Skip to main content

Live overview

Explore the Smartflow deployment model

Click a layer to see what it does, where the data boundary sits, and what users need to know.

Dedicated institution environment

Workspace

Smartflow app

The application handles upload, review, approval, and export workflows.

Boundary
Workflow activity remains inside the dedicated customer environment.
User takeaway
This is the main place users upload documents, review extracted fields, and approve work.

What this page covers

Architecture Overview explains how Smartflow is deployed inside your institution environment and how document data stays within the approved boundary.

Before you start

  • Your IT team has selected an approved deployment model.
  • Your security team has confirmed data residency requirements.
  • Your Smartflow onboarding team has confirmed the implementation scope.

Deployment model

Smartflow runs as a dedicated instance for your institution. It is not a shared document-processing environment.

The website version of this page includes a clickable architecture view. If you are reading the Markdown file directly, the deployment model has three simple layers:

  1. Users access the Smartflow application through the approved access path.
  2. Smartflow handles upload, extraction, review, approval, and export workflow.
  3. Approved services and integrations store documents, keep workflow records, run extraction, and send verified data to configured downstream systems.

Core components

  • Application - Lets users upload, review, approve, and export work.
  • Document storage - Stores uploaded documents inside the approved environment.
  • Application database - Stores workflow state, extracted values, review activity, and configuration.
  • AI extraction services - Extract fields and provide confidence scores and source evidence.
  • Review workflow - Routes work to users for human review and approval.
  • Approved integrations - Sends approved data to connected systems where configured.

Security model

Smartflow is designed around:

  1. Dedicated customer deployment.
  2. Bank-controlled data boundary.
  3. Role-based access control.
  4. Human review before high-impact data is accepted.
  5. Field-level source evidence.
  6. Activity and review traceability.

Data flow

  1. A user uploads a document.
  2. Smartflow stores the document inside the approved environment.
  3. Extraction creates structured fields.
  4. Reviewers confirm or correct the fields.
  5. Approved data becomes available for export or downstream use.

Tips

  • Use this page as the first architecture reference for IT and InfoSec reviews.
  • Pair this page with the deployment prerequisites before scheduling rollout.
  • Keep integration discussions separate from user training.

Troubleshooting

  • A security reviewer asks for network details -> Use Deployment Prerequisites.
  • A reviewer asks where document data is stored -> Confirm the approved deployment boundary with your IT owner.
  • A reviewer asks about downstream systems -> Confirm which integrations are in scope for the deployment.

Related features