What happened
The Straits Times reports that Hosan Swee, founder of The AI Burrow, built KiasuMiles – a digital assistant that tells users which credit card in their wallet will earn the most frequent‑flyer miles for a given transaction. The tool’s covers roughly 50 Singapore credit cards and more than 3,300 merchants, with data on minimum spend thresholds, monthly mileage caps, and category‑specific reward rates. Swee assembled the database by prompting AI coding models (ChatGPT, Claude, and Codex) to scrape bank and merchant websites, then linked the resulting knowledge base to popular AI chatbots. After a one‑time setup that connects KiasuMiles to a user’s chatbot account, users can ask questions such as “Which card should I use for GrabFood?” and receive a recommendation plus any relevant conditions (e.g., required monthly spend). The development relied on AI‑generated code, allowing Swee to create the service without traditional programming expertise.
Swee’s motivation stemmed from the difficulty of manually tracking the varying reward schemes of multiple shared credit cards. He turned to AI coding assistants, prompting them to compile a structured of card rules, spending categories, and mileage caps.
Using prompts like “Build a structured of Singapore credit cards with accompanying rules…”, Codex scanned bank and merchant websites, extracting relevant data in minutes rather than the hours required for manual research.
The resulting database feeds into a chatbot interface. After a one‑time connection of KiasuMiles to a user’s ChatGPT, Claude, or Codex account, the assistant can answer real‑time queries about the optimal card for a specific merchant or payment method, also surfacing any qualifying conditions.
Swee tested the recommendation logic with a limited set of cards and merchants before scaling the system. He emphasizes that while free AI tools suffice for prototyping, paid subscriptions provide faster, more capable models for larger data workloads.
Source details: straitstimes.com ↗
Why it matters
KiasuMiles illustrates how recent advances in AI coding assistants can lower the barrier for non‑programmers to build functional, data‑driven applications. By automating the collection and structuring of fragmented credit‑card reward information, the tool reduces the time and effort consumers spend optimizing their spending, potentially unlocking significant monetary value in the form of saved travel miles. The approach also demonstrates a practical use case for AI‑generated code in personal finance, a sector where data is often scattered across multiple sources and updates frequently. If similar tools proliferate, they could reshape how consumers interact with financial products, prompting banks to reconsider the transparency and accessibility of reward structures.
The case shows that AI coding tools can democratize software creation, enabling individuals without formal programming training to launch functional services.
By aggregating fragmented reward information, KiasuMiles can help users capture otherwise missed mileage, translating into tangible travel savings.
The tool’s reliance on AI‑generated code raises questions about the accuracy and timeliness of the underlying data, especially as banks frequently update reward terms.
If similar AI‑driven financial assistants become widespread, they may pressure financial institutions to standardize reward disclosures and could attract regulatory attention regarding consumer protection.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
Which description best fits "narrow AI", the kind of AI in use today?
What to watch next
Future developments to monitor include: (1) whether KiasuMiles expands beyond Singapore to cover additional markets or card issuers; (2) the sustainability of the AI‑generated as banks modify reward terms; (3) potential regulatory scrutiny over AI‑driven financial advice; and (4) the emergence of comparable low‑code AI tools for other consumer finance decisions, such as loan comparisons or budgeting.
Expansion plans: Whether the creator will broaden coverage to include more cards, merchants, or regional markets.
Data freshness: How the system will keep pace with frequent changes in credit‑card reward policies and whether automated re‑scraping will be reliable.
Regulatory response: Potential oversight from financial regulators concerning AI‑based advice on credit‑card usage.
Competitive landscape: Emergence of other AI‑powered low‑code tools targeting personal finance optimization, which could drive innovation or market saturation.