Automated AI Agent Building
Budget: ₹37,500 – ₹75,000 INR
Every month I need an AI-driven workflow that picks up my existing Google Docs and Notion pages, applies my predefined update templates, and pushes the refreshed versions straight into a custom GPT on OpenAI. Most of these files are heavy on tables, charts, and well-layered headings, so the agent must keep every structure intact while reorganising any clutter it finds.
Key actions the flow must cover end-to-end:
• Locate and read the target documents from both Google Workspace and Notion
• Apply my provided templates to update numbers, text blocks, and any embedded tables or charts
• Re-order sections when the template dictates, ensuring headings remain properly nested
• Export the revised documents and feed them automatically to the designated custom GPT instance
I am tool-agnostic—Claude, n8n, Make, Python scripts, or any combination that reliably meets the above steps is fine, as long as the whole chain runs unattended once a month and sends me a short completion log.
Acceptance criteria:
1. On demand or on schedule, the agent updates at least two sample docs per platform without manual intervention.
2. Table formats, chart links, and heading hierarchy are preserved exactly.
3. The custom GPT receives the new versions and confirms ingestion via its API response.
4. A concise log (success/failure per file) lands in my inbox.
Provide a brief outline of your proposed architecture and any libraries or no-code tools you intend to use so I can green-light the build.
Key actions the flow must cover end-to-end:
• Locate and read the target documents from both Google Workspace and Notion
• Apply my provided templates to update numbers, text blocks, and any embedded tables or charts
• Re-order sections when the template dictates, ensuring headings remain properly nested
• Export the revised documents and feed them automatically to the designated custom GPT instance
I am tool-agnostic—Claude, n8n, Make, Python scripts, or any combination that reliably meets the above steps is fine, as long as the whole chain runs unattended once a month and sends me a short completion log.
Acceptance criteria:
1. On demand or on schedule, the agent updates at least two sample docs per platform without manual intervention.
2. Table formats, chart links, and heading hierarchy are preserved exactly.
3. The custom GPT receives the new versions and confirms ingestion via its API response.
4. A concise log (success/failure per file) lands in my inbox.
Provide a brief outline of your proposed architecture and any libraries or no-code tools you intend to use so I can green-light the build.