You can now upload files in chat, and have your Agent analyze them against its entire knowledge base.
Simply toggle on the new Agentic Action, Document Analyst
Upload any text or image file(s) – transcript, contract, resume, photo, schematic – and analyze it against everything your organization knows.
Or upload several at once and compare them side-by-side.
Tools like ChatGPT and Claude can summarize your document, but they don’t have your full company knowledge and policies at their disposal.
They give generic answers.
Your agent, powered by Document Analyst, gives answers based on your knowledge — without hallucinating.
Businesses like yours already have this in action
Many of our early-access customers gave Document Analyst a try during the beta, and these are use cases where they found the most success:
- Financial analysis – Investors are uploading earnings call transcripts and pulling out KPIs, cash flow trends, and value creation signals – all filtered through their own valuation frameworks.
- Compliance pre-check – Compliance teams drop in marketing materials and get instant flags against their regulatory guidelines. Human review still happens, just faster.
- Contract review – Legal uploads vendor agreements, checks them against their standard templates, and spots what’s missing or off before anyone signs.
- Candidate screening – HR compares batches of resumes against job requirements using their own hiring criteria. Hours of screening, gone.
- Framework application – Consultants and students upload real-world case studies and apply proprietary methodologies to them on the spot.
- Technical support – Engineers drop in spec sheets or incident reports and cross-reference against SOPs and troubleshooting guides without digging through folders.
- Content transformation – Marketing takes dense technical reports and turns them into on-brand blogs and emails using their voice guidelines.
You can find more ideas and real-world examples in the Document Analyst Best Practices User Guide.
And it’s not just text
Document Analyst sees images too – and understands them in the context of your knowledge base. Here’s an example of what you can do:
- Upload a property photo and find similar listings across your database.
- Share a technical schematic and analyze it against your engineering documentation.
- Snap a product image and verify it matches catalog specifications.
Hours of work. Minutes to complete.
You know that compliance review that usually takes half a day? Or the context you’d need to grab from three different colleagues before you could even start? That’s all baked in now.
Your team spent years building institutional knowledge. Now it’s available for every single file someone uploads.
Finally, and every answer comes with citations – so you’re never blindly trusting the AI. You can verify every statement.
How to enable
Simply go to your agent’s Actions page and switch the toggle.
Your users will see an attachment icon in chat. They will be able to upload their documents, ask, and get answers grounded in your business data.
Available on Premium and Enterprise and works on all deployment types – including API.
Connect With Us
If you want to see how other businesses are using Document Analyst to transform their workflows, or share your own use cases, join the CustomGPT.ai Slack Community today.
FAQs
Q: What is the Document Analyst?
A: The Document Analyst is a powerful agentic feature that allows end-users to upload files directly into a conversation with your CustomGPT.ai agent. The agent then analyzes the document against its private knowledge base – your data, your guidelines, your standards – to deliver contextual insights that generic AI tools simply cannot provide.
Q: What makes Document Analyst different from uploading files to ChatGPT or Claude?
A: Generic AI tools analyze your document in isolation – they don’t know your business context. Document Analyst evaluates every file against your CustomGPT.ai agent’s private knowledge base, giving you analysis grounded in YOUR regulations, YOUR standards, and YOUR institutional knowledge. Generic AI gives generic answers. Your AI agent gives YOUR answers.
Q: Who is the Document Analyst for?
A: Document Analyst is designed for businesses that want to save time on document review tasks and/or monetize intelligent document analysis for their customers. Ideal use cases include compliance verification, contract review, resume screening, competitive analysis, quality inspection, and any workflow where documents need to be evaluated against established.
Q: What file types can I upload?
A: Document Analyst supports PDF (.pdf), Word documents (.docx, .doc), OpenDocument (.odt), plain text (.txt), and image formats including JPG, JPEG, PNG, and WebP.
Q: What is the maximum file size?
A: Premium plan users can upload files up to 5MB. Enterprise customers can request extended file size limits through their account manager.
Q: Is there a limit on document length?
A: Yes. Document Analyst processes up to 3,000 words per document on Premium plans. Content beyond this limit is automatically truncated. Enterprise customers can request extended word limits.
Q: Can I upload multiple documents at once?
