I am a long-term ChatGPT Plus user and the managing director of a German company. I use ChatGPT extensively for construction projects, business administration, document analysis, and complex legal matters.
With the upcoming expansion of autonomous ChatGPT capabilities, I believe one fundamental feature deserves much greater priority: reliable, persistent project memory built automatically from complete document collections.
My proposal
1. Full folder import via drag-and-drop
Users should be able to drag an entire project folder into ChatGPT, including all subfolders, PDFs, emails, images, correspondence, and other documents.
ChatGPT should handle the technical processing, indexing, and organization automatically, without requiring users to create ZIP files, split folders, or upload documents individually.
2. Automatic creation of a structured project knowledge base
ChatGPT should independently analyze the documents, identify duplicates, establish chronological timelines, recognize relationships and contradictions, and build a structured, searchable project record.
Every conclusion should remain traceable to its original source.
3. Truly persistent and reliable project memory
Future conversations should consistently build on the complete project history, including previously established facts, decisions, documents, and outstanding issues.
Users should not repeatedly have to explain information they have already provided.
4. Complete export, backup, and restoration
Users should be able to download the entire AI-organized project archive, including original files, indexes, document relationships, timelines, and work progress.
This backup should be fully restorable, allowing users to recover their complete project knowledge base without information loss.
Why this matters
I have personally attempted to prepare a very large legal project archive for ChatGPT. Even creating a ZIP file proved impractical due to the volume of material. Manually splitting and organizing the archive would take many hours and could easily result in missing information.
This is exactly the kind of work AI should handle.
The ideal workflow would be:
Drag in a complete folder → automatic processing → structured project knowledge base → persistent memory → continuous updates → complete downloadable backup.
Final thought
Autonomous AI capabilities are valuable, but autonomy without reliable memory has serious limitations.
What is the benefit of an AI assistant independently executing complex tasks if it cannot consistently remember the information, decisions, and context from previous work?
For business owners, lawyers, accountants, engineers, and other professionals, this functionality could fundamentally change how ChatGPT is used.
I believe this would be a major competitive advantage for OpenAI and a meaningful step toward making ChatGPT a genuinely reliable long-term professional assistant.
I hope the OpenAI product team will consider this proposal.
Hüseyin Yil
Managing Director, ISO-Bauelemente GmbH
Germany