BranchCheck
The principle

One engine, seven domains

BranchCheck is not a suite of bolted-together modules: a single engine runs every business domain of a branch network — which is how five of them shipped end to end, and two partly, without writing seven applications.

One engine, not seven applications

"A single engine of referentials, campaigns, evaluations, proofs, actions, validations and reporting must serve every domain. Domains must not be built as independent applications."

Requirements specification, §1. Read domain by domain, the need looks like seven products; what it asks for is an engine. A domain that needs code beyond its reporting screens is the sign the engine is not general enough — so every domain here is a set of referentials, form templates, workflow rules and entitlements.

The domains — configurations

  • Compliance · Shipped
  • Permanent control · Shipped
  • Quality · Shipped
  • Audit · Shipped
  • Network animation · Shipped
  • Commercial performance · Partly covered
  • Training & competencies · Partly covered

One common engine

  • Versioned referentials
  • Campaigns
  • Four-state evaluations
  • Offline proofs
  • Configurable workflow
  • 4D entitlements
  • KPI dictionary

The engine's seven capabilities

Each one generalizes what already runs in production — not a layer still to come.

Versioned referentials

Requirements, criticality, expected proof, effective date: a domain's referential is versioned and published, never hard-coded.

Campaigns

Goal, scope, population, success criteria: a domain's scheduling is a campaign, not a development.

Four-state evaluations

Compliant, non-compliant, not applicable, missing data: four distinct states — an aggregate that confuses them is a blocking defect.

Configurable workflow

Draft, review, validation, closure: transitions and their guards — role, justification, proof, signature — are data.

4D entitlements

Role × scope × domain × confidentiality: activating a domain is an entitlement, not a code variant.

KPI dictionary

Name, formula, source, thresholds: indicators share one definition, and every score carries its calculation trace.

The foundation

Offline proofs

Collected with no signal, sealed on submission, verifiable forever: every result stands up in front of a regulator.

What the stitching makes possible

Three chains shipped in v1.5 that cross domains with no integration to write.

A repeated failure triggers microlearning

After repeated inspection failures or a new procedure, the right microlearning module is assigned automatically — read and acknowledged on mobile, even offline.

A complaint goes back to its cause

The complaint is qualified — severity, cause, process, impact, computed recurrence — then root-cause analysis and an improvement plan, submitted for approval, close the loop; effectiveness is measured as recurrence before and after.

A commercial gap becomes a coaching visit

Comparing results with execution across four quadrants surfaces the gap; a corrective action or a coaching visit is triggered and tracked like any other action.

Seven domains, one true state each

Five domains ship and run end to end on the engine: compliance, permanent control, quality, audit, network animation. Commercial performance and training are partly covered. Each domain page says, requirement by requirement, what is covered, partial or not covered.

In the free trial, the domain screens start empty: the shared demo pre-fills inspections only.

The shared glossary — one term of the engine or a domain, one definition, three languages