Innovation

Innovation

Innovation Week AI ML Enrichment Executive Summary This white paper examines the Enhanced Transaction Experience initiative, designed to address customer needs and pain points related to understanding and locating specific transactions. By leveraging advanced search capabilities, enriched transaction details, and an intuitive user interface, this initiative aims to significantly reduce call volumes, increase customer satisfaction, and generate substantial cost savings over a five-year period. Background Current transaction inquiry volumes for Card and Retail banking are as follows:

Card: 4.0 million monthly calls, 3.2 million customers, 784,000 monthly transaction inquiry calls Retail: 4.9 million monthly calls, 3.4 million customers, 1.2 million monthly transaction inquiry calls

Combined, this represents 8.8 million monthly calls, 6.3 million customers, and 1.9 million monthly transaction inquiry calls. Notably, 800,000 customers call within 4 hours of being digitally active with a transaction inquiry. Enhanced Transaction Experience Features

Account Switching and Multi-Selection

Allows users to switch between Card and DDA accounts or select multiple accounts simultaneously

Advanced Search Capabilities

Supports search using combination terms (e.g., "Uber refund", "Food expenses July", "Travel Charges more than $150") Search parameters include date, amount, merchant, spending category, transaction type, product name, product type, account number, and keyword

Enriched Transaction Details

Provides cleansed descriptions, multiple merchant information, and merchant logos/category icons

Multiple Merchant Information

Displays primary and secondary merchants (e.g., DoorDash delivery from Chipotle, Purchase from Amazon Seller, User ride paid through Apple Pay)

Consolidated Filters

Reduces long list of choices into ~9-11 logical values Represents filters with visual icons for quick recognition Filters and search complement each other

Customer Scenarios and Resolution Example 1:

Customer wishes to understand additional details on a particular transaction from February Identifies the primary merchant, but not the secondary merchant or purpose of which the transaction was made Agent tries to help the customer with the merchant category as Travel and suggests calling the agency

New Experience:

Customer can search by primary/secondary merchant to locate transaction New experience will list 2 levels of merchants to help recognize transaction with ease Customer searches for Priceline New experience will surface primary (Priceline) and secondary merchant (United Airlines) to help identify transaction with merchant logo as additional visual cue

Example 2:

Customer wishes to find a transaction for a laptop purchase made sometime back Doesn't remember exact date, only able to recollect the amount range (~$2000, possibly from Apple Store) Agent searches and finds no transaction above the mentioned amount range Guides on how to search the transaction by applying category filter as a workaround

New Experience:

Customer: "I am trying to find a laptop purchase of approx. 1000 USD" Customer is looking for a transaction with a specific amount and purchase category Search "Shopping above $1000" and results will show all shopping transactions above $1000

Natural Language Understanding (NLU) Model The NLU model plays a crucial role in the Enhanced Transaction Experience, consisting of the following components:

Tokenizer + Featurizer Named Entity Disambiguation (NED) Intent + Named Entity Recognition (NER) Post-processing Experience API

Out of 1,436,029 NLU searches analyzed:

210,744 distinct search terms were used Top search terms include Amazon (21,328), Apple (14,744), Interest (9,438), Costco (8,878), United (7,364), Netflix (6,709), Hotel (6,661), DMV (6,108), Hulu (5,917), and Gas (5,683)

The NLU model classifies search terms into various categories:

Merchant: 63.0% No Resolution: 21.9% Amount: 7.2% Date: 2.2% Combination: 2.1% Spending Category: 1.9% Transaction Type: 0.9% Account Number: 0.4% Product Type: 0.3% Product Name: 0.2%

Implementation Roadmap The Enhanced Transaction Experience will be rolled out in six phases:

Phase 1: 1% of customers Phase 2: 5% of customers Phase 3: 15% of customers Phase 4: 20% of customers Phase 5: 50% of customers Phase 6: 100% of customers

Throughout the implementation process, the following activities will be undertaken:

Reviewing customer feedback Addressing data quality issues Addressing performance and scalability Setting up a new card cluster to isolate card traffic, allowing for scalability Conducting deep dives into the NLU model to assess the accuracy of customer inquiry resolution Reviewing call volume

Projected Impact Over a five-year period, the Enhanced Transaction Experience is projected to generate the following savings:

Card: $3.2 million Retail: $4.1 million Combined: $7.3 million

By improving the customer experience and reducing the need for transaction inquiry calls, this initiative will lead to significant cost savings and increased customer satisfaction. Innovation Week Event Description The new transaction list & search experience provides an enriched view of transactions and enhanced capabilities to search through them across multiple accounts. For every transaction customers can now view clean merchant names , logos, category icons etc which will help them easily identify the transactions; as well as search through them using natural language phrases without being restricted to keyword searches. Already live to 20% personal card customers, 100% targeted early June 2024.

