Banking Challenges on Cash Management

Logistics of cash replenishment is a complex and costly process

  • 1

    Ensuring adequate cash levels is a challenge, requiring accurate forecasting of cash demand

  • 2

    Managing the cost of cash replenishment is important

  • 3

    Banks need to manage the risks associated with cash replenishment, such as theft or loss during transportation

  • 4

    Banks need to comply with various regulatory requirements related to cash replenishment,

  • 5

    Idle cash in ATMs represents an inefficient use of resources, as the cash is not being used to meet customer demand.

  • 6

    If there is too much idle cash in the ATM, it can be costly for the bank, as they have to pay for transportation, security, and storage

  • 7

    Rapid changes on consumers behaviour

  • 8

    Vandalism of ATMs

Custom Made solutions

Problem

85% of big data projects fail (Gartner, 2017)

87% of data science projects never make it to production (VentureBeat, 2019)

“Through 2022, only 20% of analytic insights will deliver business outcomes” (Gartner, 2019)

Traditional approach

:

ExepnoCash for ATM Cash Replenishment
Optimisation

Some of the most important reasons that lead to project failure include:

Data Management

Data Integration

Technology Complexity

Insufficient resources

Challenging to create an ML workflow

Difficult to ensure continuous training and deployment

From raw data to added value

1

Ingest Data

✓ Data Warehouse
✓ Hadoop HDFS
✓ External Sources

2

Data Preparation

✓ Outlier Detection
✓ Handling Missing Values
✓ Normalisation
✓ Feature Engineering
✓ Feature Selection

3

Train Model

✓ Random Forests
✓ Neural Networks
✓ Boosted Trees
✓ Econometrics

4

Validate Model

✓ Error metrics
✓ Key performance indicators
✓ Model comparisons

5

Deploy

✓ Reporting
✓ Productionalising
✓ Automating
✓ Creating pipelines
✓ and APIs

How a machine Learns

1

Uses large volumes of data

✓ Structured ✓ Unstructured
2

Utilises advanced algorithmic models

✓ Algebra ✓ Calculous
3

Learns without being explicitly programmed

✓ No if-then-else functions
4

Discovers hidden patterns

✓ Uncovers the underlying rules

End-goal: Collective Intelligence

  • Human experts can provide context and insights: Human experts, such as bank managers and financial analysts, can provide valuable context and insights into the underlying factors that drive cash demand.
  • Machine learning models can process large amounts of data: Machine learning (ML) models can process large amounts of data and can identify patterns and trends that may not be visible to human experts.
  • Combining human expertise and machine learning: By combining human expertise and machine learning, banks can develop more accurate predictions of cash demand. For example, human experts can provide insights into the underlying factors that drive cash demand, while machine learning models can process large amounts of data to identify patterns and trends.

Options

Data-Driven Models considered

Standard Econometric Models

AR
ARMA
ARIMA
SARIMA

Machine Learning Models

Support Vector Machines
Random Forests
Gradient Boosted Trees

Neural Network Models

Feed Forward Neural Networks
Entity Embeddings
Wide & Deep
Recurrent LSTMs

Why Vertex AI?

1

Build, Deploy and scale ML models faster.

2

We have created a set of training models that can fit the data set of any bank

3

Vertex AI is evaluating and comparing models into production with the new ones to choose the most accurate

4

Reduce the complexity and management of the large-scale deployment of ML models

Google Vertex AI

Vertex AI Advantages

Top use cases of Vertex AI

Data loading and preprocessing

VertexAI provides an easy-to-use interface for loading and preprocessing data, which can be used to load transaction data and other relevant data such as weather information and economic indicators.

Read More

Model training

VertexAI includes a number of pre-built ML models that can be used to train models on the data. It also allows users to create their own custom models using the drag-and-drop interface.

Read More

Model deployment

VertexAI allows for easy deployment of trained models. This means that once a model is trained, it can be deployed in a production environment and used to make predictions in real-time

Read More

Model monitoring

VertexAI provides monitoring tools that allow users to track the performance of the deployed models over time. This helps to ensure that the models continue to make accurate predictions

Read More

OUR IMPLEMENTATION - Components

Data collection

The first step is to collect the data needed to train the model. This includes data from previous transactions, as well as any additional data that may be relevant, such as weather information and economic indicators

Data preparation

Once the data is collected, it needs to be prepared for training. This includes cleaning the data, removing any inconsistencies or outliers, and normalizing the data

Model selection

Next, a machine learning model needs to be selected. It could be a pre-built model or a custom model created by the user. The model should be appropriate for the task at hand and capable of handling the data set

Model training

Once the model is selected, it can be trained on the prepared data. The goal of this step is to train the model to make accurate predictions about cash demand

Model evaluation

After the model is trained, it should be evaluated to ensure that it is making accurate predictions. This can be done by comparing the predictions made by the model to the actual cash demand

Model deployment

Once the model is trained and evaluated, it can be deployed in a production environment. This can include publishing the model to a variety of different platforms, such as web, mobile, and IoT devices, to make it available for use in real-time

ExepnoCash

ExepnoCash offers a highly dependable cash management solution by eliminating manual calculations in the ATM replenishment process. It minimizes unnecessary cash reserves, precisely forecasts cash shortages, and lowers emergency cash deliveries. It harnesses the capabilities of Machine Learning and it shortens the duration required to efficiently move from data ingestion to deploying a new ML model in production, based on updated information. ExepnoCahs can be relied on to accurately predict cash demand and manage cash reserves, which would lead to fewer emergency cash deliveries and more efficient use of resources. Additionally, it means that the solution can be trusted to make accurate predictions, which would help banks to better manage their cash resources and provide a better customer service.
  • ExepnoCash automates, monitors, and optimizes cash management for banks
  • Provides valuable insights for managers and stakeholders
  • Allows inclusion of known events and bank holidays in advance to predict cash demand fluctuations
  • Its AI-brain constantly compares predictions with actual results to improve prediction accuracy
  • Based on the vast computational power of Google's Cloud Platform, it has an advantage over traditional econometric models.
product pdf

The Key Point of Vertex AI components:

  • Data collection and preprocessing: ExepnoCash has the capability to collect and preprocess data from various sources such as ATM transactions, historical data and external data sources like weather and economic indicators.
  • Machine learning models: ExepnoCash has pre-built or customizable ML models that can be trained on the collected data. These models are responsible for making predictions about cash demand.
  • Event and holiday inclusion: ExepnoCash would need to have a feature that allows users to input known events and holidays in advance, so that the system can take into account the expected fluctuations in cash demand.

Advantages

  • Model evaluation and improvement: ExepnoCash has a system for evaluating the predictions made by the model, comparing them to actual results, and using this feedback to improve the model's accuracy over time
  • Cloud-based infrastructure: ExepnoCash is based on the powerful cloud-based infrastructure of Gogle Cloud, to allow for the fast and efficient processing of large amounts of data.

Components

Monitor model performance

Evaluate the performance of the current model and improve as needed

Route Optimisation

Determine ICT track routing to curtail costs further

Extend the use case ATMs

Scale to the entire ATM network / Include the entire branch cash demand

Solve the Cash Replenishment challenge

Accurately predict the cash demand per ATM

Our solution:

ExepnoCash

Roadmap

Artificial Intelligence Banking Applications

Credit Risk

Money Laundering Detection

Fraud Detection

Clustering / Customer Segmentation

Churn

Cross Selling

Recommendation Systems

Natural Language Processing

Channel Optimisation

Chat Bots

Image Recognition

ATM Cash Replenishment Optimisation

contact Us