Finacle

Data and AI Suite

Accelerate AI adoption at scale

Inspiring Better Banking

www.finacle.com
 AI in banking – A growing imperative
AI is no longer a novelty in the banking sector. Banks have been exploring AI technologies for
a while now, ranging from machine learning and natural language processing to computer
vision, deep learning, and now, generative AI. However, the pace of AI adoption has accelerated
significantly in recent years, with over 90%1 of banks implementing AI in some capacity.

Generative AI, in particular, has disrupted the status quo, forcing banks to re-evaluate their AI
strategies. While the full potential of generative AI in banking is still unfolding, it has undoubtedly
brought AI discussions from tech circles to the forefront of boardroom agendas.

The IMF predicts2 a twofold increase in AI spending by financial institutions between 2023 and
2027, highlighting the growing importance of AI in the banking industry. As banks intensify their
AI investments, they are strategically allocating resources to both generative AI pilots as well as,
speeding up traditional, predictive AI programs.

The world is abuzz with excitement about AI’s potential. Banks worldwide are exploring this
transformative technology, recognizing that their success hinges on a comprehensive AI strategy, a
commitment to future vision, and the adoption of modern architectural principles.

Source: (1) IMF report – AI’s reverberations across finance, 2023

(2) EY survey – AI adoption in Financial Services, 2023

2 | External Document © 2024 EdgeVerve Systems Limited

 While the potential is clear, lack of
AI strategy is a concern
Despite significant advancements in AI technology, many banks face including ineffective and inaccurate insights, poor decision-making, increased
challenges in fully realizing its potential. One of the primary obstacles is regulatory scrutiny and higher costs, all of which can eventually result in low
inadequate data readiness. Although they possess vast amounts of data, less ROI and a stalled AI progress.
than 0.5%1 is effectively utilized, hindering AI progress. Also, over 40%2 of banks
cite a lack of proper data infrastructure as a major deterrent.

Beyond data readiness, AI governance and stewardship are other crucial

areas - where banks often fall short. Unclear governance methods, model
explainability, and poorly established responsible AI frameworks impede
meaningful and sustainable AI adoption at scale.
The The
impediments consequences
Generative AI, a relatively newer technology, introduces additional
Suboptimal data Ineffective insights,
complexities. Several concern areas regarding LLMs, including model poor decisions |
strategy | Inadequate
accuracy, drift, and bias, are gradually coming to the fore and emerging as AI governance Black-box models
significant challenges. | Emergent | Regulator and
complexities in LLMs stakeholder scrutiny
| Increased costs
Banks have a lot of ground to cover, before they start unlocking the true
and complexity | AI
potential of AI. A poorly defined AI strategy can have severe consequences, program failures

Sources : (1) MDPI Journal - Big Data-Driven Banking Operations: Opportunities, Challenges, and Data Security Perspectives , July 2023
(2) EY Survey: AI adoption among financial services, Dec 2023

3 | External Document © 2024 EdgeVerve Systems Limited

 An ideal AI strategy is built on the
following foundational virtues
Data democratization to ensure data is accessible and readily available for AI consumption
by all relevant stakeholders, fostering collaboration and innovation.

AI/ML agnosticism to ensure the strategy is vendor and technology agnostic, allowing for
flexibility and adaptability to emerging trends and advancements.
Data AI ML Model
Model transparency and explainability to make AI models understandable and democratization agnosticism Transparency and
interpretable, promoting trust and accountability. explainability

Scalability to handle increasing workloads and data volumes while maintaining optimal
performance and efficiency.

Data privacy control through encryption and masking during model training, to protect Scalability and Data privacy Continuous
user data while ensuring optimal AI performance. performance controls monitoring
and adaptability
Continuous monitoring and adaptability through robust learning and feedback
mechanisms to continuously refine model performance, keeping drift and bias under
control.

4 | External Document © 2024 EdgeVerve Systems Limited

 Accelerate AI
Journey with
Finacle

5 | External Document © 2024 EdgeVerve Systems Limited

 Finacle Data and AI Suite
Finacle Data and AI Suite is a ‘data to decisions’ suite of platforms to help The suite offers
banks unlock the full potential of AI at scale. The suite offers a set of low-code
A robust data platform to seamlessly capture, integrate, and transform
no-code platforms that empower banks to accelerate their AI journeys by
data throughout the entire data lifecycle, ensuring data quality
scaling data readiness, industrializing AI model development and harnessing
and accessibility.
generative AI technologies. The suite is designed to seamlessly infuse AI into
banks’ digital operations and deliver actionable insights across their entire A no code AI platform built exclusively for banking, empowering
ecosystem. business users to build, train, deploy, monitor and optimize AI solutions.

Finacle Data and AI Suite is a part of Infosys Topaz Suite of AI offerings, which A host of Generative AI solutions tailor made for specific banking needs,
brings the advantage of 12,000+ AI assets, 150+ pre-trained AI models and a including enterprise-centric and market-centric use cases.
‘responsible by design’ approach that maintains strong standards of AI ethics,
trust, privacy, security and regulatory compliance.

