도구마다 데모 프롬프트 하나씩. 새 에이전트 세션에 그대로 붙여 넣어 각 도구가 무엇을 할 수 있는지 확인해 보세요.
Six prompt patterns that turn vague questions into tight answers.
솔직한 답은 제품이 아니라 조합이에요. 이미 쓰는 에이전트를 고르고, 은행 읽기 권한을 주고, 일상 언어로 돈 질문을 하면 돼요.
The most common error in agent-generated spending summaries is counting both the card swipes and the checking payment that covers them. One sentence of prompt fixes it.
Bank data labels a rent payment to your landlord the same way it labels moving money to savings: a transfer. Here's how to teach your agent which transfers count as real spending.
가맹점은 청구 디스크립터를 계속 바꾸기 때문에, 하나의 구독이 거래 내역에서 두세 개의 서로 다른 가맹점으로 나타날 수 있어요. 에이전트로 이름 변경을 잡아내는 세 가지 신호 레시피를 소개할게요.
여름엔 전기 요금이 두 배, 겨울엔 가스 요금이 세 배가 돼요. 에이전트가 거짓말을 연간화하지 않도록 하는 12개월 프롬프트 패턴이에요.
All 29 AI apps that can connect to a remote MCP server in 2026, in one list: auth method, config file, and a setup guide for each. Nobody else maintains this well, so we do.
Track monthly rent, confirm it posted, spot unexpected fees. Two prompts to start.
Monthly grocery spend, across every store, split between in-person and delivery.
Every recurring charge, ranked by cost. Find the ones you forgot about.
Income minus expenses, top income sources, top expense categories, net number.
What percent of your take-home pay is staying in accounts instead of being spent.
Overdrafts, ATM fees, interest charges, foreign transaction fees, monthly maintenance.
Direct deposits from employers, pay cadence, YTD gross, any unexpected changes.
For LLC owners and freelancers: estimated tax math, when to send, how much.
What percent of your portfolio is in each position, sector, or asset class.
Dividend income received, which positions pay, YTD dividend yield.
Monthly fuel spend, across every gas station, to track commuting costs.
Every Amazon charge, across Prime, Whole Foods, digital purchases, AWS.
The classic latte-math question, now answered with real data.
Dining-out spend, split from groceries, across in-person and delivery.
Rideshare spending, across Uber + Lyft + taxis, monthly trend.