Turn quarterly PDFs into a clear earnings brief saved to Google Docs. Built for finance teams that need quick updates on the last three quarters with clear trends and outliers.

A manual run starts by reading file links from a Google Sheet. Google Drive downloads each PDF. The workflow splits the text into small chunks and creates vector data using a Gemini embeddings model. It stores this data in a Pinecone index for fast search. An AI agent uses a vector store tool to find facts from the last three quarters, compares the numbers, and writes a markdown summary. The agent writes with OpenAI while the search tool runs with a Gemini chat model. The final text is inserted into a Google Doc you select.

You will need accounts for Google Sheets, Google Drive, Google Docs, OpenAI, Google Gemini, and Pinecone. Connect credentials in n8n, point the Sheet and Doc to your files, and keep the Pinecone index settings from the template. Many teams cut hours of reading and copying and can prepare board notes, investor updates, or team briefings in minutes.

• On demand run using a manual trigger for instant insights
• Reads file links and names from a Google Sheet to control inputs
• Downloads earnings PDFs from Google Drive for processing
• Splits long documents into smaller parts to improve search
• Creates embeddings with a Google Gemini model for vector search
• Stores and retrieves data from a Pinecone vector index
• AI agent uses a vector store tool to focus on the last three quarters
• OpenAI writes a clear markdown report with trends and outliers
• Saves the final output into a chosen Google Doc
• Processes multiple files through a loop for scale

• Reduce manual reading and notes from 3 hours to 10 minutes
• Automate 90 percent of report drafting for quarterly reviews
• Cut copy and paste mistakes by 80 percent with source backed answers
• Handle 10 or more PDFs in one run with the batch loop
• Connect six systems in one flow for end to end reporting
• Produce a consistent report format every time for easier review

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Google Sheets, Google Drive, Google Docs, OpenAI, Google Gemini and Pinecone. See the Tools Required section above for links to create accounts with these services.
3. In the n8n credentials manager, connect Google Sheets OAuth2, Google Drive OAuth2, and Google Docs OAuth2. Double click each related node, choose Credential to connect with, click Create new credential, and follow the on screen steps.
4. Create API keys for OpenAI, Google Gemini, and Pinecone. Get your OpenAI key from the OpenAI dashboard. Create a Google AI key in Google AI Studio. Create a Pinecone key in the Pinecone console.
5. In Pinecone, create an index that matches the embeddings model size. Use the same index name configured in the template.
6. Open the List Of Files To Load node and set the Document ID and Sheet Name to your Google Sheet. Make sure your sheet has a File URL column and a 10Q column or update the node fields to match your headers.
7. Open the Download File From Google Drive node and map the File URL field to the sheet column. Set the file name field to the column that holds the display name. Confirm your Drive links are accessible by the connected account.
8. Check the text splitting settings in the Recursive Character Text Splitter. Default chunk size and overlap work for most earnings PDFs.
9. Open the embeddings nodes and select the Gemini embeddings model. Confirm credentials are set to your Google AI key.
10. Open the Pinecone Vector Store nodes and confirm insert mode for loading and the same index for retrieval. Make sure your Pinecone credential is selected.
11. Open the AI Agent and review the prompt. Update the company name if needed. Confirm the agent uses your OpenAI credential.
12. Open the Vector Store Tool and ensure it connects to the retrieval nodes. Confirm the tool uses the Gemini chat model as configured.
13. Open the Save Report to Google Docs node and paste the target Document URL. Keep the update operation and insert action that writes the agent output.
14. Run a test with the manual trigger. Watch the run, then open your Google Doc to confirm a new summary was inserted. Check Pinecone to see that vectors were created.
15. If you see empty results, verify the Sheet ID and tab, Drive link access, and that the Pinecone index has data. If the report is too brief, add at least three recent quarterly PDFs.