1 2 3

Finacle Finacle Data Finacle Finacle

Data and Platform AI Platform Generative AI
AI Suite Data Lakehouse | No Code MLOps | Enterprise-centric |
Data Pipeline Embedded AI Market-Centric
Use cases

6 | External Document © 2024 EdgeVerve Systems Limited

 Finacle Data
Platform

7 | External Document © 2024 EdgeVerve Systems Limited

 Finacle Data Platform
Finacle Data Platform is a comprehensive solution designed to help banks
unlock the full potential of data and accelerate AI readiness. The platform is Finacle Data Platform comprises of two key components
built on a layered OLTP-OLAP architecture which is polyglot by design and
segregated by a CQRS pattern. The platform offers low-code configuration
capabilities, facilitating easy integration from various data sources and Finacle Data Pipeline Finacle Data Lakehouse
enabling rapid data cleansing and transformation.
Automates ingestion, cleaning, and A unified data repository that centralizes
transformation across all sources. and manages both structured and
Powered by a robust data pipeline, the platform captures, extracts, curates,
Built-in adapters simplify integration, unstructured data.
and delivers high-quality data for various consumption scenarios, including
while custom ETL scripts and metadata
analytics, reporting, and AI applications. The integrated yet modular data management offer flexibility. Built on BIAN-inspired models, it
lakehouse provides a unified data estate, enabling banks to access data from ensures data standardization and
diverse sources and establish a single source of truth. Advanced cleaning and enrichment interoperability. The lakehouse’s in-
enablers (Deduplication, consistency memory, domain-based data marts
Built in security features such as access control, encryption, authentication, checks, formats and anomaly detection) facilitate rapid data access and analysis.
synthetic data generation, dynamic masking and a zero-trust architecture ensure data integrity. Furthermore,
helps banks safeguard their data. Real-time observability controls with codeless data conversions and Moreover, its RESTful APIs enable
transformation tools prepare data seamless data sharing with external
dashboards, notifications, alerts and log streaming ensure transparency.
for diverse consumption scenarios, systems.
Moreover, robust data governance through classification, lineage tracking,
maximizing its potential.
metadata management and policy enforcement guarantees data
auditability and compliance.

8 | External Document © 2024 EdgeVerve Systems Limited

 Finacle
Data Platform Capture/Emit Stream Curate

Finacle Data Platform

Consume

Rest APIs

Finacle/Non-Finacle Applications

Finacle/Non-Finacle Applications
Functional Architecture
OLTP Finacle Data Pipeline
database Business dashboards
Real-time and reports
Ingest Clean Transform
event
Files Tailored data Deduplication | Pre-modeling checks | AI/ML models
streaming extraction | Metadata Consistency checks Model selection |

Connectors
management | | Formats Aggregation | Embedded analytics
APIs/ Pre-processing Normalization | Encoding
webhooks Real-time Enterprise analytics
event sourcing engine
Event Query Engine - SQL, JSON, Spark
Developer apps
streams Batch feeds
Finacle Data Lakehouse Downstream
system
Others Unify Store Access
Structured and BIAN data In memory domain Synthetic datasets
... unstructured data sets models specific data-marts
...

Data Observability
Dashboards | Notifications | Alerts | Log streaming

Data Governance
Classification | Quality frameworks | Metadata management | Lineage tracking | Policy enforcement | Audit | Compliance

Data Security and Privacy

Access control | Encryption | Authentication | Data retention | Dynamic masking | Zero trust

External Logos used are registered trademarks of the respective organizations

9 | External Document © 2024 EdgeVerve Systems Limited

 Finacle AI
Platform

10 | External Document © 2024 EdgeVerve Systems Limited

 Finacle AI Platform
Streamline AI development lifecycle with no-code capabilities
Finacle AI Platform is a comprehensive solution designed to help banks With its extensive library of pre-built AI use cases, Finacle AI Platform enables
accelerate AI journeys. The unified, no-code platform simplifies AI development banks to seamlessly integrate AI into various business processes and customer
lifecycle, empowering both technical and business users to rapidly build, train, journeys. This empowers banks to accelerate value creation from their AI
deploy, monitor and optimize explainable AI solutions with ease. investments and drive innovation.

The platform offers a rich library of pre-built ML models and supports varied
techniques like classification, regression, clustering, and forecasting. Banks can
leverage these models to build custom ensembles or even import their own
pre-trained models. Features such as model comparators, what-if simulations
Ensemble of Combinatorial Model Model
and pattern analysis are available to refine and elevate modeling sophistication.
pre-built- approach monitoring and explainability
The platform promotes responsible AI practices - comprehensive capabilities models governance
for detecting biases and drifts, along with a patent-pending synthetic data
generation feature to safeguard data privacy during model build and training are
all available natively.

Finacle AI Platform is built on the principles of transparency, interpretability, and

explainability, fostering trust and accountability in AI models. The platform is Single click Deeper Standardization Privacy
based on open-source technologies and can be deployed on both cloud and on- integrations insights options design
premises infrastructure.