DUDE Deck 'June 2024\n\nDUDE\nSetting the product vision\n|\n\n1\nDUDE\n1\n2\nCounter Party Tagging\n14\n3\nMerchant Tagging\n26\n4\nArchived Slides\n73\n\nImproving Data Enrichment for More Clarity\nDevelop a centralized capability that takes raw, unstructured customer transaction data and transforms it into standardized data attributes for use throughout the bank.\nOUR MISSION\nOUR MULTI-YEAR VISION\nEnhance customer satisfaction and financial confidence.\nDUDE offers expanded transaction coverage and enhanced data reliability which improves user satisfaction and trust in bank services.\nTransform banking experience with real-time, transparent transactions.\x0bDUDE offers near real-time enriched data for operational and analytical use with expanded transaction coverage and enhanced data reliability.\nSignificantly cut call center load, saving $15M annually.\x0bDUDE connects directly to sources to reduce data latency, increasing customer comprehension & reduce confusion.\n\n\n\n\nWHAT OUR VISION & MISSION ENABLES\nDUDE| Integrated Product Vision, Mission, and Outcomes\n1\n\nWhat we do \x0b\x0bTransform unstructured data into usable information\n\n\nDUDE\n2\n\nDigital Utilities Data Enrichment\nTransform unstructured data into usable information for:\ncustomer insights, \nproduct experiences, and \nbusiness decisions\n“Enrich once, use everywhere.”\n\n4\n\nA product vision is…\n3\n\n\n\n…anchored in improving the customer experience\n\n…forward thinking\n…inspirational and motivating\n\nOUR VISION STATEMENT\nBe the authoritative source for money movement and merchant transaction enrichment data in the Bank.\nMake enriched data easily accessible, hold a high standard of data quality and accuracy, and provide low latency output to end users.\n3\n\nOur Functions\n\nCounterparty Tagging\nMerchant Tagging\nEnrich once, use everywhere.”\n\nDUDE\n5\n\nTRANSFORM UNSTRUCTURED DATA INTO USABLE INFORMATION\n7\n\nCUSTOMER PAIN POINTS\n\nOUR PROBLEM STATEMENT: 80% OF DATA IS UNSTRUCTURED\n\n8\n\nTo deliver trusted knowledge everywhere, Utilities has to be built to an uncompromising set of standards, and has to cater all data from SORs needed to serve query traffic\nFrom producer to consumer\n\n>20 Billion\nrecords ingested per day\n99.99% availability\nonly down 52 minutes per year\n1second E2E\nReal-time data consistency across all channels\nEngineer for every product\nwithout customization\n50-100ms\nperformance\n\n\n\n\n\n25K+ TPS\nSupport consumption at scale\n\nEnrichment: BRIE Findings\n~50 billion\ntransactions available\nData available dates to 2020\n600m-900m\ntransactions per month\n7 critical fields Identified\n96 fields to explore\nfor data build process\n130 fields total\navailable, 34 (26%) are blank\n\n\n\n\n\n\n\nNext Step\ndive deeper into deciphering remaining items\n\nETUDE Q3 Target Priorities\nQ3 Target Deliverables\nTECH MODERNIZATION\nFinalize the exit of the Legacy Data Center and complete migration of all components to the ETU SEAL ID, to streamline supportability\nImplement Pay in Four API as a first AWS use case for ETU\nBUSINESS INITIATIVES\nComplete all pre-dependency work for\xa0Multi-Card 1.0\xa0to enable integration testing and planned E2E pilot\nComplete all pre-dependency work for\xa0Rewards Transaction Earn\xa0experience to enable integration testing\nImplement\xa0Pay By Bank\xa0ACH enrichment rules in new AWS environments \xa0\nProvide One Chase Service\xa0Declined\xa0DDA Transaction details\xa0to support reduction of incorrect customer call\xa0redirects\nTBD – Kick-off Deposits 2.0 - Phase 1 providing\xa0foundational capabilities for first use case (UAT only)\nSTABILITY AND ADOPTION\nGA rollout for v4 Card Transaction History\xa0\xa0with Channels, P&I, and Aggregators (90%+ of traffic will be Primary)\nEnsure\xa0Search / Transaction List GA\xa0is stable and support\xa0Core Card\xa0experience rollout\nInvest in our App Health\nPhoton upgrade, AU re-point, and other activities needed to ensure stability\nContinue migration of consumers to DDA Transaction History and Enhanced Transaction Detail v3, as well as\xa0realignment with the System of Record to resolve Data Quality and Reconciliation gaps, in preparation for\xa0DDA as\xa0Primary\nBroad Organizational Goals\n\nWhy\nTake traffic off SOR\nConsistent data across Digital, CB, ODS, Analytics\nMonetization opportunities with transactions\nReduce operational costs\nExamples\nStability/Resiliency\n\nCUSTOMER JOURNEY WITH UNSTRUCTURED DATA\n\n\n\n\n9\n\nOUR OPPORTUNITIES\nENHANCE DIGITAL CUSTOMER EXPERIENCE\nSUPPORT DATA-DRIVEN BUSINESS DECISIONS\nSUPPORT DATA-DRIVEN BUSINESS DECISIONS\n\nUNSTRUCTURED DATA REPRESENTS THE FASTEST GROWING AND LARGEST DATA TYPE\n10\n\nChase has multiple use cases use cases involving cost-savings, revenue generation and analytical insight opportunities\n\n\nPoint-of-Sale Transactions\nMoney Movement Transactions\nApplication\nLogs\nWeb Interactions\nUnstructured Data\n\n\n\n\n\n\n\n\n\n\nEnhanced Controls\nRevenue Generation\nCost Containment\nCustomer Experience\n34\n\nOUR IMPACT\n\n11\n\nOur Work in Action\n\n12\n\nOur Strategy:\nMake enriched data easily accessible, hold a high standard of data quality and accuracy, and provide low latency output to end users\n\n1\n\n2\n\n3\nTARGET NEW TRANSACTION TYPES\nFacilitate better monitoring\xa0and reporting on new\xa0transaction types like Pay-by-bank.\n\nMONITOR AND REPORT NEW TRANSACTION TYPES EFFECTIVLEY\nOffer a standardized, consistent\xa0dataset for uniform business\xa0capabilities.\n\nBUILD TRUST AND TRANSPARENCY IN BANKING OPERATIONS\nSupport broader customer service\xa0improvement goals.\xa0\n\n\n13\n\n2024 DUDE Product Milestones\n\nQ1\nQ2\nQ3\nQ4\nBenefit\nCounterparty Tagging\n\n\n\n\nSetting up necessary infrastructure within AWS for Counterparty Tagging.\nCreating an ingestion pipeline to handle diverse data sources for Counterparty Tagging.\nDeveloping the logic for an enrichment engine to handle ACH transactions.\nEstablishing a delivery pipeline for distributing enriched data within the firm.\n\n\nMerchant Tagging\xa0\n\n\n\n\nEnhance Middleware Technology (MT) with Failover Cassandra DB changes and auto-failover for improved efficiency, performance, cost savings, and resilience.\nEnable retagging capability for historical data.\nImplement a Kafka publisher for real-time transmission of debit card transactions,\xa0\nMigrate 2 years of historical credit and debit card transaction data to the data lake.\nIntegrate the transaction feed from First Republic Bank into JPMC system, applying the same enrichment process used for JPMC\xa0transactions to\xa0First Republic Bank data.\nData Consumption: Ingestion Pipeline for Counterparty Tagging\n\uf0ea Run the Engine Enhancements\nInfrastructure: Setup for Counterparty Tagging\n\uf0ea New Card Cluster Retag\n\uf0ea Debit Card 3.0\nData Delivery and Distribution: Delivery Pipeline for Counterparty Tagging\nEnrichment: Enrichment Engine Logic for Counterparty Tagging\n\uf0eaHistorical Load to Data Lake\n\uf0ea\xa0FRB Historical Transaction Enrichment\n16\n\nVision Statement\n\nBusiness Objective\n\nTarget Users\nBe the authoritative source for money movement and merchant transaction enrichment data in the Bank.\n\nMake enriched data easily accessible, hold a high standard of data quality and accuracy, and provide low latency output to end users\n\nData analysts and \ncustomer-facing product teams\n\n\n\nDUDE\xa0has two discrete enrichment processes: (i) merchant and\xa0(ii) counterparty\n“Enrich once, use everywhere.”