Why the generic answer isn't enough
Search "how much cash should I keep in checking" and you'll get the same advice everywhere: one to two months of expenses, plus a small buffer for timing hiccups. It's a fine rule. The problem is the variable inside it. Most people don't know their monthly expenses. They know a vibe.
So the rule stays theoretical. You keep $18,000 in checking because it feels safe, when your real burn is $6,200 and two months would be $12,400. The extra $5,600 just sits there. Not lost, but not working either.
An agent connected to your accounts can turn the rule into a number. Your number. That's the whole trick of this guide.
Three places money goes idle
Idle cash hides in predictable spots. The obvious one is a checking balance well above your monthly burn. Checking pays nothing, or close to it, so anything past your safety margin earns zero by design.
Second: savings accounts at big banks. Plenty of accounts labeled "savings" still pay 0.01% to 0.05% APY. At that rate, $50,000 earns about $5 a year. A high-yield account paying around 4% would earn roughly $2,000 on the same balance.
Third, and easiest to miss: uninvested cash inside a brokerage account. You transferred money in months ago, bought some of what you meant to buy, and the rest landed in a sweep fund. It shows up when you check your portfolio, so it feels invested. It often isn't.
Step 1: Compute your real monthly burn
Everything downstream depends on one number: what you spend in a typical month. With your bank connected through BankBridge, the agent pulls this from real transactions instead of your best guess.
What's my average monthly spending over the last six months? Exclude credit card payments and transfers between my own accounts so nothing gets double counted.
Under the hood that's get_monthly_cashflow and get_spending_summary doing the work. The exclusions matter: paying off a credit card isn't spending (the swipes were already counted), and moving money from checking to savings isn't spending at all. A six-month window smooths out the weird months, and you can ask the agent to flag outlier months before it averages.
Say the answer comes back at $6,200 a month. Two months is $12,400. Add whatever buffer lets you sleep and round to $15,000. That's your checking target. Everything above it is a candidate for somewhere better.
Step 2: Scan every account against that line
Now the comparison.
List all my accounts with current balances. Which ones hold more cash than my $15,000 checking target, and by how much?
That's a single list_accounts call, and because BankBridge fetches live, the balances are current as of right now, not whenever a cache last updated. If you've connected more than one bank, the agent sees all of them in one pass, which is where this gets useful. A $4,000 cushion in one checking account and $7,000 in another don't look like a problem individually. Together they're a five-figure pile earning nothing.
Ask for the savings check in the same breath. The agent can't always see your APY directly, but it can spot the symptom: a savings balance that hasn't grown by more than a few cents of interest in months.
Step 3: Check the brokerage for uninvested cash
If you have investment accounts connected, list_holdings shows what's in them, position by position. Cash and sweep-fund positions appear right alongside your index funds.
Look at my brokerage holdings. How much is sitting in cash or a money market sweep instead of invested, and roughly how long has it been there?
This is the check that surprises people most. Rollover IRAs are the classic case: the money moved custodians years ago, landed as cash, and nobody ever placed the buy order. It's been "in your retirement account" the whole time while earning nothing.
To be fair, some brokerage cash is intentional. Money market funds inside a brokerage can pay competitive rates. The question the agent helps you answer is whether your cash is in one of those, or in a default sweep paying a fraction of it.
Step 4: Put a yearly price on it
Numbers move people more than principles do. So make the agent price it.
Assume idle cash could earn around 4% in a high-yield savings account or money market fund. For each pile of idle cash you found, show what my current setup costs me per year.
The output is a short table. The $5,600 extra in checking costs about $224 a year. $50,000 in a 0.01% savings account costs about $2,000. An IRA balance sitting in cash, measured against even a conservative long-term return instead of 4%, may be costing more than everything else combined.
This isn't investment advice, and the agent shouldn't pretend otherwise. It's arithmetic on your live balances at a rate you chose. What to do about the IRA cash is between you and your risk tolerance, or your advisor. But you can't decide anything about money you don't know is idle.
What the agent can and can't do
BankBridge is read-only. The agent can find the idle cash, compute the cost, and nag you about it every month. It cannot move a dollar. There are no transfer tools, so there's nothing to misfire.
That's deliberate, and it's the right shape for this job. Spotting the problem is the part software is good at. Opening the high-yield account and moving the money takes you ten minutes, once you know it's worth the trip.
Make it recurring
Idle cash grows back. A few good months and your checking balance drifts up again. A dividend lands in the brokerage and sits as cash. One scan fixes today; a habit fixes the year.
Fold the check into a monthly money review, or just ask on the first of the month:
Rerun the idle cash check. Compare against last month and tell me if anything drifted.
If you haven't connected a bank yet, setup takes about five minutes. After that, this entire article is four prompts.