Your documents know
more than your people can find.
AI-powered classification, auto-tagging, and intelligent document processing — natively inside Microsoft 365.
Most organisations spend years filing documents into SharePoint. They’re still spending just as long looking for them. Manual classification doesn’t scale. Neither does the risk that comes with getting it wrong.
Document management breaks down
long before anyone notices.
It starts as a filing convention problem. It becomes a compliance problem. By the time it becomes a legal problem, the document you need either can’t be found or can’t be trusted.
Search that doesn’t find
When documents are filed inconsistently — different folder names, no metadata, no content type — search returns noise instead of answers. Staff work around it with email chains and personal copies. The “single source of truth” exists only in theory.
Compliance you can’t verify
Retention schedules are only enforceable if documents are correctly classified. When classification is manual and inconsistent, records are either deleted too early or kept forever — both of which create regulatory risk. Audits become painful exercises in reconstruction.
Manual tagging that doesn’t happen
Asking users to classify documents on upload is a reasonable policy that fails in practice. Under time pressure, people skip it. Classification accuracy degrades. Remediation projects get scoped, budgeted, and then quietly deprioritised — until the next audit.
Migration debt that compounds
Every platform migration — file shares to SharePoint, SharePoint On-Premises to Online — carries classification debt forward. Thousands of documents arrive in the new environment with no metadata, no content type, and no retention assignment. The problem restarts at scale.
AI that can’t reach your documents
Microsoft Copilot and SharePoint Agents are only as useful as the content they can access and understand. Unclassified, untagged document stores produce poor AI answers — or no answer at all. Document quality is the prerequisite for AI quality.
Vendor lock-in from third-party DMS
Dedicated document management systems solve the classification problem — but introduce a parallel platform, a separate licence, and a migration dependency. Organisations that are already investing in Microsoft 365 often find they’ve duplicated capability rather than extended it.
AI document management
built inside Microsoft 365.
Capella is Antares’s document management solution built natively on SharePoint Premium and SharePoint Agents. It classifies documents automatically, applies metadata without human intervention, enforces retention policies, and deploys AI agents that can answer questions about your document estate — all without a third-party platform sitting between your content and your people.
SharePoint Premium content AI does the reading.
SharePoint Premium (the AI-powered content layer of Microsoft 365) includes document understanding models that read the structure and content of your documents — not just the filename or folder — and extract the data fields you define. When a contract arrives, it identifies it as a contract. It extracts the counterparty name, contract value, and expiry date. It writes those values back as searchable SharePoint columns without anyone touching a keyboard.
Capella pre-configures SharePoint Premium with classification models built for common enterprise document types — contracts, policies, invoices, forms, correspondence — and extends them for your specific taxonomy, industry, and regulatory requirements.
SharePoint Agents turn your document store into a resource people actually use.
Once your documents are classified and tagged, SharePoint Agents give your people a natural language interface to that content. An agent scoped to your contracts library can answer “what supplier agreements expire in the next 90 days?” without anyone running a report. An agent scoped to your policy library can surface the current version of a policy and explain what it covers.
Capella configures and trains SharePoint Agents against your classified document estate, defines the scope and permissions for each agent, and integrates them into the Microsoft Teams channels and SharePoint sites where your people already work. The agent is only as good as the underlying metadata — which is why classification comes first.
→ Acme Pty Ltd — 30 June 2026 — $240k
→ Brightline Services — 15 September 2026 — $85k
→ Northgate Consulting — 1 November 2026 — $320k
Shall I open any of these, or export the list?
From document chaos to a classified,
AI-ready document estate.
We’ve been in enough document remediation projects to know where they stall. Capella follows a sequenced approach that builds classification accuracy before deploying AI — so your agents answer correctly from day one.
We map your document landscape: volumes, formats, current taxonomy, and classification accuracy. We identify the five to ten document types that account for 80% of your risk and volume — these become the first classification models.
We build SharePoint Premium document understanding models against your actual documents. We define the metadata taxonomy, retention schedule mapping, and content type structure that classification will write to.
Models go live on incoming documents and priority existing libraries. We validate accuracy against a sample set, retrain where needed, and progressively expand library coverage. Purview retention labels are applied and governed.
