. Review the detected columns on your screen. The system uses automatic schema detection with a high confidence threshold, but you can adjust the data types manually if needed. Click Done to index the table. You will see a progress bar showing the chunking status. Attach this newly updated Knowledge Base to your specific AI Agent. This feature works perfectly for Chat, Voice, or Workflow AI. Test the bot with a sample natural language question, like asking which customers have overdue invoices, to confirm it pulls the exact right answer directly from your rows. Stop letting your bots fail because they cannot read your spreadsheets, and start turning your static data into dynamic conversational assets.
Start Your HighLevel Free Trial
Browse all GoHighLevel guides · AI & Chatbots.