\n\nDigital Utilities Data Enrichment Product | Vision & Strategy\n21\n\nThis data is generated as “exhaust” from payment processing systems—it’s\xa0not designed\xa0for analytics or customer intelligibility\n\n25\n\nDUDE: Transforming Data Into Information\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n22\n\nData comes in multiple forms…\nStructured\n\n\nSemi-Structured\n\n\nUses pre-defined data models filled with labels, numbers and values \nExcel spreadsheets, data tables, forms\n\n\nUnstructured\n1\n0\n0\n1\n1\n0\n1\n0\n1\n1\n1\n0\nUses pre-defined data models filled with labels, numbers and values \nExcel spreadsheets, data tables, forms\nUses pre-defined data models filled with labels, numbers and values \nExcel spreadsheets, data tables, forms\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n32\n\n2024 DUDE Product Milestones by Quarter\n\n\nQ1\nQ2\nQ3\nQ4\nBenefit\n1.\nData Lake\n\n\n\n\nSupport GA release for Transaction Search credit and debit card experience.\n\n2.\nDecommission Merchant Tagging 2.0\n\n\n\n\nReduce manual processes to maintain MT2.0 rule logic.\n3.\nPlatform migrations\n\n\n\n\nOperational platform migration aligns to CCB initiative to consolidate data centers\nDUDE Analytics data infrastructure will promote transaction enrichment data to AWS to enable integration with new analytics interfaces (e.g., Snowflake)\n\n4.\nAmazon Co-brand Card\n\n\n\n\nEnables Amazon Co-brand card holders to view merchant refund notifications on Amazon.com.\n5.\nCounterparty Tagging Modernization & Expansion\n\n\n\n\nEvaluate and modernize counterparty workflow and technology stack to enable key businesses cases (payments & wealth management).\nExpand counterparty tagging logic for additional transaction types (ATM and FedNow payments).\n\n\n \n\n\n\n\nEvolve automation capabilities for data quality monitoring and processing tools.\n\uf0ea Credit Card Enhancements\n\uf0ea Decommission MT2.0\n\uf0ea Operational Platform DCM\nKey\n \uf0ea Platform\n\uf03c Feature\nDRAFT\n\uf0ea Initial Design & Deployment for Debit Card\n\uf0ea Debit Card Enhancements\n\uf03c Amazon Co-brand Card\n\uf0ea Counter Party Tagging\n6. Platform Enhancements\n\uf0ea Platform Enhancements\n\uf0ea Analytics Platform Migration to AWS\n35\n\nProblem: Unstructured data represents the fastest growing and largest data type and can unlock a competitive advantage\n\nNote: 1 https://mitsloan.mit.edu/ideas-made-to-matter/tapping-power-unstructured-data\n2. 2019 survey by Deloitte\xa0\n3. https://www.baselinemag.com/analytics-big-data/structured-vs-unstructured-data-including-examples-of-both/\ndata is unstructured\n\n80%\norganizations able to take advantage of unstructured data\n\n18%\n\n25%\nCAGR in global datasphere 10 years\n33\n\n1\nDUDE\n1\n2\nCounter Party Tagging\n14\n3\nMerchant Tagging\n26\n4\nArchived Slides\n73\n\nCounterparty Transaction Enrichment | Feature Overview\nCounterparty Tagging\nCounter Party Tagging\nDescription\nEnriches non-card DDA transactions with standardized party and financial institution names and insights into transaction purpose. Coverage rate of 44% (name) and 70% (purpose).\nStatus\nMVP live;\xa0\nv1.0; targeted for Q4 2024\nTransaction Types\nACH, Wires, and Real-time payments\nEnrichments\nClean counterparty name\nCounterparty details (including financial institution)\nTransaction purpose\nWhere it’s used\nAnalytics: data source for select dashboards and ad hoc analysis\xa0\n17\n\nPrivate and confidential – For internal discussion purposes only – Data Sourced Manually – not to be shared outside the bank.\nThe Beta PBB dashboard gives insights into the growth of PBB overall and by payment flow, use case, transaction type, industry code, merchant and risk tier\nFollow the pace of growth for PBB via TANs\nView the breakdown and growth of emerging \nPBB transaction types\nFilter into deep dives (e.g., Uber, Digital Wallet – Funding & P2P)\n¹See appendix for risk tier description\nUnderstand how fast higher-risk use cases are growing\n\n\nTrack payment volume growth by risk tier weekly / monthly\n\nUnderstand QoQ Growth by Risk Tier\n4\n\n6\n\nCounterparty Transaction Enrichment | Phase 1 Architecture Diagram\nACH – transaction focus\n\nCounterparty Transaction Enrichment | Target State Architecture Diagram\n6\n\nSetup Cluster\nCreate Cluster -> Configure Compute\nMultiple internal and external sources are compiled to source transaction and customer data.\nProcess\nPre-Process\nMatching\nLocal Output\n\nBanking Reservoir & Intelligence Engine\n(ACH, Wire, Zelle, P2P, ACATs)\nAccuity\n\n\n\n\n\n\n\n\n\nDDA Transactions\nSource File Ingestion involves collection and ingestion of raw data from various sources\n2\nBRIE is data reservoir ingesting ACH\xa0transaction\xa0data from ACHCORE for\xa0near real-time (NRT) delivery. ACH Data from BRIE is mapped during pre-processing for efficient processing and enhanced for data quality.\n\nAccuity provides financial institution names (incl. Routing and Account Numbers) to facilitate standardized financial institution names. It\'s provided on a recurring\xa0basis for accurate and up-to-date information on financial institutions\' routing codes and payment data which is leveraged for accurate mapping and enriching transaction data.\n\nCustomer Reference Data is leveraged to enrich transactions further, as it helps in identifying and linking transactions to customers. This facilitates in identifying transactions associated to financial institutions, merchants or customers.\n\nDemand Deposit Account data provides additional transaction description information that allows for additional context to be provided (incl. Transaction Purpose). This is the backbone of CPT being able to enrich non-card DDA transactions with standardized transaction purposes.\n\nData will be ingested into staging area, transformed based on requirements, and written\xa0to target system (CPT 1.0) by June 2024.\n\nPull Data\nIngest Raw Data -> Write to Staging\n\nPrepare Data\nRestructure, Transform, Validate\n\nSchedule Pipeline Jobs\nCreate Job -> Schedule Pipeline Job\n\nMatching\nPreprocess Data\nRules engine matches standardizes entity name and identifies transaction purpose.\nRun rules engine\nName Entity\nRemove data anomalies and invalid records\nPurpose Entity\nLocal Output\nName Entity\nPurpose Entity\nProcessing and matching involves refining the data for improved accuracy, clarity, and utility\n3\n\nIdentify and eliminate inconsistencies in the data, such as duplicate entries, incomplete records, or irrelevant information, to ensure data quality.\nStandardize/Normalize the format of the data elements to ensure consistency across the dataset. This includes identifying information in a uniform manner.\n\n\n\nIdentify and standardize the names of entities using regular expressions (regex) and other logical constructs to match and standardize entity names, ensuring consistency and accuracy in identifying transaction parties\nIdentify the purpose of each transaction by categorizing transactions based on their nature and intent, using predefined rules and patterns to assign a standardized purpose to each transaction.\n\n\n\nStore the outputs locally in a redshift database for subsequent use, monitoring, and analysis\n\nCleanse\nRemove Anomalies & Invalid records\n\nCombine\nJoin with DDA & Customer Data\n\nEnrich\nImplement Enrichment (Name, Purpose, Etc.)\n\nTest & Improve\nContinuously test and Improve\nEnrichment is scheduled to begin in June and will be carried out until the end of Nov\'24.\xa0Continuous improvement will be carried out until Q1-25.\nDistribute\nData Lake\nKafka (Real Time Streaming)\n\nCPT 1.0 Development Phases\n-Development of CPT 1.0 will be carried out over five distinct phases that will\n\n\nKey Steps\n\nPhases\nInfrastructure (App Build)\nConsume Data\nData Delivery (Hydration)\nTesting & Optimization\nH1-2024\nH2-2024\nIntegrate Enrichment Logic\nInfrastructure Setup: Setup Bitbucket, Terraform, and Sonar.\nData Transfer & Storage: Secure file transfer and Amazon S3.\nCompute & Processing: Amazon EMR, Lambda, and AWS Step Functions.\nWorkflow Orchestration: AWS Event Bridge & Step Functions for process coordination.\nOnboarding: Secure pipeline with Databricks.