SharePoint and Microsoft Syntex content processing
How Microsoft Syntex automates document classification, metadata extraction, and content processing at scale.
Microsoft Syntex (formerly SharePoint Syntex, now also marketed as SharePoint Premium) adds AI-driven content processing to SharePoint Online — automatically classifying documents, extracting structured metadata, generating documents from templates, performing eSignature, and applying retention / sensitivity at scale. For organisations with high-volume document workflows, Syntex is one of the higher-impact Microsoft 365 capabilities of the past several years.
Document understanding
The core capability: train an AI model to classify documents and extract specific fields from them. When new documents arrive at a SharePoint library, Syntex applies the model — categorising each document and extracting field values into SharePoint columns.
Two model types:
Unstructured document processing
For text-heavy documents where layout varies — contracts, agreements, policies, articles. The model learns from example documents you provide. For each new document:
- Classifier identifies the document type — "this is a contract."
- Extractors pull specific fields — start date, end date, parties, total value, jurisdiction.
The extracted fields land in SharePoint metadata columns. Filtering, searching, and aggregating across documents becomes structured.
Structured / freeform document processing
For form-style documents where field locations are more consistent — invoices, applications, receipts. Uses AI Builder form processor under the hood. Provide 5–10 example documents; train the model; apply to incoming documents.
Prebuilt models
Microsoft ships ready-to-use models for common document types:
- Invoices — vendor, invoice number, date, line items, total.
- Receipts — vendor, date, total, line items.
- Business cards — name, company, contact details.
- Contracts — parties, dates, terms (basic).
Prebuilt models are a fast starting point without training.
Content assembly
Generate documents from templates and structured data:
- Word templates with placeholder fields.
- Data source for the placeholder values — Dataverse, SharePoint list, Power Automate input.
- Output document generated with values filled in.
Use cases: contract generation from a customer record, proposal generation from product configurations, personalised letters from a customer list. Replaces manual mail-merge or third-party document-automation tools.
eSignature
Native digital signing in Microsoft 365:
- Request signatures on PDF or Word documents.
- Multiple signers, sequence-controlled.
- Tamper-evident signed documents stored in SharePoint.
- Audit trail of signature events.
Alternative to DocuSign or Adobe Sign for in-tenant signing workflows. Some organisations use Syntex eSignature for internal use, third-party for customer-facing.
Content processing flows
For automated handling at scale:
- Auto-classify documents as they arrive based on a trained model.
- Auto-apply sensitivity labels based on detected content.
- Auto-apply retention for specific document categories.
- Trigger workflows — Power Automate flows fire when classification happens.
Image and video processing
Lighter capability but useful:
- Image tagging — automatic tags for images uploaded to SharePoint.
- OCR on images and PDFs for content searchability.
- Video transcription for accessibility and search.
Often consumed at the pay-as-you-go transaction level — image processed = transaction.
Where Syntex fits with Microsoft 365 Copilot
Syntex and Microsoft 365 Copilot are complementary:
- Syntex makes content machine-understandable — classification, extraction, structured metadata.
- Copilot reasons over content — answers questions, summarises, drafts.
Better Syntex-applied metadata = better Copilot grounding quality. For Copilot-heavy use cases, Syntex investment pays back in answer quality.
Licensing
Syntex has shifted to a pay-as-you-go model:
- Document understanding transactions per file.
- Content assembly per document generated.
- eSignature per signature request.
- Content processing per asset.
Some capabilities also have per-user licensing for organisations with heavy steady use.
When Syntex is the right investment
Most valuable for:
- High-volume document processes — accounts payable processing thousands of invoices, claims processing.
- Contract management with significant document review.
- Regulatory document classification at scale.
- Customer-facing document generation at volume.
Less valuable for:
- Low document volumes — manual processing is fine.
- One-off document scenarios.
- Highly variable document types that resist consistent classification.
For organisations matching the heavy-volume profile, Syntex often pays back the licensing cost within months in saved manual effort.