FinovateEurope 2022 Session: “Keynote Address: How to Use Data Analytics and AI to Create Human- centric Financial Products”
Speaker: Inma Martinez, digital pioneer and AI scientist, G7, OECD & EU Digital Transformation
Findings
- Inma Martinez’s keynote speech outlined the potential of AI-enabled data analytics to create human-centric financial products. AI models can generate deep insights and reclassify customer segmentation more granularly than unaided human analysis.
- She emphasized the strengths and limitations of AI and the crucial role that humans play in using tech to personalize the customer experience.
- While AI models possess enormous computational power—far beyond human capabilities—to process complex datasets, they currently lack the cognitive power to understand the context of decision-making.
- She noted that large tech companies such as Meta are hiring philosophers and psychologists for their machine learning teams to enhance their algorithms.
- In a subsequent Q&A session moderated by Insider Intelligence Principal Analyst Eleni Digalaki, Martinez explained that chatbots currently available in customer servicing offer a frustrating user experience because they are decision-tree-based and rely on predefined rules to drive conversations.
- Martinez said that generative chatbots which produce original conversations are possible today—but no one has implemented them yet for customer service. This would be a labor-intensive undertaking with unclear ROI. It also would require a skilled workforce that would be difficult and expensive to assemble due to the current talent wars.
What this means
- Humans play a crucial role in teaching algorithms about the context in which humans operate to enable them to fully understand why they make decisions.
- Algorithms are not psychologically trained and struggle to understand the emotion behind decisions because “humans don’t live in a computational world.”
- One example that Martinez offered is of humans training computer vision algorithms during online verification processes which ask them to prove their identity by tagging various objects in images.
Our Take
- While AI-powered automation will replace inefficient processes—such as claims management in insurance—fears that AI will supplant skilled human workers are largely misplaced.
- We expect that the next paradigm of AI use cases in financial services will be centered in “human plus digital”: Applying human expertise and cognitive functions to the computational power of AI models to derive insights at scale from large, complex data sets.
- Financial employees will require upskilling and reskilling to identify potential AI use cases and—as in the chatbot example–to learn to train it effectively to create hypotheses about what’s driving human behavior, as well as understand what lies behind humans’ highly nuanced use of language.
- AI’s potential to augment the client experience is particularly relevant in wealth management: Wealth managers can arm their advisors with AI-powered solutions and insights to personalize advisor-client interactions.
- For example, Merrill Lynch’s Client Engagement Workstation centralizes client insights from millions of interactions that its advisors can use to personalize customer communications and recommend tailored financial products and investment plans.
Go deeper: Check out some of our articles about AI and personalization