\nData Integration: Ingest ACH, third party, and customer data.\nField Mapping: Precise selection and filtering of fields through analysis of data structure.\nPre-Processing: Transform data structure and identify/map key ACH fields.\nEnrichment: Implement logic for additional fields (incl. Financial Institutions, Transaction Type, etc.)\nData Governance: Adherence to data governance and metadata registration.\nPipeline Setup: Set up UAT, PROD, and Data Pipeline.\nData Loading: Load historical\xa0 into Data Lake.\nAnalytics Enablement: Enable analytics via Tableau and Snowflake.\nMonitoring Setup: CloudWatch monitoring and Tableau dashboards.\nRe-enrichment Process: Processes for re-enriching historical and future transactions.\nQuality & Adaptability: Continuous testing and system adaptability.\nRe-enrichment Implementation: Deploying processes to fix and update transaction data.\n\n-\n\nCounterparty transaction enrichment: platform modernization will eliminate current performance issues for a business-critical data set\n\n\nCounter Party Enrichment\n(Current Status) \n\nProvides clean names and details for counterparties, such as financial institutions and transaction purpose for ACH, Wire and Zelle transactions\nMVP launched in October 2021\nBuilt in partnership between FAST team and DUDE product\nCoverage: ~280MM transactions per month\nHigh-demand from end users across CCB customer-facing and analytics teams\n\n\nProblem statement\nTransaction data source: MANTAS is a Global Financial Crimes Compliance data base not meant to support current use cases\nData latency: batch file data loads result in 11-15 day processing delay\nEngineering capacity shortage: one (1) engineer supporting platform modernization work\n\n\nRecommended solution\nInvest in engineering capacity: add one (1) engineering team with AWS and big data experience\nReduce processing latency: modernize tech stack with AWS services\nIngest new data source: Utilize Business Reservoir Intelligence Engine for real-time transaction data to improve data scalability and agility\n\n\n\nCOUNTERPARTY ENRICHMENT\n\nRevenue Generation\nExpense Reduction\nEnhanced Controls\nConsumer Bank\n \n \n \nWealth Management\n \n \n \nConnect Commerce (Payments)\n \n \n \nHome Lending\n \n \n \nAuto\n \n \n \nUniversal Search\n\n \n\n\n\n\n\n\nBENEFITS TO BUSINESS\nItem\nCurrent FTE (4)\nAllocation\nTimeline\n\n\n\n\n\nQ2\nQ3\nQ4 \nLate 2024\nData Center Migration\n2.0\n\n\n\n\nRTE support\n1.0\n\n\n\n\nCPT Modernization\n1.0\n\n\n\n\n\n\n\nCURRENT ROADMAP & ENGINEERING ALLOCATION\n\nTeam is currently over allocated\n2\n\nEnd Users\nPayments\nBankers\nAnalysts\nWealth Management\nDifferent Chase LoBs (CIB, AM etc.)\n\nData Lake\nKafka Publisher\n\nVarious Distribution Channel for End users\n\n\n\n\n\nData will be securely transmitted to end-users through the Data Lake and Kafka via ETU\n4\n\nThis process is set to end in Feb\'25 when data will be available within the Data Lake.\nEstablish a pipeline from CPT to the Data Lake so that analysts including the pay-by-bank insights team can consume enriched data for analytics via platforms like Tableau and Snowflake.\n\nSet up environments for testing and production, ensuring data security, establishing a structured storage framework, and adhering to governance processes for cybersecurity and application stability.\xa0\n\nSet up a Kafka publisher-consumer mechanism to ensure that enriched data is seamlessly integrated into Chase\'s customer-facing platforms, enhancing the customer experience by providing clearer transaction\n\nInitiate PTX\nPermit to Build/Deploy/Operate\n\nOnboard SEAL\nEstablish Data Publisher Roles\n\nRegister Data Pipeline\nData Schema, Business Meta Data\n\nPublish Data\nWrite data to Managed Data Lake\n\nCounterparty enrichment examples\n\nMe-to-Me Identification\nMe-to-me flagging identifies transactions where customers move money between their own accounts (ex: checking to brokerage account)\n\n\nPurpose Enrichment\nPurpose enrichment provides purpose and context for a given transaction\n\nInbound Check Identification\nCounterparty tagging provides name identification using historical ACH / Wire / RTP transactions\n1: GROUPO MERC SERV LLC ID: 801378\n2: AQ WELL WORLD LLC\n3: INTUIT PAYM SOLU LLC\nGroupon\nAqua Wellness World\nIntuit Payment Solutions\nSender: Allie Bernand \nReceiver: Carrington Mortgage \nDescription: Autopay\nMortgage Payment\nSender: Martinez Carlos\nReceiver: Carlos R Martinez\nMe2Me: TRUE\nInbound Check\x0bAccount #:12345678\nACH / Wire / RTP History\x0bAccount #: 12345678\x0bName: MIAMI AQUATICS\nInbound Check\x0bAccount #:12345678\x0bName: MIAMI AQUATICS\n\nStandardized Financial Institution Names\nCounterparty tagging provides standardized institutional names for use across multiple applications\n1: HSBC\n2: HSBC Bank\n3: HSBC Bank USA, National Association\nHSBC Bank USA, National Association\n\n\n\n\n\n4\n\nCounterparty Tagging enriches non-card DDA transactions with standardized party and financial institution names and insights into transaction purpose.\n\nCPT | Product Strategy:\n\n1\n\n2\n\n3\nIMPROVE CUSTOMER EXPERIENCE AND EXECUTIVE INSIGHT\n\nFacilitate better monitoring and reporting on new transaction types like Pay-by-bank.\n\nMONITOR AND REPORT NEW TRANSACTION TYPES EFFECTIVLEY\nOffer a standardized, consistent dataset for uniform business capabilities.\nBUILD TRUST AND TRANSPARENCY IN BANKING OPERATIONS\n\nSupport broader customer service improvement goals.\xa0\nCPT 1.0 targets ACH,\xa0Zelle, Wires, and Pay by Bank customers, offering a standardized dataset for uniform data and consistent business capabilities, aligning with broader customer service improvement goals.\n52\n\nWe\'ve moved from an ambiguous / resource focused discussion to proactive partnership with the Pay By Bank team\n\n\n\nSource Files Ingestion\nExtracted and Loaded 1.5M+ transactions from BRIE to Dev\n\nProcessing & Matching\n\nDistribution\n\nVia APIs\nData Lake\nKafka Publisher\n\nMatching\nPreprocess Data\nAwaiting enrichment rules from D&A\nRun rules engine\nName Entity\nRemove data anomalies and invalid records\nPurpose Entity\nUAT Environment Underway\n\nBanking Reservoir & Intelligence Engine\n(ACH, Wire, Zelle, P2P, ACATs)\nAccuity\n\n\n\n\n\n\n\n\n\nDDA Transactions\nLocal Output\nName Entity\nPurpose Entity\n\n\n\n\n\n1\nJun\'24\nNov\'24\nFeb\'25\nOn Track\n\nOn Track\nAwaiting\nCounterparty Transaction Enrichment\nAnalytical\nOperational\nPay by Bank Insights Team\nCustomer Facing (channels)\n\nSource Files Ingestion\nMultiple internal and external sources compiled to source transaction and customer data\n\nProcessing & Matching\n\nDistribution\n\nEnd Users\nPayments\nBankers\nAnalysts\nWealth Management\nDifferent Chase LoBs (CIB, AM etc.)