11 | External Document © 2024 EdgeVerve Systems Limited

 Finacle
AI Platform Embedded Customer Data Hub | Payments | Wealth Management | Other Finacle and non-Finacle products
AI

Functional Architecture
AI
governance
Integration End Point Integration
through API/ Data drifts
Batch Connected devices | Webhooks | APIs | Reports | Dashboards | ....
Model drifts

Continuous Decision & Prediction AI Engine Model

AI / Loop ML Ops - Training, Deployment, Monitoring Explainability | Transparency Building
Engine

Pre-Processing
Flexi Data
Marts Cashflow forecasting | Wealth management - Hyper personalized recommendations |
Deposit churn | Corporate credit risk rating basis MD&A assessments | Smart deposits

Synthetic data generation

Finacle
Finacle Data Pipeline | Finacle Data Lakehouse | BIAN-inspired data models |
Data
In memory, domain-based data marts
Platform
OLTP database | Files | APIs/webhooks | Event streams | Others
(Finacle/Non Finacle systems)

12 | External Document © 2024 EdgeVerve Systems Limited

 Gen AI powered
combinatorial
approach
Simplifying model
selection and build

13 | External Document © 2024 EdgeVerve Systems Limited

 An array of embedded AI use-cases,
available out of the box

Retail Banking | Corporate Banking | Wealth Management | …

NPA prediction for Cash flow Forecasting Corporate credit risk Wealth management
Customer deposit
early intervention and Smart deposits for corporates rating basis MD&A - Hyper-personalized
churn prediction
default mitigation assessments recommendations

and more…

*Indicative list, not exhaustive

14 | External Document © 2024 EdgeVerve Systems Limited

 Finacle
Generative AI

15 | External Document © 2024 EdgeVerve Systems Limited

 Finacle’s strides with Generative AI
Proving value to create value for the industry
Finacle has been making rapid strides with Generative AI. The technology
Finacle Knowledge AI Assistant simplifies navigation and search across
plays a foundational role in our AI platform, empowering users with an

1
multiple document repositories. By understanding diverse user queries
interactive interface to select and build the right models for their AI use
and leveraging RAG, the AI assistant provides contextual summaries and
cases. Our Generative AI offerings encompass a diverse range of AI assistants
relevant information.
designed for both enterprise and customer-centric applications.

Finacle Support AI Assistant streamlines ticket resolution for support

Finacle’s Generative AI assistants leverage advanced techniques such as
teams. Trained on a vast repository of tickets and resolutions, the bot
prompt engineering, model fine-tuning, and Retrieval Augmented
Generation (RAG). This enables them to effectively process and generate 2 can combine solutions from multiple incidents to address new issues
effectively. It can also access a universal set of incident-response
context-aware, human like responses. These AI assistants are built on the
repositories for enhanced resolution mechanisms.
foundations of a variety of transformer-centric large language models,
including all the major open-source and enterprise editions. However, the AI
Finacle’s market-centric Generative AI offerings include collaborative
assistant solutions are agnostic to the choice of LLMs, and hence offer flexible
partnerships with banks to co-create several use cases addressing key focus
options.
areas such as customer engagements, loans processing, reporting and
compliance, data analysis, and more.
Finacle’s enterprise-centric Generative AI assistants are designed to
revolutionize operations and step up efficiencies. An array of AI assistants
tailored for varied requirements of banks are available. Here are two
prominent examples –

16 | External Document © 2024 EdgeVerve Systems Limited

 The AI revolution in banking is here.
To maximize AI opportunities and
respond with agility, banks and financial
institutions must adopt a holistic approach
to their data and AI strategy. This involves
reimagining their technology platforms
with AI-infused capabilities to meet
emerging requirements.

Join the Finacle Digital

Banking Community,
and Accelerate Your AI
Journey, Today.

17 | External Document © 2024 EdgeVerve Systems Limited

 [email protected]

www.finacle.com

www.linkedin.com/company/finacle

twitter.com/finacle

. . . . .......................................................................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............................................................................................................................................................................................................................................

For more information, contact [email protected] www. finacle.com

. . . . .......................................................................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............................................................................................................................................................................................................................................

© 2024 EdgeVerve Systems Limited, a wholly owned subsidiary of Infosys, Bangalore, India. All Rights Reserved. This documentation is the sole property of EdgeVerve Systems Limited (“EdgeVerve”). EdgeVerve believes the information in this document or
page is accurate as of its publication date; such information is subject to change without notice. EdgeVerve acknowledges the proprietary rights of other companies to the trademarks, product names and such other intellectual property rights mentioned
in this document. This document is not for general distribution and is meant for use solely by the person or entity that it has been specifically issued to and can be used for the sole purpose it is intended to be used for as communicated by EdgeVerve
in writing. Except as expressly permitted by EdgeVerve in writing, neither this documentation nor any part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, printing, photocopying,
recording or otherwise, without the prior written permission of EdgeVerve and/ or any named intellectual property rights holders under this document.