A: Yes. Premium plans support uploading 1 file per prompt by default. Enterprise customers can enable multi-file uploads – up to 3 files per prompt by default, with a maximum of 10 files available upon request. Note that the word/token limit applies to all files combined, not per file. Files are processed in the order uploaded until the limit is reached.
Q: How does the word limit work with multiple files?
A: The 3,000-word limit (or your custom enterprise limit) applies to all uploaded files combined, not per file. Files are processed in the order they’re uploaded. If your combined content exceeds the limit, later files may be partially or fully truncated.
Q: How does Document Analyst handle images?
A: Document Analyst features Advanced Vision Intelligence. When you upload an image, the AI uses your specific question to deeply understand the visual content – far beyond generic image description. This enables precise analysis of complex visuals like technical schematics, architectural blueprints, and product photos.
Q: What image use cases does Document Analyst support?
A: Supported use cases include analyzing architectural plans and blueprints against building codes, comparing product photos against catalog specifications, interpreting charts and graphs from reports, reviewing screenshots of software or error messages against troubleshooting guides, understanding scientific and technical diagrams, and matching real estate or product images against your visual knowledge base.
Q: Can Document Analyst search for similar images in my knowledge base?
A: Yes. If your knowledge base includes vision-processed images, Document Analyst can compare uploaded photos against them – enabling powerful use cases like finding similar real estate properties, matching products across catalogs, or identifying visual patterns.
Q: How do I enable Document Analyst?
A: Go to your agent settings, click Actions, find Document Analyst, and toggle it On. That’s it – your users will immediately see an attachment icon in the chat interface.
Q: How do users upload files?
A: Users click the attachment icon next to the chat input field, select their file, and send their question. The agent analyzes the document and responds with contextual insights.
Q: Does the agent explain its reasoning?
A: Yes. Document Analyst provides chain-of-thought reasoning, showing you how it reached its conclusions and which sources from your knowledge base informed its analysis.
Q: Are uploaded files stored permanently?
A: No. Files are only stored temporarily during processing. They are not added to your agent’s knowledge base and are not retained after the session.
Q: Which plans include Document Analyst?
A: Document Analyst is available on Premium and Enterprise plans. Standard plan users do not have access to this feature.
Q: How does Document Analyst affect my query usage?
A: Document Analyst uses advanced multi-step reasoning, which requires more computational resources. Each document analysis counts as 10 queries (1 base query + 9 agentic action cost). Enterprise customers can negotiate custom pricing based on volume.
Q: Is Document Analyst available via API?
A: Yes. Document Analyst is fully supported through the CustomGPT.ai API, enabling seamless integration into your existing applications and workflows.
Q: Does Document Analyst work on mobile devices?
A: Yes. Document Analyst works everywhere your CustomGPT.ai agent is deployed – web, mobile, embedded widgets, Slack, and API integrations.
Q: Can enterprise customers get higher limits?
A: Yes. Enterprise customers can request extended limits for file size, document word count, and image resolution. Contact your account manager to discuss custom configurations.
Q: Is custom pricing available for high-volume usage?
A: Yes. Enterprise plans offer volume-based pricing with custom agentic action modifiers. High-volume users can achieve significantly lower per-analysis costs.
Q: How secure is Document Analyst?
A: Security is a top priority. Uploaded files are processed in a secure environment and stored only temporarily during analysis. Files are never added to the agent’s permanent knowledge base or shared with other users.
Q: Who can see the files I upload?
A: Only you. Uploaded files are accessible only within your conversation session. Neither the agent’s owner nor CustomGPT.ai administrators can view files uploaded by end-users.
Q: What happens if I upload an unsupported file type?
A: The system will display an error message and prompt you to upload a supported file format (PDF, Word, ODT, TXT, or image files).
Q: What happens if my document exceeds the word limit?
A: Content beyond 3,000 words is automatically truncated. The agent analyzes only the first 3,000 words. For longer documents, consider splitting them into sections or contact your account manager about extended limits.
Q: What if my file is larger than 5MB?
A: Files exceeding 5MB will be rejected with a “File too large” message. Compress your file or split it into smaller sections. Enterprise customers can request higher limits.