AI in Financial Services
Photo courtesy of Mr. Tharakesavulu Vangalapat

"The convergence of artificial intelligence with traditional industries represents more than technological advancement," reflects Tharakesavulu Vangalapat, Sr. Director of Data Science at Broadridge Financial Solutions."We are witnessing a fundamental restructuring of how enterprises understand prediction, optimization, and decision-making at scale".

His observation arrives during a remarkable transformation across sectors where machine learning intersects with established business operations. The global artificial intelligence market, valued at $196.63 billion in 2023, projects expansion to $1.81 trillion by 2030, according to Grand View Research. Professionals who translate academic breakthroughs into commercial applications occupy an increasingly vital position. Vangalapat's career trajectory illustrates this evolution, spanning nearly two decades across Fortune 500 companies where he has architected systems generating measurable financial returns while advancing the theoretical foundations of applied machine learning.Vangalapat joined Broadridge Financial Solutions in June 2021, assuming responsibility for enterprise-scale AI strategy across business units serving institutional investors and asset managers. His initial project involved constructing a Global Demand Forecasting Model incorporating generative AI and agentic frameworks that predict Assets Under Management and Net Flow metrics for investment firms. The system combines time-series analysis with natural language processing of market sentiment indicators. Market forecasting in asset management requires synthesizing disparate data streams where economic indicators, regulatory changes, competitive positioning, and historical performance patterns all influence fund flows through complex interactions.The platform supports $60 million in assets under management and generates $4-5 million in annual recurring revenue during its first operational cycle, with long-term growth projections targeting $60 million. The integration of generative AI capabilities enables the platform to process unstructured market sentiment data from news sources, analyst reports, and social media, while agentic AI components autonomously adjust forecasting parameters based on changing market regimes. The broader machine learning market in financial services reached $28.94 billion in 2024 and projects expansion to $214.06 billion by 2034, representing a compound annual growth rate of 22.1%, according to Market.us research. This growth stems primarily from increasing demand for fraud detection systems, algorithmic trading platforms, risk management solutions, and personalized customer experiences, positioning Broadridge to capture significant market share within this rapidly expanding sector.Advanced Analytics Transform Corporate Governance and Compliance

Corporate proxy voting presents substantial analytical complexity for institutional investment firms. Asset managers overseeing diversified holdings face the challenge of assessing thousands of shareholder proposals each year across their portfolio companies. The process traditionally required manual review of lengthy proxy statements to assess proposals regarding executive compensation, board composition, environmental policies, and other governance matters. Vangalapat designed and developed a Customer Policy Vote Prediction Engine that combines machine learning, natural language processing, and generative AI to automate shareholder voting analysis across large-scale proxy statements.

This innovation generated over $100 million in cumulative client impact, according to performance metrics tracked since deployment. The revenue figures reflect both direct fees and the strategic value clients derive from enhanced analytical capabilities. The system processes hundreds of pages of SEC filings, extracting relevant voting guidance and predicting institutional investor preferences with measurable reliability. Generative AI transforms document analysis from a labor-intensive process into a knowledge synthesis operation. The proxy prediction platform integrates large language models with traditional machine learning classifiers, creating a hybrid architecture that leverages the strengths of each approach.

Financial institutions process enormous volumes of regulatory documents, from SEC filings to internal compliance reports. Manual extraction of structured data from these documents consumes thousands of labor hours while introducing error risks that can trigger regulatory penalties. Vangalapat led the development of an Intelligent Document Processing framework for SEC filings, including Forms DEF 14A and 10-K, leveraging generative AI and agentic AI capabilities. The system automates data extraction across hundreds of pages using advanced natural language processing, large language models, and autonomous agent workflows that orchestrate multi-step document analysis tasks.

The innovation eliminated thousands of manual processing hours, saving $400,000 to $500,000 annually while reducing human error rates over 90%. His optimization of AI pipelines and cloud architecture resulted in $25,000 monthly reductions in infrastructure costs. The document processing framework extends beyond simple optical character recognition to semantic understanding of regulatory language through generative AI models. Agentic AI workflows enable the system to autonomously navigate complex document structures, identify specific disclosure types, extract numerical data with appropriate context, and flag anomalies requiring human review. Machine learning models trained on thousands of historical SEC filings work in concert with generative AI components that interpret nuanced regulatory language and synthesize information across multiple document sections.

