Docs/How Neuron works

Automatic mapping

Reduce manual schema work without hiding uncertainty.

What a mapping describes

A mapping connects a source representation to a business concept. For example, Salesforce Account and ERP BillToCustomer may both contribute information about Customer. The original fields remain unchanged.

Architecture example · What a mapping describes
  1. Salesforce.Account + ERP.BillToCustomer
  2. Proposed Customer concept
  3. Review ambiguous meaning
  4. Retain source mappings

Signals for proposed mappings

  • Table names, field names, types and sample values.
  • Foreign keys, API schemas and existing relationships.
  • Document structure, descriptions and metadata.
  • SOP terminology and the Workspace’s business purpose.

Automation with review

The intended model is for Neuron to propose and maintain mappings, with business review when meaning is ambiguous. PlantCode might represent a physical Plant or a Business Unit; the system should surface that ambiguity.

No numeric confidence score is shown in this release. A high-confidence label or automatic approval must not be inferred from a green canvas connection.

Available now

Users explicitly connect a source to a task’s context item. PatternLab does not currently inspect schemas to propose mappings, maintain them automatically or present a mapping-review queue.