Unveiling the Dance of Data and AI with Keith Schleicher

DANCING WITH THE STARS: AI & DATA (BUT NO MIRROR BALL) KEITH SCHLEICHER

The opinions expressed in this presentation are those of the author and do not reflect the opinions of his current or former employers.

AGENDA Personal Journey with Data, Analytics, ML & AI Review of Key AI/ML concepts Turning AI/ML into VALUE Data and AI as Dance Partners

BEGINNINGS Long before the Internet or mobile sports apps lived the box score Near daily consumption led to a love of statistics and real data Eventually a BS in Mathematics and a MS in Statistics

30 YEARS OF DATA IN ACTION Statistical Modeling Machine Learning Models AL/ML Use Cases Credit Risk Analysis Model Validation/ Governance Commercial Sales/Service Marketing Analytics Financial Crimes Analysis Finance/HR/Legal Model Governance Data Science Capabilities AI Governance

DISTINCTIONS BETWEEN AI & ML Artificial Intelligence (AI): Machines performing tasks requiring human-like intelligence. Artificial Intelligence (AI) Machine Learning (ML): A set of algorithms learning from data to improve tasks. Machine Learning Deep Learning: Neural networks with many layers for complex pattern recognition. Deep Learning Large Language Model (LLM): Deep learning models for understanding and generating human language. LLM

GENERATIVE AI Generative AI (GenAI) refers to algorithms that produce new data by learning underlying patterns in training data using LLMs CREATIVITY AUTOMATION VERSATILITY Produces original content (e.g. text, images and code) Reduces time and effort associated with complex tasks Disruption occurring across multiple industries

GO BEYOND THE HYPE How do I translate Data, AI, ML, etc. into meaningful value? How do I build excitement for AI in my organization? How do I deliver on the promise of AI?

VALUE STORIES DATA INSIGHTS ACTIONS OUTCOMES Successful projects in AI/ML start with a clear problem, a willing sponsor and good data.

EXAMPLE USE CASE Sales/Service representatives spend hours per day on calls with potential and current customers While calls are recorded, most calls are manually summarized by a human and entered into a system (e.g. Salesforce, ServiceCloud, etc.) AI can transcribe and summarize the calls, increasing the information captured and allowing the human to spend more time with customers

APPLYING THE FRAMEWORK DATA INSIGHTS ACTIONS OUTCOMES Call recordings and transcripts Pareto Analysis of Common Defects Process/Tech Improvements Higher Satisfaction Scores/ Usage Commonly Asked Product Questions FAQs for Key Products Fewer Inbound Calls

GENERATIVE AI RISKS Privacy Transparency Hallucinations Security Intellectual Property Computational Effort Change Management Bias

SOME WORK STILL TO DO Prompt Generate an image of an adult woman walking her small dog in a city park. The dog must be on a leash.

DATA AS A DANCE PARTNER The AI/ML tools thrive on large volumes of data (even unstructured data) Many of the innovations in LLMs arise by expanding the underlying training data While AI/ML might be getting a lot of publicity, DATA is FOUNDATIONAL for organizational success.

DATA MANAGEMENT Quality Governance Data Platforms (and the underlying processes) are critical to the AI/ML journey Lineage Security Integration Scalability

SUGGESTIONS (DATA MANAGEMENT) DATA QUALITY/ GOVERNANCE TALENT/ COLLABORATION SCALABLE INFRASTRUCTURE Ensure consistency and accuracy of internal data Hire diverse talent Continued investment in Data Lakes/ Warehouses for structured and unstructured data Invest in training and development Establish clear Data Governance policies and procedures Evaluate and pursue automation opportunities Encourage cross-functional collaboration Foster a Data-driven culture

ACKNOWLEDGEMENTS Data is Everybody s Business by Barbara Wixom, Cynthia Beath & Leslie Owens ChatGPT & MS Co-pilot (AI Generated Image example)

THANK YOU!