How It Works

Save the Excel model. JSON In. JSON Out.

Business logic does not belong in application code.

CLE separates decision logic from application code and allows business experts to maintain decisions in Excel while applications consume the results through APIs.

CLE is a compiler and runtime for Excel-based business decision models.

Three Things. That’s It.

1. Business experts maintain the decision model in Excel

Pricing rules, validation rules, approval criteria, classifications, calculations.

The Excel workbook is the source code of the decision model.
The dependency graph is the executable artifact.

2. Your application sends JSON

POST /api/decision
Content-Type: application/json

[{
  "country": "Italy",
  "weight": 205,
  "currency": "EUR",
  "requestedQuantity": 300,
  "purchasePrice": 1.35
}]

Your ERP, CRM, PIM or application submits an existing business object.

3. CLE returns JSON

HTTP 200 OK
(annotated for clarity)

[{
  "country": "ITALY",       // standardized
  "iso_code": "IT",         // added
  "weight": 205,
  "weightClass": "heavy",   // added
  "currency": "EUR",
  "requestedQuantity": 300,
  "purchasePrice": 1.35,
  "purchaseTier": 3,        // calculated
  "marginFactor": 1.81481,  // calculated
  "salesPrice": 2.45        // calculated
}]

By default, CLE follows a JSON-In / JSON-Out enrichment model and returns the original business object with additional decision results.
For advanced scenarios, an optional custom output mode allows CLE to generate entirely different response structures.

CLE is designed for bulk decision execution.

All business objects within a request are evaluated by the same decision model. A request may contain one object or thousands of objects of the same type. This allows complete product catalogs, order lines, validation tasks or user interface evaluations to be processed efficiently in a single API call.

What Happens Inside CLE?

From the outside, CLE appears simple. Internally, however, CLE does not execute spreadsheets.

Instead, CLE analyzes the workbook and compiles it into an executable dependency graph.

flowchart TD

    A["Excel Workbook<br/>Source Code"]
    B["Dependency Analysis"]
    C["Compile"]
    D["Dependency Graph<br/>Executable Artifact"]
    E["Decision Runtime"]
    F["REST APIs"]

    A --> B
    B --> C
    C --> D
    D --> E
    E --> F

    classDef source fill:#E3F2FD,stroke:#1976D2,color:#000;
    classDef runtime fill:#FFDFE5,stroke:#FF5978,color:#000;
    classDef artifact fill:#FFF9C4,stroke:#D4AF37,color:#000;

    class A source;
    class D artifact;
    class E,F runtime;

Why a Dependency Graph?

Real-world business decisions are rarely just a single decision table.

A pricing model may depend on multiple formulas, lookup tables, intermediate calculations and business rules. CLE identifies these relationships and transforms them into an executable dependency graph.

A single pricing decision may combine formulas, lookup tables, reference data, classifications and multiple intermediate calculations.

CLE identifies these relationships and transforms them into an executable dependency graph.

flowchart TD

    I1["Input"]
    I2["Lookup A"]
    I3["Lookup B"]

    C1["Calculation 1"]
    C2["Calculation 2"]

    D1["Decision"]

    O1["Output A"]
    O2["Output B"]

    I1 --> C1
    I2 --> C1

    I1 --> C2
    I3 --> C2

    C1 --> D1
    C2 --> D1

    D1 --> O1
    D1 --> O2

CLE is Stateless By Design

CLE is a stateless decision runtime. It stores no business data, requires no database and maintains no user sessions. Every request is evaluated independently based solely on the provided JSON object.

Benefits:

  • horizontal scalability
  • deterministic execution
  • cloud-native deployment
  • container deployment
  • simplified operations
  • predictable performance
Scroll to Top