\n\nVia APIs\nData Lake\nKafka Publisher\n\nMatching\nPreprocess Data\nRules engine matches standardizes entity name and identifies transaction purpose\nRun rules engine\nName Entity\nRemove data anomalies and invalid records\nPurpose Entity\nVarious Distribution Channel for End users\n\nBanking Reservoir & Intelligence Engine\n(ACH, Wire, Zelle, P2P, ACATs)\nAccuity\n\n\n\n\n\n\n\n\n\nDDA Transactions\nLocal Output\nName Entity\nPurpose Entity\n\n\n\n\n\nCounterparty transaction enrichment modernization roadmap will provide near real-time non-POS DDA enrichment capabilities at scale\nJun\'24\nNov\'24\nFeb\'25\n24\n\nBuilt a cloud infrastructure for our vision and upstream partnerships. The foundation will allow us to scale quickly across a variety of use cases (incl. PBB)\n3\n\n2023 & Q1-24\nQ2-24 & Q3-24\nH2 2024\n\n\nGetting started\nIntegrate with Data Source\nBuild Enrichment Logic & Deliver\nCPT 1.0 POC launched in Q3’23\xa0\nDefined schema & taxonomy standards with key stakeholders\nAligned on Key Use Case\nIntegrated BRIE Dev Environment\nIntegrate with upstream data sources\n\xa0ACH Data via BRIE\nCustomer Reference Data\nThird Party Reference Data (Accuity)\nSet up AWS Infrastructure and environment\n\n\nBuild enrichment logic\nDeliver enriched data to the Data Lake\nIntegrate with ETU for Operational use Case\n\n\n\n\nDraft\n\nMitigating PBB Bank Exposure through CPT 1.0\nRising demand for PBB accelerates development of CPT 1.0 with increased funding/resources\n\nSituation\n\nKey Items\n\nDescriptions\nComplication\nResolution\nRising demand of PBB makes insights/reporting more critical\nPay-by-Bank Insight Team established to analyze ACH PBB transactions\nPBB Team relies on excel-driven manual approach to manage TAN Data\nInefficiencies and risks in current ACH/TAN reporting process\nInsufficient data to make data driven decisions\nNeed for real-time analytics for executive decision making\nProvide real-time analytics via Data Lake to support executive insight\nBuild CPT 1.0 with modernized infrastructure that future-proof non-reporting environment use cases(i.e. Channels)\nExecutives need visibility to monitor the growth of this payment method for product roadmap prioritization.\n42\n\nCounterparty transaction enrichment: platform modernization will eliminate current performance issues for a business-critical data set\n\n\nCounter Party Enrichment\n(Current Status) \n\nProvides clean names and details for counterparties, such as financial institutions and transaction purpose for ACH, Wire and Zelle transactions\nMVP launched in October 2021\nBuilt in partnership between FAST team and DUDE product\nCoverage: ~280MM transactions per month\nHigh-demand from end users across CCB customer-facing and analytics teams\n\n\nProblem statement\nTransaction data source: MANTAS is a Global Financial Crimes Compliance data base not meant to support current use cases\nData latency: batch file data loads result in 11-15 day processing delay\nEngineering capacity shortage: one (1) engineer supporting platform modernization work\n\n\nRecommended solution\nInvest in engineering capacity: add one (1) engineering team with AWS and big data experience\nReduce processing latency: modernize tech stack with AWS services\nIngest new data source: Utilize Business Reservoir Intelligence Engine for real-time transaction data to improve data scalability and agility\n\n\n\nCOUNTERPARTY ENRICHMENT\n\nRevenue Generation\nExpense Reduction\nEnhanced Controls\nConsumer Bank\n \n \n \nWealth Management\n \n \n \nConnect Commerce (Payments)\n \n \n \nHome Lending\n \n \n \nAuto\n \n \n \nUniversal Search\n\n \n\n\n\n\n\n\nBENEFITS TO BUSINESS\nItem\nCurrent FTE (4)\nAllocation\nTimeline\n\n\n\n\n\nQ2\nQ3\nQ4 \nLate 2024\nData Center Migration\n2.0\n\n\n\n\nRTE support\n1.0\n\n\n\n\nCPT Modernization\n1.0\n\n\n\n\n\n\n\nCURRENT ROADMAP & ENGINEERING ALLOCATION\n\nTeam is currently over allocated\n54\n\ngo/dude-figma\n\n\n\n\n43\n\nSource Files Ingestion\nMultiple internal and external sources compiled to source transaction and customer data\n\nProcessing & Matching\n\nDistribution\n\nEnd Users\nPayments\nBankers\nAnalysts\nWealth Management\nDifferent Chase LoBs (CIB, AM etc.)\n\nVia APIs\nData Lake\nKafka Publisher\n\nMatching\nPreprocess Data\nRules engine matches standardizes entity name and identifies transaction purpose\nRun rules engine\nName Entity\nRemove data anomalies and invalid records\nPurpose Entity\nVarious Distribution Channel for End users\n\nBanking Reservoir & Intelligence Engine\n(ACH, Wire, Zelle, P2P, ACATs)\nAccuity\n\n\n\n\n\n\n\n\n\nDDA Transactions\nLocal Output\nName Entity\nPurpose Entity\n\n\n\n\n\nCounterparty transaction enrichment modernization (CPT 1.0) roadmap will provide near real-time non-POS DDA enrichment capabilities at scale\nJun\'24\nNov\'24\nFeb\'25\n41\n\nTeam Roles and Responsibilities\nPOC Development of Counter Party Tagging ACH/TAN Data Enrichment\nData Enrichment Requirements\nAggregator Name (e.g. Plaid)\nPayment Processor Name (e.g. Paypal)\nMerchant End Point\nSender Financial Institution\nReceiver Financial Institution\nPayment Processor/Aggregator Identification\nProvide additional information on parties involved with transaction\nEntity Type (i.e. C2B/C2C)\nUse Case\nTransaction Type\nIndustry Resolution\nTransaction Context\nProvide additional context on the transaction to understand use case, industry, etc.\nSender Name\nReceiver Name\nSender/Receiver Name Standardization\nProvide clean and consistent names for transaction senders and receivers\nBusiness\nData and Analytics\nDigital Utility Data Enrichment (DUDE)\nDefine data enrichment requirements \nReview and provide feedback on data enrichment POC throughout development\nProvide final sign-off on POC output for insights use\nDevelop of analytical plan to produce transaction enrichment POC\nIdentification of data sources for use in enrichment logic\nDevelop enrichment logic in analytical non-production environment and execute for POC development and refinement\nShort Term: Execute enrichment logic manually for insights in parallel of DUDE operationalization of logic\nEstablish data feeds of all necessary source data for developed enrichment logic into CPT infrastructure\nOperationalize and maintain enrichment logic and integrate with CPT utility\nImplement enrichment logic in production environment Provision output data into a consumable analytical environment\nGiven the data quality of ACH transactions today, a scalable, centrally managed/maintained and automated data enrichment solution is needed to address the business’ needs to provide consistent and accurate insights related to PBB transactions. The Digital Utility Data Enrichment (DUDE) team owns the Counter Party Tagging (CPT) utility which is a central data enrichment utility for deposit transactions which is intended for a multitude of both customer and internal facing use cases across the bank. The development of data enrichment logic to be ingested and maintained in the CPT utility will serve as the foundation of all PBB insights in our target state architecture.\n56\n\nDUDE Role & Responsibility\n\n\n\n\nDUDE is a Product\nDUDE provides Rule Updates\nDUDE are not Consultants\nDUDE does not do Analysis\n\n\n\n\n68\n\nCounterparty Tagging is already being leveraged to generate value in a variety of use cases across CCB D&A\n\n\n\n\n\n\nCounterparty Tagging\nThe Business Bank leverages Counterparty Tagging as the main data asset in the automation of Bionic Banker, which provides insights to bankers on customer deposit flows. This used to be a manual process pulling from LDA and Mantas data and manually tagging counter parties on a monthly basis and is now fully automated using Counterparty Tagging\nCCB Strategy Corporate Development used Counterparty Tagging to analyze the number of customers receiving personal loans and the size of these loans. Records were identified as personal loan payments based on the granular purpose enrichment offered by the tagging that allowed the team to identify “Personal Loan Payments”.\nCCB Strategy Corporate Development enhances the capabilities of their Fintech Dashboard through the entity enrichment offered by Counterparty Tagging to track the performance of various fintech competitors for strategic monitoring\nEnriches non-card DDA transactions with standardized party and financial institution names and insights into transaction purpose\n\n\n\n\nPurpose\nEntity\n\n51\n\n1\nDUDE\n1\n2\nCounter Party Tagging\n14\n3\nMerchant Tagging\n26\n4\nArchived Slides\n73\n\nA preprocessing layer is applied before transaction enrichment.\xa0\nPreprocessing includes a data quality check to confirm eligibility for transaction enrichment.\nKey merchant fields: Merchant name, city, state, postal code, category code, country code, merchant ID, Visa transaction\xa0ID.\n\nTransaction Enrichment includes a waterfall process:\xa0\n\nRaw cleansed:\xa0Partially cleansed transactions that include removal of store numbers, special characters,\xa0symbols and insignificant words and parsing out intermediary prefixes.