With a classified document estate, SharePoint Agents are configured and deployed to target Teams channels and SharePoint sites. Ongoing managed service covers model accuracy, taxonomy governance, and new document type onboarding.
Six capabilities that close the gap
between SharePoint and a real DMS.
SharePoint Premium provides the engine. Capella provides the configuration, models, governance framework, and ongoing service that make it work for your specific document estate.
Documents classified automatically on upload — content type, retention category, and custom metadata columns populated without user action. Applies to new documents and can be run against existing libraries in batch.
Key fields extracted from document content and written to SharePoint columns — counterparty names, dates, values, reference numbers, policy owners. Makes document metadata machine-readable and reportable without manual entry.
Retention labels applied automatically based on document type and classification — aligned to your records schedule and Australian regulatory obligations. Integrated with Microsoft Purview for records management, holds, and defensible deletion.
SharePoint Agents configured to answer natural language queries against classified document libraries. Users ask questions; the agent surfaces the right document, version, and summary — without training or a search syntax to learn.
Classification models applied to existing document libraries in prioritised batches — not just incoming content. Migration debt addressed systematically, with accuracy validation before each library is considered complete.
Classification accuracy tracked continuously. Models retrained on edge cases and new document variants via fortnightly managed service sprints. Taxonomy changes governed through a structured change process — not left to drift.
Capella versus the alternatives.
A direct comparison.
There are three ways organisations typically tackle document classification. Here’s where each fits, and where each falls short.
| SharePoint out of the box |
DIY SharePoint Premium |
Capella (Antares) |
Third-party DMS (Objective, OpenText) |
|
|---|---|---|---|---|
| Auto-classification on upload | No | Possible — requires model build | Yes — pre-built and configured | Yes |
| Structured data extraction | No | Possible — requires field mapping | Yes — mapped to your taxonomy | Yes |
| Retention label automation | Manual only | Possible — Purview config required | Yes — integrated with Purview | Varies by vendor |
| AI document discovery (natural language) | No | SharePoint Agents available — config required | Yes — configured and deployed | Limited — depends on vendor roadmap |
| Backlog remediation (existing libraries) | Manual project required | Possible — requires planning | Yes — structured batch approach | Varies — migration complexity |
| Stays inside Microsoft 365 tenancy | Yes | Yes | Yes — no third-party data flows | No — separate platform |
| Ongoing accuracy monitoring | Not applicable | Typically not included | Yes — managed service included | Varies — support contract |
| Additional platform licence required | No | SharePoint Premium licence | SharePoint Premium licence (Microsoft) | Yes — full DMS platform licence |
Capella is the right fit when you’re already invested in Microsoft 365 and want classification accuracy without a parallel DMS platform. SharePoint Premium licence costs are a Microsoft commercial matter — Antares will advise on the right licence tier for your document volumes during the discovery engagement.
High-consequence document environments
need more than good intentions.
In most organisations, a misfiled document is an inconvenience. In regulated industries, it’s a liability. We’ve deployed document classification solutions across the sectors where classification accuracy isn’t a nice-to-have.
“We needed to know that our policy documents, correspondence, and claims records were classified correctly before we could rely on AI to surface them.”
Classification accuracy for financial services document estates typically requires handling 15 to 30 distinct document types across multiple business units. Capella builds models sequentially — highest volume and highest risk first — rather than attempting full coverage from day one. NRMA Insurance is among the organisations that have trusted Antares with their Microsoft 365 environment.
“Student records, research outputs, HR documentation, and board papers all live in SharePoint. They have completely different retention obligations.”
Education institutions carry some of the most varied document obligations of any sector — student records, enrolment agreements, staff contracts, research data, board minutes, and accreditation documentation all with different retention and access requirements. Antares works across tertiary and K–12 education, including UNSW Sydney, CQUniversity, Haileybury College, and Catholic Schools Broken Bay.
“Our client case files, funding agreements, and service contracts span 20 years. We needed a way to make sense of them without a full-time records manager.”
