Isbank Deputy CEO Sezgin Lüle, whose bank won the World’s Best Mobile Banking App award, explains how LLMs and partnerships enable hyper-personalized financial services and foster a “super-app” ecosystem.
Global Finance: Could you elaborate on the most impactful recent application of AI or generative AI within İşbank?
Sezgin Lüle: Our bank’s legacy AI, including rule-based natural language processing models, handled over 108 million conversations annually with 11 million customers. Now, we’re transitioning to generative AI [GenAI], integrating large language models [LLMs] into our mobile app to assist clients. We have already deployed three agentic bots, with more to come.
We have our Investment Assistant, an LLM-powered GenAI offering investment advisory services (exchange rates, market analysis, portfolio construction) that uses internal and public data, providing a conversational experience. It’s in production, accessible to 15 million customers.
Another is our Greeting Assistant, which guides users through the application’s functionalities conversationally, addressing around 800 menu options. It also enables customers to easily find information on the bank’s website. Although it is still under development for full transaction fulfillment, it significantly improves user experience.
Finally, we have the Loan Assistant, a GenAI assistant that helps users plan and manage their overall financing needs. It presents personalized loan offers, runs payment plan simulations, and seamlessly completes the process by transferring approved funds to the customer’s account.
GF: Why did you choose these three verticals?
Lüle: Investment transactions drive most of our mobile-app traffic, thanks to rich functionality, including buying global stocks, fractional investments, and gold and silver. We wanted to make it easier for customers to make investment decisions.
For the second one, we aim to improve the user experience by moving beyond traditional interfaces toward a conversational, functional approach. You might say, “Which credit card earns miles?” The assistant retrieves the information directly from the bank’s website and guides the user to the application menu.
The final one addresses a strategic priority in Turkey: addressing the low savings rate—70% of gross national income comes from spending, not saving, leading people to rely on overdrafts, installment loans, and credit cards. This financial behavior makes the third bot crucial.
GF: What are the key strategies and initiatives driving your “super-app” development and your collaboration with startups and emerging technologies?
Lüle: Our app, with 16 million users, aims to be a personalized finance hub. We partner with startups to offer diverse nonbanking services, creating an ecosystem that benefits us (varied offerings) and startups (access to our customer base). Examples include car value assessments and home services. For seven years, our “Workup for Entrepreneurs” program has supported over 25% of Turkey’s startup ecosystem. We also have a corporate venture capital fund and run yearly incubation programs for digital, AI-driven Turkish startups, integrating them as future SMEs.
We see startups as partners, not disruptors, recognizing their use of data, AI, and cloud technology. Collaborating early allows us to learn from their agility, stay relevant, and explore new portfolios, as we acknowledge the shift in future savings.
GF: How is Isbank moving from data analytics to true hyper-personalization?
Lüle: We use customer data with legacy AI for product recommendations. GenAI, however, will enable hyper-personalization, moving beyond general AI suggestions. For instance, instead of a general card offer, GenAI can recommend a cashback credit card for dog-related purchases for dog owners. We continually conduct A/B testing and interpret millions of data points from digital interactions. Our AI-powered personal finance management tool helps customers manage their budgets. For example, it can suggest a spending limit for restaurants if the user wishes. This approach drives hyper-personalization, ensuring advice is tailored. The AI can remind customers about subscription increases, like Netflix’s annual fees, showing how transaction-based personalization adds value. This ongoing effort aims to position us as a precise bank that truly matters to our customers.
