What challenges did DTI Group face before AI implementation?

Manual extraction, reconciliation, validation, and re-entry of information from unstructured documents slowed onboarding, claims handling, invoicing, and other workflows. Traditional rule-based automation could not reliably interpret unstructured inputs or connect different information types across complex, multilingual processes.

What was the impact of the AI solution?

The platform analyzes documents in seconds with accuracy approaching 99%. In one loan-processing scenario, processing time was reduced by 80% per case. In a high-volume banking deployment, pages were processed in an average of 2.5 seconds, below the required service-level thresholds of 6 to 9 seconds. It helped one retail bank customer process more than 50,000 loan agreements annually.

Source

Microsoft

DTI Group uses Azure AI Document Intelligence and Microsoft Foundry to deliver agentic document processing (COGNAiO)

Use case typeDocument processing automationUpdated 6 days ago

DTI Group, based in Switzerland, built COGNAiO on Azure and Microsoft Foundry to analyze documents via meaning and context, validate information, and connect outputs to downstream systems. The platform supports document-heavy workflows across regulated sectors and is designed to automate extraction and validation across documents, emails, images, and data feeds. It uses Azure Document Intelligence, Azure Kubernetes Service, and Azure Database for PostgreSQL, and allows users to describe processes in natural language rather than code.

Organization: DTI Group
Industry: Professional Services
Location: Switzerland
Published: July 2026

Reported outcomes

+99%

document accuracyQuality & accuracy

−80%processing time per loan case2.5 secondspage processing time6 secondsservice level threshold9 secondsservice level threshold50,000 agreementsloan agreements processed annually

Strategic outcomes

Speed & agilityAutomated end-to-end document workflowsRisk & complianceBuilt-in governance for regulated environmentsMarket & geographic expansionExpanded into additional marketsNew business modelShifted from solution specialist to platform company

Catalog median for quality & accuracy deployments: +90% across 282 reported metrics. Compare benchmarks →

Primary read

Use case focus

Showing 2 of 2

1Document processing automation
2AI agents

Manual extraction, reconciliation, validation, and re-entry of information from unstructured documents slowed onboarding, claims handling, invoicing, and other workflows.
Traditional rule-based automation could not reliably interpret unstructured inputs or connect different information types across complex, multilingual processes.

DTI Group built COGNAiO, an AI-native platform for agent-based document intelligence at scale.
Its agents interpret, validate, and act on information in documents, emails, images, and data feeds, connecting inputs to downstream actions.
The platform combines Azure Document Intelligence with Microsoft Foundry, Azure Kubernetes Service, and Azure Database for PostgreSQL.
It supports natural-language process specification and includes logging, explanations, confidence thresholds, and human oversight for regulated environments.

Technologies

Azure AI Document Intelligence Microsoft Foundry Azure Kubernetes Service Azure Database for PostgreSQL

The platform analyzes documents in seconds with accuracy approaching 99%.
In one loan-processing scenario, processing time was reduced by 80% per case.
In a high-volume banking deployment, pages were processed in an average of 2.5 seconds, below the required service-level thresholds of 6 to 9 seconds.
It helped one retail bank customer process more than 50,000 loan agreements annually.

Architecture

COGNAiO runs on Azure with Microsoft Foundry for agentic document intelligence, Azure Document Intelligence for document understanding, Azure Kubernetes Service for scalable application hosting, and Azure Database for PostgreSQL for persistence. The platform uses agents to interpret, validate, and act on information across documents, emails, images, and data feeds, with built-in governance, logging, explanations, and confidence thresholds for human oversight.

Sources & evidence1

Groundedness: 5/5Type: Customer StoryPublished: Jul 7, 2026Publisher: MicrosoftEvidence: PrimaryConfidence: High

Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Customer evidence