Skip to main content

Land small, prove fast, expand - Explore the Squirro AI Agent Catalog – Download Now

Book a Demo

Information Management · Banking · Government · Healthcare · Legal

Your Content Grows 50% Every Year. Your Classification Team Doesn't.

Untagged documents pile up. Manual teams fall behind. Self-tagging gets abandoned. Squirro classifies every document accurately and automatically at any scale, from the moment it enters your organization.

Save Money Deploy in 4–8 weeks Medium Effort
Book a Demo Explore the Full Agent Catalog

Trusted by Industry Leaders

Zwickroell Henkel Siemens OCBC Bank Candriam Bertelsmann IDB Invest Security Benefit

Measurable Impact

Colleagues working through requests together
57% average annual content growth rate
Over 95% classification accuracy, automatically, at ingestion

Evidence in Action

Watch Squirro Classify Untagged Documents from Ingestion to Audit Trail.

Document Classification & Taxonomy Agent - live demo, ready to play
Prefer to see it on your own data? Book a Demo

How it works

Fifty Percent YoY Content Growth. Same Taxonomy Team.

Manual tagging teams fall behind by design: content arrives faster than people can process it. Squirro classifies at ingestion against one enterprise taxonomy, writes the result back to every document, and sends the ones it is unsure about to a person instead of guessing.

01

INGESTION

Documents ingested from any source

Documents from SharePoint, OpenText, network drives and email archives land in one index, each with its source system kept alongside it.

Document Browser The live index — every document with the classification the pipeline wrote onto it
Search the live index by title, reference or topic… All document types 0 of 15 shown
SYN-0001 Supplier Agreement — IT Infrastructure ServicesSharePoint · Technology ContractConfidential94%
SYN-0004 Invoice — Cloud Services Q2SAP S/4HANA · Finance InvoiceInternal91%
SYN-0008 Correspondence — Financial Regulator Q3Exchange · Legal & Compliance Government CorrespondenceConfidential85%
SYN-0011 IT Security Incident Report — Case 2025-112ServiceNow · Technology Incident ReportInternal95%
SYN-0007 Phase III Trial Summary — Cardiovascular StudySharePoint · R&D · Pending Clinical SummaryRestricted72%
Anything under the 80% confidence threshold goes to a person before it is final
02

TAXONOMY

One taxonomy applied at ingestion

Every document gets a type, topic, department, sensitivity and retention class, with the model's confidence recorded next to each decision.

Document Browser The live index — every document with the classification the pipeline wrote onto it Classifying…
SYN-0001 Supplier Agreement — IT Infrastructure Services ContractConfidential94%
SYN-0001
Supplier Agreement — IT Infrastructure Services
Classification
Document typeContract
Model predictedContract · matches
TopicIT Procurement
DepartmentTechnology
SensitivityConfidential
Retention7 years
Source systemSharePoint
The model's raw prediction is kept in its own field rather than overwriting the curated label
03

QUALITY

Low-confidence calls go to a person

Anything under the confidence threshold lands in a review queue. A reviewer confirms or reclassifies it, and that decision is written back to the index.

Review Queue Below the 80% confidence threshold or manually flagged — actions write back to the index
3Pending review
12Auto-classified
0Reviewed
80%Threshold
SYN-0007Phase III Trial Summary — Cardiovascular StudyClinical Summary72% Confidence 72% is below the 80% threshold and “Clinical Summary” has low corpus coverage (n=1). Confirm or correct the classification; the confirmation becomes a labeled example for the next training run. ConfirmClinical SummaryReclassify
SYN-0013Annual Report Draft — FY2024Report78% Confidence 78% is below the 80% threshold and “Report” has low corpus coverage (n=1). Confirm or correct the classification; the confirmation becomes a labeled example for the next training run. ConfirmReportReclassify
SYN-0006Legal Opinion — GDPR Data TransferLegal Memo88% Manually flagged for review. Confidence is above threshold — confirm the classification to move it to reviewed status. ConfirmLegal MemoReclassify
Confirming writes review_status back onto the item and the counters recompute.
A person improving the model, not the model improving itself.
04

GOVERNANCE

Classifications written back to every item

Type, sensitivity and retention sit on the document itself, so search, retention and access rules all read the same labels.

Model & Ground Truth The model and ground truth behind every label, and what the metadata leaves behind on each item
Written back to every itemqueryable by anyone
FieldWritten byExample
doc_typeAI Studio model outputContract
topicClassifier / taxonomy conceptsIT Procurement
departmentLLM pipelet / custom fieldTechnology
sensitivityPrivacy layerConfidential
classification_confidenceClassifier score0.94
review_statusCustom fieldauto
taxonomy_concept[]Taxonomy concept URIsContract / Supplier Agreement
What this does not do yet
Classification-specific audit export is not built. Operational activity logging exists; a records-management audit trail does not.

Integrations

Connect Seamlessly with Your Existing Data

Break down data silos and integrate all of your existing data, creating a single source of truth.

Taxonomy-connectors

Your Expansion Path

Start Here. Scale Across Information Intelligence.

Classification is the foundation. Everything built on top – search, compliance, governance, migration – gets better as the taxonomy matures.

  1. 01 To start

    Document classification & taxonomy agent

    Automatic classification, Synaptica-powered taxonomy, self-learning accuracy, full audit trail. Live in 4–8 weeks.

  2. 02 Next

    Enterprise search unification agent

    Classification feeds directly into unified semantic search, making documents findable across every repository, by any dimension.

  3. 03 Next

    Automated content governance agent

    Retention categories applied automatically. Sensitive content identified and access-controlled. Regulatory requirements met without manual intervention.

  4. 04 Vision

    Full information architecture intelligence

    A connected content layer spanning classification, search, governance, and knowledge graph enrichment, where every document is organized, findable, and governed from the moment it arrives.

Our Platform

Connect Once, Govern Once, Reuse Everywhere.

Most enterprises rebuild the same plumbing for every AI project — connectors, governance, retrieval, audit. Squirro builds it once. You extend it.

Squirro Platform Copy-1

One Platform. Limitless Workflows. Ready to Deploy.

Ready to Deploy

Start Where the Friction Is.

Document taxonomy and classification, enterprise search unification, and automated content governance each create friction differently in every org. We work with organizations to map the most strategic path to value and to deploy quickly where the friction is sharpest.

Share your details with us to discover the use case, discuss how it fits your stack, or explore a deployment.

Browse the Catalog