AI Brand Setup: Document Upload & Extraction
AI Brand Setup: Document Upload & Extraction
Available from: v1.0.54
NurtureHub's AI Brand Setup now accepts PDF and Word documents as a source of brand intelligence. Upload brochures, service guides, team bios, or marketing materials and the platform will extract and structure the content, merging it into your agency's brand profile alongside data from your website crawl.
Overview
When you set up your brand profile, NurtureHub crawls your website to learn your agency's tone, services, and identity. Document upload extends this capability: any written materials your agency already produces can now feed directly into that same profile.
This is especially useful for agencies whose website doesn't fully reflect their brand — for example, agencies that rely on printed brochures, detailed landlord or vendor guides, or staff bios that live in internal documents rather than on their site.
Supported File Types
| Format | Extension |
|---|---|
.pdf | |
| Word Document | .docx |
Files that do not match these types are rejected at upload with a validation error. Other formats (e.g. .doc, .pptx, .txt) are not currently supported.
How It Works
- Upload — Select one or more PDF or Word files in the Brand Setup screen.
- Extraction — The file is parsed server-side:
- PDFs are processed with
pdf-parseto extract raw text. - DOCX files are processed with
mammothto extract readable text content.
- PDFs are processed with
- AI structuring — Extracted text is analysed by the AI layer, which identifies brand-relevant content: tone of voice, service descriptions, team information, key selling points, and similar signals.
- Profile merge — Structured data from the document is merged into the existing brand intelligence profile, supplementing (not replacing) any data already gathered from your website crawl.
- Incremental uploads — You can upload additional documents at any time; each one is processed independently and its data is added to the profile.
What to Upload
Any document that communicates how your agency presents itself will be useful. Good examples include:
- Vendor or landlord guides — Service overview documents sent to new clients
- Company brochures — Printed or digital marketing materials
- Team bios — Staff profiles or about-us documents
- Area guides — Location-specific content that reflects local expertise
- Lettings or sales packs — Information packs describing your process and fees
Documents do not need to be formatted in any particular way. The extraction process handles varied layouts and structures.
Relationship to Website Crawl
Document uploads are supplementary to the website crawl. Both data sources contribute to the same brand intelligence profile:
- The website crawl runs automatically when you connect your website during Brand Setup.
- Document uploads can be added before or after the crawl completes.
- If both sources contain information about the same topic (e.g. your fee structure), both are retained and the AI layer uses the combined picture when generating email content.
Incremental Processing
Multiple documents can be uploaded across separate sessions. Each upload is processed independently — you do not need to re-upload previous documents when adding a new one. The brand profile grows incrementally as new documents are added.
Limitations
- Only
.pdfand.docxformats are supported; other file types are rejected. - Scanned PDFs (image-only, non-searchable) may yield little or no extracted text, as
pdf-parseextracts text layer content only. - Very large documents may take longer to process; upload progress is shown in the interface.