\xa0\n\nRule based:\xa0Merchant rules are matched based on the regular expression and merchant category codes\xa0in the rule.\n\nEnhanced String Distance:\xa0includes a set features tuned to match merchant transaction data to\xa0merchants and compiled into a decision tree.\nHow Merchant Enrichment is Done:\nTurning raw transaction data into actionable insights\n\n14\n\nHow is Merchant Enrichment Done? - Future state\n\n\nPotential to replace up to 1.17k regex rules\nExpected to save ~515 hours/year of rule maintenance\nExpected to save approximately $1.2MM cost savings from call reduction\nAdditional coverage improvement could save up to $5MM\xa0(for additional analytical usage of the data)\nCounter Party Tagging\n15\n\nMerchant tagging 3.0 provides enriched merchant information from credit and debit card point of sale transactions for both Chase customers and insights for internal teams\nWhat\'s new?\nLower latency:\xa0streaming data rather than batch processing for both pending and posted transactions.\nData Consumption:\xa0data is now consumed through events and APIs.\nNew Data Model:\xa0more flexible handling of merchant hierarchy and intermediaries through transaction-level and merchant level outputs.\nMerchant logos: merchant logos will be displayed alongside enriched transactions.\nNew data model provides more flexible merchant information including:\nUp to four merchant entities can be tagged on a transaction\nNew merchant role codes for each entity → digital wallet, ordering/delivery app, product or service.\xa0\nThe "merchant of record" will be the first merchant entity provided on a transaction.\n30\n\nMerchant Transaction Enrichment | Feature Overview\n\nMerchant Tagging \nMerchant Tagging \nDescription\nCleans merchant names and enriches individual transactions with an 80-85% coverage rate\nStatus\nCredit Card: v3.0 Live\n\nDebit Card: v2.0 live; v3.0 Q1 2024\nTransaction Types\nPoint-of-sale credit and debit card\nEnrichments\nClean Merchant Name\nMerchant details:\ncontact information, \ngeolocation, \nintermediary info, and \ncorporate parent\nWhere it’s used\nTransaction history & details in: Web, Mobile, Servicing, Toggle\nPay-in-4\nPersonalization & Insights recommendations and eligibility\nAnalytics: Branch Reassignment, Commercial Bank consumer spend\n64\n\nAs a result, customer monetary transaction data can be difficult to use and understand\n\n26\n\n27\n\nCustomer-facing use case: Target state power X-Chase Transaction Search so customers don’t have to remember exactly which account they used\n\n\n28\n\nEnhanced String Distance\nFalse\nTrue\nFinal list of decision tree features\nJaccard Weighted Words Similarity\nBinary if first 3 characters match\n# of overlapping bigrams between strings\nDifference in Character Length\n# of overlapping trigrams between strings\nEmbedding Cosine similarity between MCC & SIC\nJaccard Similarity of the phone number\nJaccard Similarity between first 10 characters of both strings\nDifference in number of bigrams between the two strings\nEnhanced String Distance is comprised of a set of features tuned to match transaction merchant data to merchants, and compiled into a decision tree\nThe decision tree creates rules based on cutoff points for each of the features. We limit the tree depth to 6 in order maximize coverage and accuracy while minimizing complexity\nRule: Jaccard Weighted Words less than 0.7\n\nChance of being a match: 30%\nChance of being a match: 80%\n\n\n\n\n\nAdditional rules\nAdditional rules\n\nMerge raw transactions with Third Party Merchant Data on the first four digits of the Zip Code \nKeep Candidates with at least one of the following\nA Jaccard Score over 0.5\nMatching Phone Numbers\n\ns\nThe first step in tagging a transaction is finding candidate merchant in a geographically limited area\n"Raw" transaction details\n\n\nMerchant name\nMCC\nZIP\nNY PIZZA SUPRE\n5812\n10001\nThird Party Merchant Data match\n\nName\nName match score\nFAMOUS PIZZA\n0.40\nNY PIZZA SUPREMA\n0.90\nCAFE BRAVO\n0.00\nSTARBUCKS\n0.00\n. . .\n. . .\n29\n\nMerchant data is enriched using a waterfall of methods\n\n\n63\n\nCounterparty Tagging enriches non-card DDA transactions with standardized party and financial institution names and insights into transaction purpose\n\n\n\n37\n\n38\n\n39\n\n40\n\nCounterparty transaction enrichment: platform modernization will eliminate current performance issues for a business-critical data set\n\n\nCounter Party Enrichment\n(Current Status) \n\nProvides clean names and details for counterparties, such as financial institutions and transaction purpose for ACH, Wire and Zelle transactions\nMVP launched in October 2021\nBuilt in partnership between FAST team and DUDE product\nCoverage: ~280MM transactions per month\nHigh-demand from end users across CCB customer-facing and analytics teams\n\n\nProblem statement\nTransaction data source: MANTAS is a Global Financial Crimes Compliance data base not meant to support current use cases\nData latency: batch file data loads result in 11-15 day processing delay\nEngineering capacity shortage: one (1) engineer supporting platform modernization work\n\n\nRecommended solution\nInvest in engineering capacity: add one (1) engineering team with AWS and big data experience\nReduce processing latency: modernize tech stack with AWS services\nIngest new data source: Utilize Business Reservoir Intelligence Engine for real-time transaction data to improve data scalability and agility\n\n\n\nCOUNTERPARTY ENRICHMENT\n\nRevenue Generation\nExpense Reduction\nEnhanced Controls\nConsumer Bank\n \n \n \nWealth Management\n \n \n \nConnect Commerce (Payments)\n \n \n \nHome Lending\n \n \n \nAuto\n \n \n \nUniversal Search\n\n \n\n\n\n\n\n\nBENEFITS TO BUSINESS\nItem\nCurrent FTE (4)\nAllocation\nTimeline\n\n\n\n\n\nQ2\nQ3\nQ4 \nLate 2024\nData Center Migration\n2.0\n\n\n\n\nRTE support\n1.0\n\n\n\n\nCPT Modernization\n1.0\n\n\n\n\n\n\n\nCURRENT ROADMAP & ENGINEERING ALLOCATION\n\nTeam is currently over allocated\n44\n\n45\n\nMerchant Transaction Enrichment | Feature Overview\n\nMerchant Tagging \nMerchant Tagging \nDescription\nCleans merchant names and enriches individual transactions with an 80-85% coverage rate\nStatus\nCredit Card: v3.0 Live\n\nDebit Card: v2.0 live; v3.0 Q1 2024\nTransaction Types\nPoint-of-sale credit and debit card\nEnrichments\nClean Merchant Name\nMerchant details:\ncontact information, \ngeolocation, \nintermediary info, and \ncorporate parent\nWhere it’s used\nTransaction history & details in: Web, Mobile, Servicing, Toggle\nPay-in-4\nPersonalization & Insights recommendations and eligibility\nAnalytics: Branch Reassignment, Commercial Bank consumer spend\n46\n\nMerchant Tagging is already being leveraged to generate value in a variety of use cases across CCB D&A\n\n\nMerchant Tagging\nFAST’s Matching as a Service project used the merchant roll-up offered by Merchant Tagging data for the ‘Merchant Linking to BB Accounts’ use case to identify merchants that are not already in ‘Chase Merchant Services (CMS)’ platform and provide customer insights to those newly matched merchants by leveraging the customer transaction data.\nConsumer Bank’s branch reassignment project is leveraging Merchant Tagging’s enhanced geolocation data to provide more granular and accurate details on a customer’s preferred branch based of their card-present transaction history\nStandardizes point-of-sale (credit and debit card) merchant information to enable value-add features for Chase customers and merchant-level insights for the business\n\n\n\nThe Commercial Bank leverages Merchant Tagging to track consumer spend as part of standard prospecting and client management activities. Through the ease of aggregation offered by Merchant Tagging, wholesale bankers are better able to understanding how a client/prospect\'s retail sales are trending month-over-month\n\n\n\n\n47\n\nCounterparty Transaction Enrichment | Feature Overview\nCounterparty Tagging\nCounter Party Tagging\nDescription\nEnriches non-card DDA transactions with standardized party and financial institution names and insights into transaction purpose. Coverage rate of 44% (name) and 70% (purpose).