NFP organisations typically carry classification debt from growth-phase rapid document creation and limited records management investment. Capella’s managed service model means you get ongoing classification governance without needing to hire a dedicated records manager internally. Mission Australia is among the NFP organisations that have worked with Antares on their Microsoft 365 platform.
“Property documents accumulate across the asset lifecycle — due diligence, development approvals, maintenance records, leasing agreements. Finding the right document at the right time matters.”
Asset-intensive organisations accumulate documents across multi-decade asset lifecycles with complex chains of custody. Classification by asset type, location, and lifecycle phase makes AI-powered document discovery genuinely useful at the operational level. AVEO Group is among the property organisations working with Antares on their Microsoft 365 environment.
What organisations ask us
before they start.
SharePoint Premium (formerly Microsoft Syntex) is Microsoft’s AI-powered content layer for SharePoint and Microsoft 365. It includes document understanding models that read the content and structure of your documents — not just the filename — and extract key data fields automatically.
When a document is uploaded to SharePoint, SharePoint Premium identifies its document type, extracts metadata values (such as contract party names, invoice numbers, or policy dates), and writes those values back as searchable, governable SharePoint columns. Capella uses SharePoint Premium as its classification and extraction engine, pre-configured with models for common enterprise document types and extended for your specific taxonomy.
SharePoint Agents are AI assistants scoped to specific SharePoint sites or document libraries. They answer questions about the documents in those libraries, surface related files, summarise document content, and guide users through approval or review workflows — all from within SharePoint or Microsoft Teams.
In a Capella deployment, SharePoint Agents act as the conversational interface to your classified document store. Staff can ask “find our latest supplier agreements for Victoria” or “what are the retention obligations for this document type?” and get accurate, governed answers based on your actual content — not a generic AI response drawn from the internet.
SharePoint Premium provides the AI capability. Capella provides the deployment. That includes pre-built document classification models for common enterprise and industry document types, a structured taxonomy and metadata framework, configured retention policies aligned to Australian regulatory requirements, SharePoint Agent setup and training, integration with Microsoft Purview for records management, and an ongoing managed service covering model accuracy monitoring and taxonomy governance.
Most organisations can activate SharePoint Premium in a day. Getting classification accuracy above 90% across their real document estate — with the models trained on their actual document variations — takes a different kind of work. That’s the gap Capella fills.
No. Capella is built entirely on Microsoft’s native platform — SharePoint Premium, SharePoint Agents, and Microsoft Purview all run inside your Microsoft 365 tenancy. No documents, metadata, or AI outputs are sent to third-party platforms.
Your existing Microsoft Entra ID access controls, data residency settings, and compliance policies all apply without modification. There’s no new vendor with access to your content — classification processing happens within the same Microsoft AI infrastructure that underpins the rest of your Microsoft 365 environment.
Yes. SharePoint Premium can process existing document libraries as well as incoming files. For organisations with large backlogs of unclassified content, Capella includes a remediation phase that applies classification models to existing libraries, identifies coverage gaps, and progressively builds metadata across your document estate.
Large libraries are processed in batches with accuracy validation at each stage before proceeding. We prioritise by volume, risk, and regulatory obligation — so the documents that matter most are classified first, and you don’t have to wait for full coverage before the business starts seeing value.
Capella works best in organisations that have significant volumes of recurring document types — contracts, policies, forms, reports, invoices, correspondence — and where manual classification is creating compliance risk, search failure, or operational overhead.
It is particularly well suited to regulated sectors: financial services, health, education, legal services, local government, and not-for-profit. The higher the consequence of a misfiled or untagged document, the stronger the case for automated classification. If your organisation has fewer than a few thousand documents and one or two document types, a simpler approach may be more appropriate — and we’ll tell you that in the discovery engagement before any work begins.
A classified document estate starts with knowing what you have.
Our Document Discovery Workshop is a two-week fixed-fee engagement. We map your document landscape, identify the highest-risk unclassified content, assess your SharePoint Premium readiness, and produce a prioritised Capella deployment roadmap — with clear scope and cost for each phase.
You leave with a concrete plan. You decide whether to proceed, and on what timeline.
No obligation. No sales pressure. An honest conversation about whether automated document classification is the right fit for your organisation right now.