How a Major Bank modernized wholesale banking �to deliver �self-service with Prophecy

In wholesale banking and many other industries, data can be a critical advantage in better understanding customers and streamlining workflows. Speeding access to this data and federating usage across teams is key to delivering on this data investment.


At a major wholesale bank, it was taking too long to build the data assets needed by the business while demand for new data continued to grow. To unlock access to data, the team focused on enabling citizen data engineers in the business with self-service tools to federate not just data access but data transformation.

Using Prophecy Data Transformation Copilot, the team has been able to reduce complex pipeline development times and bring tbusiness subject matter experts into the process, scaling the resources for data transformation and capturing valuable business insights in standard, re-usable code.

Join us for this session to
-Learn about the key role data transformation plays in unlocking data
-Hear how speeding data transformation is key to better serving customers
-How taking a new approach to data transformation with Prophecy can speed access to data by up to 10x

1. Matt Turner Director Product Marketing Prophecy @ matt_turner_nyc How a Major Bank modernized wholesale banking to deliver self-service with Prophecy Big Data London September 19, 2024

2. We’re Baaaack ! (thanks genAI ) Randy Bean Wavestone 2024 Data and AI Leadership Executive Survey genAI strategy IS data strategy Adam Selipsky (then) CEO AWS Re:Invent 2023 June 2024 87.9% say investment in data & analytics is a top priority 2018 Scary CDOs say they are the “belle of the ball”

3. Enable users Process data Product analysis MORE is needed by enterprises Data engineers Data analysts Data scientists Structured Semi-structured Unstructured Business intelligence Generative AI Precision ML Reports Oversubscribed Blocked

4. Development Observability Orchestration Execution Data Existing solutions lack performance or usability Legacy ETL Cloud Data Platform Pros Enables all users Higher productivity Cons Locked-in Low performance Pros Code power High performance Cons Fewer users Low productivity git airflow observability sql dwh spark data Metadata

5. The data transformation iron triangle Powerful Intuitive Intelligent

6. 55% Reduction in the time it took to complete a task What would you like to do today? Copilots with AI change the game

7. Prophecy is the productivity layer for users Data engineers Data analysts Data scientists Data transformation copilot Business logic - code on git SPARK SQL

8. Enable every user Ease of use Remove barriers High productivity Low code Drag & drop Spark & SQL Data analyst Data engineer

9. Makes recommendations Converts natural language to business logic Complete pipelines Generate tests Writes documentation Suggests fixes for errors More productivity for every team member Higher Productivity per User Focus on Analytics

10. Wholesale Banking Use Case

11. Wholesale Banking Banking services to corporations and financial institutions Focus on partnerships and co-creation of products Data-driven transformation of frontline to streamline processes and better serve customers

12. Wholesale Banking Transformation James Bickerton, Global Head of Client Development, HSBC August 2022, McKinsey Insight How Data and Analytics are Transforming the Wholesale Bank at HSBC HSBC is not just modernizing and automating ways of working, we’re simplifying day-to-day tasks for bankers and equipping them with tools to understand their clients holistically.

13. Data Assets Drive Wholesale Banking Massive amounts of data (19PB+) Complex, authoritative data on parties and products Data team delivers data assets Enable downstream customers drive innovation and deliver new products

14. The Challenge To support business growth sustainably and deliver cost benefits, need to build Data Assets in a cheaper, faster, and better way Significant amount of elapsed time and costs associated with the end-to-end delivery of Data Assets. Average of 5-6 weeks build lead time High resource utilization during build High ‘Barrier to Entry’ skill set required

15. Current State Data asset development relies heavily on engineering resources during the development and test phase Requestor Low tech skills Designer Med/high tech skills Builder High tech skills Business Requirements Data Exploration Asset Design Asset Testing Asset Development Test Release Production Release Production Validation Deployment Asset Testing Multiple steps, multiple tools Back and forth between subject matter experts and technical team Heavy engineering resource involvement

16. Refining with Prophecy New process to eliminate unnecessary handoffs and the high technical barrier to deliver data assets Reduce steps, single tool Reduce back and forth with visual interface Eliminate dependency on engineering resources Requestor Low tech skills Designer & Builder Med/high tech skills Business Requirements Data Exploration Data Set Design & Develop Automated Data Set Tuning Deployment Business Validation

17. 17 Omics PROs RCTs Registries Aetion® Discover Real-world data Claims Devices EHRs Labs Aetion Data Model (ADM) 17 With Prophecy, we can develop perfect pipelines fast Scale in Prophecy Visual Development Collaboration Reusable Validation Components Pipelines Datasets Outputs Git

18. Data Asset Creation Transformed Speed pipeline development with intelligent visual interface and reduce dependency on engineering resources Prophecy Data Transformation Copilot technology optimizes the process and brings other improvements. Reduced ‘Time to Market’ Improved resource efficiency Near real-time review during build Automatically generated spark/Scala code

19. After first phase: ~55% reduction in data asset build time ~50% in Data Asset development costs 5-10x increase in data asset development Additional benefits Lower ‘Barrier to Entry’ skills set level required to build data assets Automated Data Asset unit testing Results

20. Data Transformation Copilot Artificial intelligence Compiler Visual interface

21. Migration Copilot code on git Spark Data Platform read read join clean write read read join clean write import point sources, targets to cloud data run, and fix any transform mismatch visually 1 2

22. Integrated and comprehensive Single pane of glass with existing systems of record, without adding another system of record Data Transformation Copilot Development Metadata Deployment Governance Observability Business logic Spark code SQL code Airflow code Git, CI, CD system Storage Compute SPARK SQL Cloud Data Platform Cloud Data Platform

23. At a glance HQ in Palo Alto, office in Bangalore 100+ employees in US, UK, and India Raised over $65M , Series-B 2023 5+ years in business Trusted by the largest enterprises Investors Partnerships

24. Questions? Thank You!

25. Continue the conversation at our stand X224 Request Personalized Demo

26. Resources If Your Data Is Bad, Your Machine Learning Tools Are Useless by Thomas C. Redman https://hbr.org/2018/04/if-your-data-is-bad-your-machine-learning-tools-are-useless Wavestone 2024 Data and AI Leadership Executive Survey https://wwa.wavestone.com/app/uploads/2023/12/DataAI-ExecutiveLeadershipSurveyFinalAsset.pdf “ GenAI is part of data strategy”, Adam Selipsky , (then) CEO AWS, Re:Invent 2023 https://youtu.be/mM2HMK3ufTo?t=4704 Big Data Is Back and Is More Important than AI by Adam Conway https://www.linkedin.com/pulse/big-data-back-more-important-than-ai-adam-conway-0v5ac/