The Current #3: A case study in consumer AI fintech

The Current is a bi-weekly series from NEA on the developments impacting consumer technology. Each installment examines a trend, disruption, or opportunity with consumer data. Posts are concise, informative, and always current.

***

If the speed that generative AI has penetrated domains like search, customer service, or media has felt supersonic, adoption in financial services appears more grounded. While many financial institutions have already effectively integrated predictive AI, generative presents a whole new set of gymnastics to navigate. Data privacy regulations, for instance, complicate the accessibility of financial training data. Or anti-discrimination laws in preventing what the CFPB calls ‘black-box’ models demand financial institutions be able to provide comprehensive logic behind any credit decision. Or the Investment Advisers Act which mandates compensated practitioners who advise others on securities investments must register with the SEC limits the financial use cases for unregulated entities.

Not all innovations in consumer fintech stick. Decades before mobile banking, Bank One launched Channel 2000 which enabled customers to pay bills and view their account through a home television screen (JP Morgan, 1980)

Implementation may be measured, but it’s still inevitable. Given the high regulatory barriers to enter and operate, it is worth identifying the most potential sources of value. Our consumer panel of US adults offers direction in two ways: value proposition and feasibility. As a case study on the broader application of generative AI to fintech platforms, the respondents answered questions about one universal category in specific, tax-filing (it is April after all).

Value proposition

Fintech relationships are infamously sticky. To understand what value proposition most resonates with consumers, our panel selected which hypothetical feature would most persuade them to adopt a new platform if an alternative to what they use to file taxes today offered it. As a case study, the options less represent specific product features as they do holistic value propositions behind integrating generative AI in consumer fintech, specifically:

• AI assistant to check, correct, and error-proof tax return as a proxy for precision

• AI assistant to help consumer maximize refund, minimize liability as a proxy for cost-saving

• AI assistant to simplify and explain the complexity of tax decisions like deductions, credits as a proxy for technical sophistication

• AI assistant to search, discover, and submit the necessary documents and information as a proxy for efficiency

The interest in technical sophistication and precision, representative of nearly two-thirds of the panel, runs contrary to a conventional wisdom that just focuses on the bottom line. Rather it underlines exciting applications of one of generative AI’s core superpowers – the ability to bridge the gap between user-friendly natural language interfaces and specialized, complex problem domains.

Feasibility

Once products are effectively built and accessible in this category, I see two main barriers to consumer adoption. The first is trust; fundamentally, do consumers sufficiently trust AI competence with their financial livelihoods?

It depends, at least for our panel in evaluating five probing examples:

My interpretations are:

• Trust seemingly erodes when the stakes increase

• Precision and execution earn more confidence than judgment

The second barrier is privacy. Personal finance is, after all, personal. Our respondents weighed what data they would consent to sharing acknowledging access enhances the utility, product experience, and personalization of a consumer fintech AI platform – its ability to maximize a tax refund, tailor financial advice, be aware of potential credits or write-offs, even adjust the conversational tone to fit a user’s preference.

Panel openness (or lack thereof) demonstrates a difficulty in unlocking deeper levels of personalization. Consumers appear moderately open to sharing core financial data with understandable product outcomes; seemingly ancillary data that could inform deeper generative recommendations, behavior, or product experience will require more hand holding and trust building to access.

Reach out to hworland@nea.com to continue the conversation.