Personalisation and AI in Financial Services

In an era where Netflix knows what you'll watch next and Amazon predicts your purchases, financial services personalisation is still playing catch-up. This post discusses how to think about personalisation and capture opportunities with Gen AI.

Companies have to think about personalisation throughout the product lifecycle - activation, retention and up-sell. Build a foundation by leveraging existing customer data, and then external data before capitalising on Gen AI.

Leverage customer data to solve pain points.

Financial personalisation isn't just about showing different products to different users – it's about solving real customer pain points like building financial wealth. Companies like Credit Karma, Revolut, Monzo and Qapital have laid the groundwork through features such as:

• Scores: The promise of leveraging scores such as credit scores to help users make positive behaviour changes helps companies activate, retain and up-sell. For example, Credit Karma provides personalised product recommendations to help consumers improve their credit scores, primarily to activate and retain.

• Gamification of Rewards: Gamification can work across the user journey. Fintechs are starting to leverage social aspects of people’s financial lives. US-based fintech for example, Qapital helps users save with partners or friends through joint savings goals and has helped 1.8 million customers save $1bn. Companies are helping consumers with spending and savings goals through automation of credit card limits, regular savings and setting up virtual cards for different types of spending. This kind of automation can be gamified.
  

• Pricing: Segmentation of pricing pre-dates the emergence of fintechs. Credit card and insurance companies have used data to charge customers different prices. Software allows companies to move pricing from segmentation to personalisation. Insurtechs such as Marshmallow leverage technology to track driving behaviour through a score with the reward of better insurance premiums.

"Gen AI's real power lies in reducing friction in financial decisions. From automated mortgage applications to intelligent product recommendations, we're moving toward a future where financial services adapt to users, not the other way around."

Sadaf Zahid, Product Consultant @ Credit Karma

Automation can go further with gamification. Companies such as Monzo are leveraging ‘lucky draw’ to incentivise consumers to automate regular deposits.

First leverage internal and external structured data before leveraging unstructured data

Fintechs are becoming increasingly better at leveraging internal data sources, but untapped opportunities with external data exist. The combination of the two can fine-tune LLMs.

Product experiences discussed above primarily leverage internal sources such as behavioural and transaction data, but companies have yet to fully use external data. Companies have to show benefits to customers of sharing their data and expand sources of data to fine-tune LLMs to take advantage of Gen AI.

• Show users the benefits of sharing data: Unless companies get personalisation right with existing customer data, they won’t see a lot of traction in getting customers to give them other data such as through Open Banking. For example, companies are now using Open Banking data to connect to accounts to set up recurring payments.

• Expanding sources of financial data: Currently, fintechs tend to focus on users’ bank information to create a complete picture of the financial information. However, they are often missing wealth information such as investment accounts, whereas companies like Plaid provide Open Banking API which can help access brokerage information.

• Data stack: Gen AI can pull together various sources of structured and unstructured data such as interactions with support and transaction data. Large companies will have to assess how various data and tech stacks work together to realise the potential of Gen AI. Smaller companies will have to figure out how to get customer data to fine-tune LLMs to enrich Gen AI products.

Introduce native Gen AI products or AI layer within existing products?

Introducing AI experiences within existing products versus building native Gen AI products is a complex decision, especially for incumbents.

Gen AI is expensive. Think about the expensive API calls to Open AI and building your LLMs takes time. The value of Gen AI is still in the experimental stage. Hence, building the right product and experience layer is an important decision for product teams. Product managers need to answer the following questions, before jumping straight in:

• Use Case: Identify a clear use case by asking the usual business and customer outcome questions. Will Gen AI help fintech users find the best financial product in a marketplace business or file a mortgage application on a user’s behalf? The first use case may be easier to implement in an existing marketplace, but for the latter use case, building a new product does not complement the existing product experience.

• Execution: Once a use case is clear, the solution must solve a problem in a meaningful way. For example, a lot of SAAS companies have added an AI layer into an existing product, but only superficially solves the user’s problem. For example:

  • Notion’s AI summarises a user’s text, which may not be the most important use case to maintain a user’s notion page.

  • In my personal experience, Perplexity has replaced half of my Google search queries to find the most relevant content and follow-up questions. Google’s AI-generated responses on top of search results are helpful but do not offer Gen AI’s experience to ask follow-up questions. Primarily using Gen AI for search is a business model and technology shift for Google.

• Time to market: Adding an AI layer to an existing product is faster, but the potential to innovate may be much lower.

• Operational Impact: Large institutions may have to streamline operations for Gen AI native personalisation. For example, AI agents may be needed to streamline complicated processes such as mortgages before recommending them through Gen AI.

• Accuracy and hallucinations: Financial companies of any size cannot afford hallucinations to avoid customer mistrust or regulatory scrutiny.

The Future of Financial Services

Gen AI's real power lies in reducing friction in financial decisions. From automated mortgage applications to intelligent product recommendations, we're moving toward a future where financial services adapt to users, not the other way around.

What's your take on Gen AI in financial services? Are you seeing other approaches to implementing AI in financial products? Let's connect and discuss the future of fintech personalisation.

#FinTech #GenerativeAI #ProductStrategy #Innovation #FinancialServices #ProductLeadership