A customer starts paying a cloud infrastructure provider every month. On its own, that tells a bank very little. It could be a weekend side project or a work expense they’ll claim back later.
Now add two more facts: the same customer is paying for business software, and there’s no revenue coming in and no company formed yet. Read together, those three facts describe someone getting ready to start a business. It’s a moment any bank would want to be part of, and it shows up in the account well before the customer walks into a branch asking for a business account. The next best action writes itself: be the first bank they meet, with a business account ready to trade on day one.
This is the idea behind customer signals at Bud. A single transaction is a data point. A signal is a small set of indicators that, read together, point to one clear interpretation, and every signal is paired with a next best action, the specific thing the bank should do about it.
We’ve written before about the kinds of insight that live in transaction data. Here we go a level deeper, into five more of the signals we’ve built and the reasoning behind each one.
Five signals, and what to do about them
Fit for HELOC: Debt sitting in the wrong place
What Bud reads: The customer holds a mortgage, is carrying unsecured debt, and their credit utilization is high.
Any one of these is ordinary. Most homeowners have a credit card, and plenty run a balance. Together they describe a homeowner paying credit card rates on borrowing that could sit against the equity in their home at a much lower cost.
Next best action: Offer a home equity line now, before a competitor gets there first.
The customer’s monthly cost goes down, and the bank moves an unsecured exposure onto a secured product. The timing matters because utilization tends to climb before anyone notices, and a CRM that records product holdings will show a mortgage and a card without ever showing the balance creeping up month on month.
Loan stacking: Borrowing spread across too many places
What Bud reads: Several loans and buy now, pay later plans being repaid at the same time, including repayments to non-bank lenders.
Each of those repayments leaves the customer’s account as an outgoing payment, so the bank can see lenders it has no relationship with. Individually they look like routine bills. Lined up together, they show a customer juggling different rates and different due dates every month.
Next best action: Offer a consolidation loan at a better rate, one that costs the customer less overall.
The customer gets one payment and a lower rate. The bank brings borrowing back in-house that was going to other lenders, and gets a clearer view of the customer’s real commitments while it does.
Not their main bank: The signal built from what’s missing
What Bud reads: No rent, utilities or grocery spend on the account over 90 days and no bills identified.
This one is worth pausing on, because nothing in the account looks wrong. It’s open, it’s probably funded, and on a standard activity report it counts as active. What gives it away is the spending that should be there but isn’t. Everyone pays for food and housing somewhere, so when none of it runs through this account, the customer’s financial life is happening at another bank.
Next best action: Make a switch offer, such as a direct deposit bonus, help moving bill payments across, or rewards on everyday essentials.
Deposits are the cheapest funding a bank has, and primary-bank status decides where they land. A customer who already holds the account is far closer to switching than a cold prospect, and this signal finds them.
Avoidable overdraft: A timing problem that looks like a spending problem
What Bud reads: Frequent overdraft and NSF activity, a high total in fees paid, and a pattern that recurs in the same week of every month, just before payday.
The repetition is what makes this signal useful. A customer who overdraws at random is in a different situation from one who runs short in the same few days every month. The second customer usually has enough income overall – their bills and their paycheck just don’t line up.
Next best action: Offer a small recurring loan that bridges the customer to their next paycheck.
Overdraft fees can look healthy on a revenue line, but every month the customer learns that their bank charges them at the moment they can least afford it. A cheaper bridge removes that friction and gives them a reason to stay.
Affordability risk: When the right action protects the customer
What Bud reads: A high debt-to-income ratio, a customer who has recently started taking cash advances, and payments going to debt collections.
These three together show debt repayments and cash advances outgrowing income, and they usually appear well before a missed payment shows up in the bank’s own records.
Next best action: Reach out with a reduced limit, a payment plan, or fees waived.
This signal is a useful reminder that the next best action isn’t always a product. Sometimes the most valuable move is to step in early. A customer offered a workable plan now is more likely to recover, and the bank avoids a far more expensive collections process later.
Reading the signal is half the job
A signal only pays off if the action reaches the customer while it’s still true. A HELOC offer that arrives a quarter late lands after the card balance has already been refinanced somewhere else. A switch offer sent to a customer who has since closed the account is wasted. So Bud’s signals update as new transactions arrive, and the action goes wherever the customer next meets the bank.
In the app, that might be a personalized prompt the customer sees the next time they log in. In the branch or on a service call, the banker sees the signal and the suggested action on the customer’s record before the conversation starts. In marketing, the team can turn the signal into a live segment and run a campaign against everyone who matches it today. In other words – the signal appears in every channel.
Most banks already own a CRM that stores product holdings and contact history. It can tell you a customer has a mortgage. It can’t see that their card utilization has been climbing for three months, because that lives in transaction behavior the CRM was never built to read. Bud sits on top of the systems a bank already has and adds that missing layer. We don’t replace your CRM – instead, we deliver what it promised and could never deliver, with AI.
Underneath, every indicator comes from models built specifically for banking data. That’s how a payment to a cloud provider gets recognized as business infrastructure, and how a run of small repayments gets traced back to the lenders behind them.
More where these came from
The six signals here are a sample of just some of those that are available in the Bud platform, with everything from life events, debit and credit and share of wallet, to early warning signs and primacy at risk being just a few examples. Each one follows the same pattern of a few indicators, one interpretation and a specific next move, and each runs on the data your bank already holds today.
You can see more of our customer signals here. And if you’d like to know which of them are showing up across your customer base, we’d be happy to show you. Book a demo today.