Patent Innovation Across Multiple Technical Domains
Vangalapat has developed a portfolio of seven granted patents addressing fundamental challenges in artificial intelligence deployment across diverse sectors. The innovations encompass intelligent lighting systems, predictive maintenance algorithms, smart device integration, and agricultural monitoring applications. Independent researchers have cited his patent work sixteen times, demonstrating substantive impact on subsequent investigations in these technical areas. Each patent tackles specific operational problems while contributing algorithmic approaches that advance the broader field of applied machine learning.

Vangalapat's transition to Signify (formerly Philips Lighting) North America in 2016 marked a decisive shift toward consumer-facing artificial intelligence applications. Leading a team of researchers and engineers, he spearheaded the development of the Interact LightPlay application, which utilized computer vision and machine learning to enable dynamic lighting control through mobile devices. The platform achieved over one billion user interactions globally, demonstrating commercial viability for AI-powered consumer products. The work generated three patents related to interactive color selection and coded light communication systems.

One patent, filed in 2021, details methods for predicting data points using similarity functions that dynamically adjust machine learning model weights. The innovation addresses a persistent challenge in applied AI where maintaining prediction accuracy becomes difficult when deployed models encounter data distributions differing from training sets. Google Scholar records indicate the patent has been cited in subsequent research examining adaptive machine learning frameworks. Computer vision techniques developed for lighting control translate to agricultural monitoring with modifications accounting for different environmental variables and optimization objectives.

One particularly influential patent addresses predictive maintenance for smart lighting systems, technology he developed during tenure at Signify from November 2016 to June 2021. The system uses anomaly detection algorithms to proactively identify faults, firmware issues, and sensor malfunctions in connected lighting and IoT ecosystems. Vangalapat navigated collaboration between Signify North America Research and MIT's Computer Science and Artificial Intelligence Laboratory successfully, producing diagnostic models that reduced system downtime by approximately 25% while minimizing false alarms that erode maintenance team confidence in automated alerts.

Advanced Infrastructure and Generative AI Integration
Vangalapat architected a comprehensive operational framework at Broadridge, establishing continuous integration and deployment pipelines, automated quality assurance systems, and performance tracking mechanisms that accelerated implementation timelines by approximately 40%. The MLOps platform addresses the complete lifecycle of machine learning systems, from initial development through production deployment and ongoing maintenance. Continuous integration pipelines automatically test model updates against validation datasets, ensuring that changes improve rather than degrade performance. Deployment automation with rollback capabilities allows rapid iteration while minimizing risks associated with production releases.

The emergence of large language models and generative AI systems after 2022 created opportunities for enterprise technology leaders to develop next-generation financial platforms. Vangalapat led the deployment of secure generative AI infrastructure at Broadridge, integrating the latest technologies, including OpenAI GPTs, Claude, AWS Bedrock, and Llama models with comprehensive audit logging, compliance frameworks, and guardrail mechanisms. His team developed cutting-edge agentic AI workflows that combine multiple specialized models, each handling specific tasks while maintaining security boundaries between components.

These agentic systems represent a shift from simple question-answering to autonomous multi-step reasoning, where AI agents can plan actions, use tools, and iterate toward complex objectives. The architecture enables sophisticated multi-step processes while ensuring sensitive data never leaves controlled environments. Enterprise generative AI differs from consumer applications primarily in operational requirements, with Vangalapat's implementation incorporating input filters, output validators, and usage monitoring to balance innovation with risk management imperatives required from institutional investors.

Beyond commercial work, Vangalapat has completed over 60 manuscript reviews for journals and conferences in artificial intelligence and machine learning. Peer review represents essential infrastructure for scientific progress, ensuring published research meets methodological standards and makes genuine contributions to collective knowledge. The IEEE International Conference on Computing, Communication, and Automation recognized his review contributions in 2025, and he was elevated to Senior Member of IEEE, a distinction that acknowledges significant professional accomplishments and technical expertise. Review assignments arrive through reputation within research communities where conference program committees and journal editors invite reviewers based on publication records, demonstrated expertise, and track records of thoughtful, constructive feedback.