\nStatus\nMVP live; \nv1.0; targeted for Q4 2024\nTransaction Types\nACH, Wires, and Real-time payments\nEnrichments\nClean counterparty name\nCounterparty details (including financial institution)\nTransaction purpose\nWhere it’s used\nAnalytics: data source for select dashboards and ad hoc analysis \n\n48\n\nCounterparty Transaction Enrichment | Process overview\n\n\n\n\nICDW\n\n\nTransaction Purpose \nTable\n\nTransaction Entity Table\nEnrichment Process\n\n\n\n\n\n\n49\n\nCounterparty Transaction Enrichment | Process overview\n\n\n\nICDW\n\n\nTransaction Purpose \nTable\n\nTransaction Entity Table\nEnrichment Process\n\n\n\n\n\n\n50\n\nCounterparty Transaction Enrichment | Feature Overview\nCounterparty Tagging\nCounter Party Tagging\nDescription\nEnriches non-card DDA transactions with standardized party and financial institution names and insights into transaction purpose. Coverage rate of 44% (name) and 70% (purpose).\nStatus\nMVP live; \nv1.0; targeted for Q4 2024\nTransaction Types\nACH, Wires, and Real-time payments\nEnrichments\nClean counterparty name\nCounterparty details (including financial institution)\nTransaction purpose\nWhere it’s used\nAnalytics: data source for select dashboards and ad hoc analysis \n\n53\n\nAdditional Counterparty enrichment examples\nMe-to-Me Identification\nMe-to-me flagging identifies transactions where customers move money between their own accounts (ex: checking to brokerage account)\nPurpose Enrichment\nPurpose enrichment provides purpose and context for a given transaction\nInbound Check Identification\nCounterparty tagging provides name identification using historical ACH / Wire / RTP transactions\nSender: Allie Bernand \nReceiver: Carrington Mortgage \nDescription: Autopay\nMortgage Payment\nSender: Martinez Carlos\nReceiver: Carlos R Martinez\nMe2Me: TRUE\nInbound Check\x0bAccount #:12345678\nACH / Wire / RTP History\x0bAccount #: 12345678\x0bName: MIAMI AQUATICS\nInbound Check\x0bAccount #:12345678\x0bName: MIAMI AQUATICS\nStandardized Financial Institution Names\nCounterparty tagging provides standardized institutional names for use across multiple applications\n1: HSBC\n2: HSBC Bank\n3: HSBC Bank USA, National Association\nHSBC Bank USA, National Association\n\n\n\n\n\n\n\n\n\n\n1: GROUPO MERC SERV LLC ID: 801378\n2: AQ WELL WORLD LLC\n3: INTUIT PAYM SOLU LLC\nGroupon\nAqua Wellness World\nIntuit Payment Solutions\nName Enrichment\nName enrichment provides clean, standardized, colloquial names for both sender and receiver\n55\n\nConceptual Architecture Diagram for Data-as-a-Service Model\n\n\n57\n\nWhat is Merchant Transaction Data Enrichment?\nNew Data Model\n Increased flexibility with handling different roles \nmerchants have within a transaction.\nData Consumption\nUsers can consume enriched merchant transaction \ndata through events and APIs.\nData Literacy\nEnriched merchant transaction data enables Chase customers\n and internal teams to easily interpret merchant transaction data.\nReduced Latency\nUtilizing data streaming in near real-time to enrich \ntransactions earlier at the time of authorization (pending) \ninstead of waiting for a transaction to settle (post).\nMerchant Logos\nMerchant logos are displayed alongside \nenriched merchant transaction data.\nFinancial Literacy\nEnriched merchant transaction data helps customers \nmake smarter decisions with their money.\nThe newest version of the DUDE team’s product also known as “MT3.0” is simply enriched merchant transaction data. This data enrichment enhances the customer experience by providing clean merchant transaction information from credit and debit card point of sale transactions to both Chase customers and internal teams for analytics insights.\n58\n\nMerchant Transaction Data Enrichment 3.0 in Action\nMerchant Transaction Enrichment 3.0 can be found on the Enhanced Transaction Detail experience for credit card point of sale transactions. Debit card enrichment is in development and is expected to be in production in 2024.\n\n\nLogo\nPhone Number\nGeolocation\nAddress\nName\nRaw transaction data frequently presents unnecessary characters making it difficult to understand.\nRaw Data\nEnriched Merchant Transaction Data Points\nhelp customer’s recognize transactions and dispute incorrect one’s directly with the merchant.\n\n\n\n\n\n\n\n59\n\nUp to four merchant entities can be tagged on a transaction\nNew merchant “role” codes identify the role of each entity, e.g.:\nDigital wallet\nOrdering / delivery app\nProduct or service\nConsuming products can and should feel free to display whichever roles\n are most applicable to the customer needs addressed by their product.\nThe “Merchant Tagging 3.0” data model will provide more flexible merchant information\n60\n\nThe handling of some roles will change from existing behavior\n“Raw” data\nMerchant Tagging 2.0\n\nMerchant Tagging 3.0\n\n\nMerchant DBA name\nLevel 1 name\nLevel 2 name\nMerchant 1\nRole\nOther roles\nRole\nEnriched Response\nObject\n\nPPADOBE INC\nAdobe\nPayPal - Adobe\nAdobe\n1 (Company)\nETUDE Merchant ID: T245\nPayPal\n4 (Digital wallet)\nETUDE Merchant ID: N12\n\nCHIPOTLE #24601\nChipotle\nChipotle\nChipotle\n2 (Location)\nETUDE Merchant ID: V98765432\n\nLocation record contains geolocation data when available\nChipotle\n1 (Company)\nETUDE Merchant ID: T555\n\nCompany record allows for easy aggregation across all Chipotle records\n{"enrchmv":[{"etudeMerchantId":"V98765432\n","merchantRoleTypeCode":"3","enrichmentRuleId":"936","merchantSequenceNumber":9},{"etudeMerchantId":"T555","merchantRoleTypeCode":"14","enrichmentRuleId":"932","merchantSequenceNumber":15}]}\n\nDOORDASH CHIPOTLE\nChipotle\nDoorDash - Chipotle\nDoorDash\n5 (Ordering / delivery app)\nETUDE Merchant ID: N45\nChipotle\n13 (Ordering / delivery seller)\nETUDE Merchant ID: T555\n\nNote that ETUDE Merchant ID is the same as above\n\nPPDOORDASH CHIPOTLE\nChipotle\nDoorDash - Chipotle\n\nDoorDash\n5 (Ordering / delivery app)\nETUDE Merchant ID: N45\n\nChipotle\n13 (Ordering / delivery seller)\nETUDE Merchant ID: T555\n\nPayPal\n4 (Digital wallet)\nETUDE Merchant ID: N12\n\nPrime Video*JJ24601VH\nAmazon.com\nAmazon Prime Video\nPrime Video\n3 (Product or service)\nETUDE Merchant ID: T192837\nAmazon.com\n1 (Company)\nETUDE Merchant ID: T5\n\n{"enrchmv":[{"etudeMerchantId":"T000556","merchantRoleTypeCode":"3","enrichmentRuleId":"936","merchantSequenceNumber":9},{"etudeMerchantId":"T63","merchantRoleTypeCode":"1","enrichmentRuleId":"936","merchantSequenceNumber":11}]}\n61\n\nMerchant roles\nRole ID\nRole description\nExamples\n9\nBuy Now Pay Later\nKlarna, Affirm\n8\nPeer-to-Peer service\nFacebook Pay, Venmo\n5\nOrdering / delivery app\nGrubhub, DoorDash\n6\nOnline marketplace\nEtsy, eBay\n7\nTravel booking platform\nPriceline, AAA\n20\nDelivery fulfiller\nGrubhub\n13\nOrdering / delivery seller\nChipotle, McDonald’s\n15\nTravel booking seller\nHotel Claude\n18\nMarketplace seller\nJamie\'s Diamonds\n17\nP2P payee\nJamie Dimon\n19\nBuy Now Pay Later merchant\nMissoma\n3\nProduct or Service\nAmazon Kindle, Apple retail store\n2\nLocation\nStarbucks, CVS\n1\nCompany\nNetflix, Costco, Walgreens\n16\nPass-through wallet\nApple Pay, Google Pay\n11\nPayment processor\nStripe, Square\n4\nDigital wallet\nPayPal, Amazon Pay\n62\n\nMerchant tagging 3.0 provides enriched merchant information from credit and debit card point of sale transactions for both Chase customers and insights for internal teams\nWhat\'s new?