The AIMomLab

I've spent 11 years building the fraud and credit decisioning systems banks use to score you in milliseconds. I know exactly how a 30-day-late mark gets coded into your file and how long it sits there wrecking your score. And I still missed a credit card payment last spring because the due date snuck up on me during a work trip. That's the moment I stopped winging it and actually built a real autopay system. Took me 15 minutes with an AI budgeting assistant doing the grunt work. Why one missed payment does more damage than you think Payment history is 35% of your FICO score, the single biggest factor. Not utilization, not length of history, payment history. And here's the part most people don't know: creditors typically don't report a late payment to the bureaus until you're 30 days past due. So you have a grace window before it hits your file. But once it's 30 days late, it can sit on your credit report for seven years. Seven years, for one bad week wher...

Everyone keeps asking me which apps I use to run my side hustle. Wrong question. I don't use apps, I use agents. There's a difference, and once you see it, you can't unsee it. An app waits for you to open it. An agent works while you're making dinner, driving carpool, or asleep. I build risk-decisioning systems for banks during the day, so I've spent 11 years watching the gap between "tool" and "system that acts on its own." I brought that same thinking home, and it's the only reason I have a side hustle at all with two kids under 8. Why "3 apps" thinking was costing me hours My old setup was a graveyard of tabs. A writing app, a scheduling app, a research app. Every task needed me as the middleman, copying output from one place and pasting it into another. That's not automation, that's data entry with extra steps. The shift happened when I stopped asking "what app does X" and started asking "what can ac...

I spent 11 years building the decisioning systems banks use to approve or deny credit. So when my sister-in-law told me she was closing her oldest credit card to "clean up her file," I nearly spit out my coffee. Let's fix a few of these myths before they cost you real money. Myth 1: Checking your own credit score hurts it This one refuses to die. Here's the actual mechanic: there are two types of inquiries, soft and hard. When you check your own score through your bank app or a monitoring service, that's a soft inquiry. It never touches your FICO or VantageScore calculation. Hard inquiries only happen when a lender pulls your file because you applied for actual credit, a mortgage, a car loan, a new card. The scoring models literally can't tell the difference between you checking your own score and, say, your mom checking it for you. They just don't count soft pulls. Check your score weekly if you want. It changes nothing. Myth 2: Closing old, unused ...

My kid asked me what I do for work and I said "I build robots that talk to other robots to decide things." She was satisfied with that answer. Most whitepapers I read on multi-agent systems could learn something from a six-year-old's tolerance for plain language. I have spent 11 years building the AI and risk-decisioning systems banks use to approve or deny credit and catch fraud before it drains someone's account. Somewhere in the last few years, everything I build got relabeled "agentic." Same architecture, fancier name. So let me tell you what a multi-agent system actually is, using the stuff I have actually shipped, not the stuff in a vendor deck. Forget the Definition, Here's the Mental Model A multi-agent system is a group of narrow, specialized programs that each do one job well, hand off their output to each other, and someone or something coordinates the handoffs. That's it. No soul, no consciousness, no robot uprising. Think of it li...

I've spent 11 years building the decisioning models that banks use to score you for credit cards and flag your transactions for fraud. So when a budgeting app tells me its AI "learns your spending patterns," I don't hear marketing. I hear a categorization model probably running on merchant category codes and a logistic regression bolted onto a chatbot. I wanted to know which ones actually work once the pitch deck ends. So I linked three apps to my real checking account, one credit card, and a joint account with my husband, and I lived with them for a month. The setup: same data, three different brains All three apps pulled from Plaid, which means they're all working from the same raw feed most fintechs use, ACH transactions, card auth and settlement data, and account balances refreshed a few times a day, not in real time no matter what the app claims. The difference isn't the data. It's what each app's model does with it. I ran Copilot, Monarch, an...

I spent 11 years building the models banks use to decide if you get approved or declined. When my sister-in-law asked me last month how to bump her score before a mortgage application, I didn't send her some blog post with 47 vague tips. I told her exactly what to do, and it moved her score 41 points in about three weeks. The trick isn't glamorous and is not a loophole. It's understanding how FICO actually calculates utilization, and then exploiting the timing of your statement date instead of your due date. Almost nobody does this because almost nobody knows the difference. Your due date is a lie, your statement date is the truth Here's the thing most people get wrong. They think paying their credit card bill by the due date is what matters for their score. It's not. What matters is the balance sitting on your account when the issuer reports to the bureaus, which happens on your statement closing date, not your payment due date. Those are usually 20-25 days a...

My kid asked me last week why the "computer" said no to her lemonade stand loan application (long story, don't ask, it involved a play bank app on my old tablet). I gave her a real answer because that's the only kind I know how to give. Then I realized I've never written that answer down for adults either. So here it is. I've spent 11 years building the risk-decisioning systems banks use to approve or deny credit and flag fraud. I've sat in the room when a model got shipped that quietly started rejecting a group of applicants it shouldn't have. I've also sat in the room when a smarter agent caught fraud that would have cost real money. Let me walk you through what's actually happening when an AI agent looks at your loan application. It's Not One Decision, It's a Relay Race People imagine a single AI reading your application and going thumbs up or thumbs down. That's not how it works, not at any bank I've seen. It's a pi...