Documentación

Customer Scoring

Summary

Dedomena AI's Customer Scoring Neuron predicts customer value from transactional deposit and withdrawal activity. It converts recent and historical banking behavior into a score from 1 to 5 and a clear segment: top, high, medium, or low.

The system is designed for customer evaluation, segmentation, and behavior prediction using only transactional banking data. It analyzes balances, transaction amounts, deposit and withdrawal patterns, currencies, user category, account age, and transaction recency to estimate customer potential. The result helps teams prioritize engagement, personalize offers, and improve targeting with a consistent scoring layer.

???+ warning This score is an analytical prediction and should be considered alongside other business checks before making decisions. The algorithm runs on the fly: Dedomena AI does not store submitted data. You must copy the results to avoid losing the information.

Customer Scoring provides a practical way to rank customers, compare behavioral profiles, and support data-driven growth, retention, and engagement workflows.

Key Properties

The system's functionality is built on several key innovations:

  • Simplified Input Requirements: Unlike other models, this system only requires transactional banking data from the users being assessed. The API works with standard transaction fields such as amounts, balances, currency, user creation date, user category, and transaction date, reducing integration complexity.

  • Privacy-Oriented Processing: Submitted data is processed on the fly through the API. Dedomena AI does not store the transactional payload, helping organizations use the scoring service while limiting data exposure.

  • Advanced Behavioral Modeling: The data is transformed into behavioral indicators that capture deposit and withdrawal activity, transaction recency, account age, balance evolution, transaction volume, and activity across different time windows. These signals allow the model to identify meaningful customer patterns from raw transactional activity.

  • Customer Value and Behavior Models: The final scoring is determined by proprietary models that estimate customer value and behavioral consistency. The output provides a score from 1 to 5 and a clear segment: low, medium, high, or top.

  • Scalability and Updates: The system ensures scalability and seamless integration through an API. It supports scoring one or multiple users in a single request and applies usage controls based on the subscription plan associated with the token.

  • Automation and Operational Simplicity: The scoring process is automated from request validation to feature generation and prediction. The response is compact and easy to integrate into segmentation, targeting, CRM, and engagement workflows.

Use Cases

Here are the key use cases for the Customer Scoring Neuron:

  • Customer Segmentation: Identifying customer segments and tailoring products and services to meet their specific needs.

  • Marketing Strategies: Optimizing campaign audiences by prioritizing customers with stronger transactional activity, higher value signals, or greater engagement potential.

  • Customer Relationship Management: Helping CRM teams personalize outreach, offers, and service levels based on each customer's score and behavioral profile.

  • Retention and Reactivation: Identifying customers whose recent activity, transaction recency, or behavioral patterns suggest a need for retention actions or reactivation campaigns.

  • Portfolio Prioritization: Ranking customers across a portfolio so business teams can focus attention, incentives, and resources on the most relevant segments.

  • Behavioral Monitoring: Detecting changes in deposit and withdrawal activity over time and using those signals to prioritize follow-up actions.

  • Product Personalization: Supporting recommendation, bundling, and cross-sell strategies by matching products or benefits to customer value and activity patterns.

  • Business Growth: Driving strategic growth by using customer scores to improve targeting, increase satisfaction, and make engagement workflows more efficient.

Dedomena AI’s Customer Scoring Neuron is the ultimate solution for enterprises, retailers, fintechs, and organizations across sectors, seeking a comprehensive, secure, and innovative tool for customer evaluation and segmentation.

API & Endpoints

:material-application-brackets-outline: /v4/scoring

POST/v4/scoring

This endpoint predicts a customer score from 1 to 5 and classifies users into 4 categories: top, high, medium, and low.

Headers:

  • token: available via free trial or subscription.

Parameters:

  • user_type (str): Type of user. Right now the only possible value is PERSON.

Request Body:

It should be a JSON array with the following fields:

  • user_id (str): Unique anonymous id that identifies the user.

  • transaction_state (str): State of the transaction. Optional. If omitted, v4 infers success from the transaction amount sign.

  • transaction_currency (str): Currency of the transaction.

  • user_balance (float): User balance at the time of the transaction.

  • transaction_amount (float): Amount of the transaction. Positive values are deposits and negative values are withdrawals.

  • user_created (str): Date and time when the user was created.

  • user_category (str): User category. Allowed values are API and WEB.

  • transaction_date (str): Date and time of the transaction.

image

Request Body - JSON Array
json
[
    {
        "user_id": "user_123",
        "transaction_state": "SUCCEEDED",
        "transaction_currency": "USD",
        "user_balance": 1000.0,
        "transaction_amount": 50.0,
        "user_created": "2024-01-01T12:00:00",
        "user_category": "WEB",
        "transaction_date": "2024-06-30T12:00:00"
    },
    {
        "user_id": "user_123",
        "transaction_state": "COMPLETED",
        "transaction_currency": "USD",
        "user_balance": 900.0,
        "transaction_amount": -100.0,
        "user_created": "2024-01-01T12:00:00",
        "user_category": "WEB",
        "transaction_date": "2024-06-30T12:30:00"
    }
]
Response

200

json
[
    {
        "user_id": "user_123",
        "score": 4.35,
        "category": "top"
    }
]

400

json
[
    {
        "detail": "Data not valid"
    }
]

401

json
[
    {
        "detail": "Could not validate credentials"
    }
]

405

json
[
    {
        "detail": "User type not allowed"
    }
]

406

json
[
    {
        "detail": "the number of observations in the payload exceed the available quote"
    }
]

422

json
[
    {
        "detail": "Unprocessable Entity"
    }
]

429

json
[
    {
        "detail": "The maximum monthly number of calls or predictions for your account has been exceeded"
    }
]

Metrics

Dedomena AI model's performance metrics ensure high accuracy and reliability:

Summary:

Customer Scoring | Dedomena AI Documentation | Dedomena AI