释放GBS潜力：人工智能的变革性力量

The Strategic Imperative With advances in AI and the advent of AI Agents, a key question is being asked… Internal Transformation Barriers GBS Marketplace Challenges •Technology landscape complexityand integration challenges•Poor data qualityin disparate systems•Organizational resistanceto new operating models•Skills gapsin emerging technologies & AI capabilities •Legacyshared services models hittingscalability challenges•Rising labor costsin traditional offshore locations•Talent retention challengesin repetitive roles•Increasingdemand for 24/7 real-time services “How should Ire-thinkmyGBS organizationinthe age ofAgentic AI?” Disruptive Forces Critical Questions •Agentic AI capable ofend-to-end process ownership•Automation rates are rapidly increasingacross GBS orgs•Customer expectationsfor instant, accurate responses•New entrants withAI-first operating models •Is our GBS modelfuture-proof?•What capabilities do we need tobuild vs. buy?•How do weupskill our existing workforce?•Should weoptimize our GBS foundationfirstor jump toAgentic AI? The Agentic GBS Delivery Model:Evolution of GBS Finance organizations have historically invested to reduce the cost to serve through automation and outsourcing. With AgenticAI, the next evolution ofthe Finance organization will reduce cost to serve further, and reorient AGENTIC AI DELIVERY MODEL What’snext? Perform Automated Accounting Overview ZORA AIPerform Automated Accounting Process Perform Automated Accountingleverages ML models to continuouslymonitor data quality issues andautonomously perform corrections,to reduce the time spent on manuallyintensive month end financial activities. The solutionautonomously comparesand cross references datafrommultiple sources to ensure data quality.It canobtain data from varioussourcesand reconcile it across systemswithout human intervention. Approve Corrections Maintain Data andProvide Feedback Detect DataAnomalies RecommendCorrections Generate MasterData The admin can viewandapproveproposed corrections.This step can also bedone in bulk formultiple anomalies. The PerformAutomatedAccounting dashboarddisplaysall datatransactions pulledfrom theorganization’s ERPsystem to createmaster data. Detect different typesof data anomalies inthe transactions,including errors andmissing data fields. Italso providesgraphical analysisoverview and detailsof anomalies. The corrections areupdatedin theorganization's coreERP. Real-timefeedback is providedon task updates viaAPIs. Provide details foreach anomaly andrecommendscorrections. Key Features •ML-driven engagement for masterdata governance•Autonomous detection oftransaction data anomalies•Self-learning mechanism toautomate system updates•Real-time feedback andmonitoring for timely and accurateposting of data transactions Value Delivered An average of75%reduction in reconciliation errors Reduction in the time spent on the financial close cycle by an average of2-4 days Saves time, ensures complete and accurate data Perform Invoice Management Overview ZORA AIPerform Automated Accounting Process Perform Invoice Management leveragesGen AI and ML models toautomateinvoice processing.Bypasses manualreview and data entry toautonomouslyreview and validate invoicesbyidentifying patterns, relationships, andanomalies in data sets. Invoice DataExtraction Invoice Data Analysis Invoice Validation It providesend-to-end audit trail andvisibilityinto invoice processing withembedded process controls,significantly reducing invoice processingtime and cost while increasingcompliance. Once approved, theinvoice is pushed tothe ERP system forposting and payment. ZORA AIPerformmaps valuestostandardized dataformats. For any missinginvoice data, it predictsthe best fit value basedon theorganization’shistoricaltransactions The user can easilysee and update thedata flagged forreview. Once allchanges are made,the invoice can besubmitted forapproval. All invoices go to acentral mailboxwhere ZORA AIcanextract key data andattributes Key Features •Intelligent extraction of invoicedata•Real-time data integration to viewand rectify invoice exceptions•Automated multi-level approvalworkflows•Compliant transactions using ML-models to validate invoice details Value Delivered Up tofive timesincrease in processing efficiency through embedded controls Reduction in invoice processing cost by20-30%, eliminating manual data entries Detection of duplicate payments, data entry errors, and anomalies Consider your business needs when deciding whether to build, buy, or subscribe/operate AI agents …based on various decision criteria and needs Token $/Compute Agent Reuse Customisation needs(generalist vs. specialist) Closed-loop learning Time Constraints Governance Vendor Accountability Available in-house expertise Integration Ease …based on various decision criteria and needs 10 key foundations needed for AI success Considerations acrossStrategy,People,Process,DataandTechnologywill be