Latest / Fintech Conversations with Fexingo: Banking, Payments, and Financial Software Companies / How Fintechs Use AI for Real-Time Subscription Management
Transcript
- Lucas: If these conversations are useful for what you're building or running, we're glad you're here. Today we're talking about something that quietly drains billions of dollars from consumer and business accounts every year: subscription creep. Luna: Subscription creep — that slow accumulation of monthly charges you forget you signed up for. I think most of us have been there. Lucas: Right. And fintech is starting to fight back with ai powered subscription management tools. The poster child here is Truebill, which rebranded to Rocket Money after being acquired by Rocket Companies in 2021. Luna: I've actually used it. It connects to your bank accounts via Plaid, scans your transactions, and identifies recurring charges. Lucas: Exactly. The AI layer does two things. First, it categorizes each transaction with high accuracy — distinguishing a one-time Uber ride from a monthly Uber One membership. Second, it tracks price changes. If your Netflix subscription goes from $15.49 to $17.99, the app flags it. Luna: And they claim users save an average of $240 a year? That's real money. Lucas: It is. And the business model is interesting — it's freemium. The basic cancellation service is free, but they take a cut of the savings if they negotiate a lower bill for you. That aligns incentives. Luna: But this isn't just for consumers. I've seen B2B versions too — companies like Spendesk and Airbase are embedding similar logic into corporate cards. Lucas: Great point. The B2B angle is huge. Think about a marketing team that has twelve SaaS tools — Canva, HubSpot, SEMrush, whatever — and no one remembers who set them up. AI scans the corporate card statements, identifies recurring vendor charges, and flags unused licenses. Luna: It's like a digital spring cleaning for your software stack. Lucas: Exactly. And the technology behind it is getting more sophisticated. Instead of simple keyword matching, these models use natural language processing to read transaction descriptors — which are often garbled — and map them to known subscription services. Luna: Right, because the merchant name in your bank statement might say 'SQSPOTIFY'. The AI has to normalize that. Lucas: Yes. And it learns over time. If a new charge appears that looks like a subscription — regular amount, regular interval — the model flags it for the user to confirm. That's a real-time feedback loop. Luna: So we're moving from reactive — checking your statement at the end of the month — to proactive: an alert the day the charge posts. Lucas: Exactly. And that's where open banking APIs come in. The ability to read transaction data in near real time is what makes these apps possible. Plaid, Yodlee, Finicity — they provide the pipes. Luna: But there's a regulatory side too. The Federal Reserve is working on a rule that would require recurring payment providers to make it easier for consumers to cancel — the so-called 'click to cancel' rule. Lucas: That's a big deal. The proposed rule, which the Fed is expected to finalize later this year, would require that cancelling a subscription be as easy as signing up. No more having to call a customer service line. Luna: Fintechs like Rocket Money are already doing that — they'll even cancel on your behalf. But the rule would make it standard across the industry. Lucas: Right. And that changes the economics for subscription businesses. If churn becomes frictionless, companies have to compete on value, not on retention tricks. Luna: Which is ultimately better for consumers. But it also creates an opportunity for fintechs that can help businesses manage that churn — by offering usage analytics, engagement scores, that kind of thing. Lucas: Good point. So the same AI that helps consumers cancel can also help businesses retain. If a SaaS company sees that a customer hasn't logged in for 60 days, they can send a targeted offer before the next billing cycle. Luna: That's proactive retention. And it's all powered by the same transaction data. Lucas: Let's talk about the scale of the problem. According to a 2024 survey by C+R Research, the average American spends about $219 per month on subscription services — but underestimates that figure by more than $100. Luna: That's a huge gap between perception and reality. Lucas: It is. And that's just consumers. For businesses, the waste is even larger. A 2023 report from Productiv found that companies waste about 30% of their SaaS spend on unused or underused licenses. Luna: So fintechs that can plug into both personal and corporate accounts have a massive addressable market. Lucas: Absolutely. And we're seeing consolidation. Rocket Money now has over 3 million users. There's also Trim, which was acquired by OneMain Financial, and newer entrants like Scribe and Bobby. Luna: But the real innovation might be in the AI itself — moving from simple categorization to predictive analytics. Can the model tell you which subscriptions you're likely to actually use based on your past behavior? Lucas: That's exactly where it's heading. Some apps are already doing 'subscription health scores' — they analyze your usage patterns and recommend downgrading from a premium plan to a basic one if you're not using the features. Luna: That's smart. It's not just about canceling; it's about optimizing. Lucas: Right. And the AI gets better with more data. If you connect your bank accounts, credit cards, and even your email receipts, the model can build a complete picture of your recurring financial commitments. Luna: Privacy is a concern, though. Giving a third-party app read access to all your transactions is a big ask. Lucas: It is. And that's why the top apps are transparent about their data practices and use encryption. Rocket Money, for example, uses 256-bit encryption and doesn't store your bank login credentials. Luna: Still, consumers need to be careful. I'd only recommend apps that are Plaid-enabled or use similar trusted data aggregators. Lucas: Agreed. Look for apps that are SOC 2 compliant and have clear privacy policies. Luna: So what's the next frontier for AI in subscription management? Lucas: I think it's dynamic negotiation. Imagine an AI agent that can automatically negotiate your internet bill or your phone plan by comparing it to competitor offers in real time. Luna: That would be a game-changer. Some apps already do that for cable bills, but doing it across all subscriptions — that's the holy grail. Lucas: And with large language models, the AI can actually chat with customer service reps on your behalf. Early versions of this are already being tested. Luna: That feels like something out of a sci-fi movie, but I can see it happening within the next two years. Lucas: For sure. And if today's conversation was actually useful to you — whether you're a consumer trying to save money or a builder thinking about the tech — the way these episodes stay ad-free is listener support. You can find us at buy me a coffee dot com slash fexingo. Luna: Yeah, it's a small way to keep the show going. And we appreciate every bit of support. Lucas: Thanks for listening. So to wrap up: ai powered subscription management is a win-win — consumers save money, fintechs build sticky products, and businesses get a nudge to deliver real value. Luna: And the technology is only getting smarter. I think we'll look back on the days of 'set it and forget it' subscriptions as a relic of the early internet. Lucas: Absolutely. Until next time.