\nLower latency:\xa0streaming data rather than batch processing for both pending and posted transactions.\nData Consumption:\xa0data is now consumed through events and APIs.\nNew Data Model:\xa0more flexible handling of merchant hierarchy and intermediaries through transaction-level and merchant level outputs.\nMerchant logos: merchant logos will be displayed alongside enriched transactions.\nNew data model provides more flexible merchant information including:\nUp to four merchant entities can be tagged on a transaction\nNew merchant role codes for each entity → digital wallet, ordering/delivery app, product or service.\xa0\nThe "merchant of record" will be the first merchant entity provided on a transaction.\n65\n\nThe evolution of computing power and storage has created an opportunity\n\nNote: Salimbeni, Renzo. (2016). SKILLMAN Sector Skills Alliance for Advanced Manufacturing in the Transport Sector.\n70\n\nDigital Utilities Data Enrichment | Product Vision\nTransform unstructured data into usable information for:\ncustomer insights, \nproduct experiences, and \nbusiness decisions\n\n71\n\n1\nDUDE\n1\n2\nCounter Party Tagging\n14\n3\nMerchant Tagging\n26\n4\nArchived Slides\n73\n\nThe takeaway…\nDUDE: Transforming Data Into Information\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n67\n\nCounterparty Transaction Enrichment | Process overview\n\n\n\nICDW\n\n\nTransaction Purpose \nTable\n\nTransaction Entity Table\nEnrichment Process\n\n\n\n\n\n\n73\n\nBuilt a cloud infrastructure for our vision and upstream partnerships. The foundation will allow us to scale quickly across a variety of use cases (incl. PBB)\n\n2023 & Q1-24\nQ2-24 & Q3-24\nH2 2024\n\n\nGetting started\nIntegrate with Data Source\nBuild Enrichment Logic & Deliver\nCPT 1.0 POC launched in Q3’23 \nDefined schema & taxonomy standards with key stakeholders\nAligned on Key Use Case\nIntegrate with upstream data sources\n ACH Data via BRIE\nCustomer Reference Data\nThird Party Reference Data (Accuity)\nSet up AWS Infrastructure and environment\n\n\nBuild enrichment logic\nDeliver enriched data to the Data Lake\nIntegrate with ETU for Operational use Case\n\n\n\n\nDraft\n18\n\nToday…\n\n…we are covering the “Why”\n…not the “How”\n72\n\nDigital Utilities Data Enrichment\nTransform unstructured data into usable information for:\ncustomer insights, \nproduct experiences, and \nbusiness decisions\n“Enrich once, use everywhere.”\n\n20\n\nKey Assumptions: Fully staffed team with relevant skillset in place by June & enrichment logic is delivered by June\n\n\n\n\n\nOct\nSep\nAug\nJul\nJun\nMay\nApr\nMar\nFeb\n\nNov\nDec\n\n\n\nSetup\nImplementation\nOptimization\nInfrastructure\nSetup\nStorage & Compute\nApp Containerization & Event Orchestration (Lambda, Step functions)\nContinuous Maintenance & Security\xa0Updates\nData Consumption\nBRIE Onboarding (Approval/Authentication)\nIntegration into Data Pipeline (incl. scheduling)\nEnrichment\nPre-Processing\nEnrichment Implementation\nContinuous Improvement (Optimization)\nData Delivery & Distribution\nCompliance (incl. PTX Governance, Audit, Cyber Security)\nPipeline\xa0& Historical load\nUAT (Rigorous Testing)\nPROD Deployment\nTesting and Optimization\nMonitoring Dashboard (incl. Cloud Watch, Splunk)\nMeasure Accuracy & Coverage\nContinuous Improvement Monitoring\nCPT 1.0 Fully Staffed Roadmap Scenario (for discussion purposes)\n\n\n\nPhase\nSetup & Training: Cloud technology is new to most JPM Engineers but with less than a year of experience, we now have the expertise to setup an end-to-end architecture within a complex IT infrastructure .1\nOnboarding: Onboarding required months of architecture planning and several layers of approval (incl. App/Info owner, DUC) but we’ve built a strong foundation/partnership with our data providers enabling us to access sensitive transaction data for our enrichment. 2\n\nHighlights\nPre-Processing: Our new system of record has hundreds more attributes in a semi-structured schema with limited data dictionary\nCompliance: Combination of regulatory requirements, security concerns, audit processes, and need for continuous updates to protect against growing cyber attacks makes it long/arduous to become a Data Lake Publisher3\nEnrichment: Limited insight into enrichment methods makes it difficult to asses feasibility (e.g. LLM)\nPipeline: Building a firm approved data pipeline not only requires governance, but also requires additional training and engineering team engineer training on firm approved technology (i.e. JADE DPL Framework)\n\nChallenges\n\nA\nA\nB\nB\nC\nD\nC\nD\nE\nE\nF\nF\nJan\nFeb\n17\n\nCounterparty Tagging 1.0\n\n36\n\nSITUATION\nWhat is the situation?\n\nFRB was acquired by JPM on May\'23 with final integration set for May\'24.\xa0\n\nInitial plans were to integrate FRB transactions into the existing BAU\xa0process to meet timeline.\nCOMPLICATION\nWhat is the problem?\n\nRecent upstream design changes altered file formats, naming conventions, and increased file counts (from 15 to 16).\n\nThis is incompatible with our current systems and introduces a number of risks.\nRESOLUTION\nHow can we resolve this?\n\nForego enrichment of historical FRB transactions.\xa0\n\nSignificant development costs/risks outweigh marginal value of historical transaction enrichment\nDUDE has determined the best path forward for the FRB historical load is to leave the FRB-enriched transactions as is.\n66\n\nWe have an amazing opportunity to affect great change for Chase!\n\n\nDude the Fish\n69\n\nA Product Vision is…\n\n…anchored in improving the customer experience\n\n…forward thinking\n…inspirational and motivating\n\n\n\n31\n\n23\n\nOur Design Principles Follow Command Query Responsibility Segregation (CQRS) Pattern\nUtilities DOs\nProvide fast read-only access to aggregated as well as non-aggregated data across products (e.g., balances across a customer’s set of accounts)\xa0\nHave enrichment services to enrich data it gets from the SORs with additional data that does not exist in a SOR (e.g., enriching transaction data with merchant details to support transaction search)\nPerform data reconciliation and quality functions to ensure data is available for consumption not only in a timely fashion, but also to meet strict requirements on accuracy\nLeverage APIs for consumers to get data(e.g., the channels, agent desktop, banker tablet), or publish events (e.g., P&I for transactions so they can determine eligibility for Pay-in-4)\nRepresent a set of FACTS\t\nAPIs exposed to consumers are for INQUIRY only\nUtilities DON’Ts\nMust not include any product-specific, channel-specific or business process-specific logic, and they do not include any use case customizations\nDo not offer any consumer-specific interfaces – all consumers consume data the same way, using the same APIs\nNot consumer-specific, they do not address one consumer’s need; the data in a utility is needed by multiple consumers\nNot Data marts, nor data lake, no reporting; there is limited history in utilities\nDo not make predictions or inferences\nNot systems of record; do not take any COMMAND traffic\n\nMajor Business\xa0\nInitiatives\n\nTech\xa0\nModernization\n\nDriving Adoption\n\nDigital Utilities Data Enrichment Roadmap\nQ2’24\nQ3’24\nDUDE | Milestones\nEnrich FRB Transactions (May)\nLoad MT 3.0 Data to Data Lake (Jun)\nProvision UAT Environment (May)\nProvision PROD Environment (Aug)\nCreate Retagging Topic/API\xa0(May)\nDeploy Debit Card Publisher (May)\nImplement PBB Enrichment (Jun)\nGoing Primary with Card: New Card Cluster Retag\nDUDE-4668\nGoing Primary with Connected Commerce: CPT 1.0 Enrichment Engine\nDUDE-4616\nPublic Cloud AWS: MT 3.0 Card Data to Data Lake\nDUDE-4672\nPublic Cloud AWS: CPT 1.0 Infrastructure\nDUDE-4617\nMT 3.0 Debit Card Publisher \nDUDE-4345\n2\n3\n5\n6\n7\n4\nMT Debit FRB Integration\nDUDE-4685\n1\n19'

${JIRA Description}

${Terms Bullets}

${Notes}