One will get through. A phishing email that uses a real client’s name, a real bank and a real deadline can reach a firm that has done everything right. What matters is the next hour. That covers who checks what, and what you tell the bank, the client and the person who clicked.
Maya Okafor’s firm got its one on a Monday. Her Columbus firm has six people and does books and taxes for small businesses. Someone got into her account two weeks earlier, and the firm spent the week since locking every door it could find. At 11:40, an email from a bank arrived anyway. It reached Jess, at the front desk, and Jess clicked.
This part of the series is the rest of that Monday and the Tuesday after it. It covers what stopped the email from becoming a second break-in, where the sender got its details, and who Maya told. It also covers the question most owners ask last: whether the law says you have to tell anyone at all.
If you’ve just clicked something you shouldn’t have, you’re in ordinary company. Careful people click. The protection that counts is the kind that still holds after they do.
Everything below holds whichever app opens your mail, Point included. What Point does is covered separately.
Four minutes at the front desk
The email looked right. It came from Kim, the loan officer, at the bank. It said Dale Huang’s upload had failed overnight. It asked the firm to send the file again through a secure link. The logo was right. Kim’s signature was right. Dale’s name was spelled the way Dale spells it.
Jess clicked the link at 11:41. A page opened with the bank’s logo and a button: “Sign in with your work email to view this message.” She clicked that too. A sign-in page opened that looked exactly like Google’s.
She typed her address. Then the page asked for her password, and she typed that as well.
The next screen asked for a six-digit code sent to her phone. Jess picked up her phone and waited. Nothing came.
That’s where she stopped. All week, signing in had meant her phone asking for her face. Since Thursday, she hadn’t typed a code once. This page wanted the old way. At 11:44 she stood up and walked to Maya’s office.
Why the password got through and the sign-in didn’t
Whoever built the page now had Jess’s password. They didn’t have a way in.
On Thursday, the firm had moved every account to passkeys. The account underneath covers why that matters. A six-digit code would have worked on the fake page as well as the real one. A passkey looks at the site’s real address first, and a fake page can’t pass that check. So the page asked for the one thing it could catch, and Jess’s account doesn’t use it anymore.
Maya and Jess still treated the password as lost. Jess changed it at Maya’s desk. Then they opened Google’s Security Checkup for her account, which lets you “check recent security events.” The only sign-ins were Jess’s own laptop and phone. Her connected apps were the same short list Priya had cleaned up on Friday.
By 11:58 Jess’s account was confirmed clean. The page had a password that no longer opened anything.
Who sent it, according to Gmail
The next question was who had sent it. Gmail shows that, if you know where to look. Google’s Gmail Help page on authentication puts it in one line:
Below the sender’s name, click the Down arrow.
That panel shows “Mailed by” and “Signed by” for an authenticated message. Google adds that a question mark next to the sender’s name means the message isn’t authenticated.
Maya clicked the arrow. There was no question mark. Both lines showed a domain, and the message had passed every check. The domain just wasn’t the bank’s. It was the bank’s name with a hyphen and the word “secure” added. Someone had registered it, set it up properly, and sent mail from it.
A message that passes authentication really did come from the domain it names. It doesn’t mean the domain belongs to who you think. Gmail’s checks and the gray padlock both told the truth about the wrong thing.
How a stranger knew about Dale’s loan
The domain explained the sender. It didn’t explain the details. Whoever wrote the email knew Dale, knew Kim, knew there was a loan and knew a document was due.
Priya found the answer in about ten minutes. For the eleven days before the morning they found it, a rule in Maya’s account had forwarded every message containing “invoice,” “payment” or “wire” to an outside address. Two weeks ago, Dale had copied Maya on an email to Kim about the lease. Its subject line was “Lease payment schedule and documents.” It named the bank, Kim, Dale, and the end-of-month deadline. The rule had sent a copy out.
So the attacker hadn’t gotten back in. Every door was still shut. They were using what they’d read before the doors were shut, and that copy was never coming back. Like any email once it’s sent, Dale’s thread couldn’t be recalled from somebody else’s server.
That changed what Maya needed to say, and to whom.
What you tell the person who clicked
Maya started with Jess, because Jess was still standing in her doorway.
She told her that she’d done the thing that mattered. She noticed something was off, she stopped, and she came straight over. Four minutes from email to office is fast. A firm where people hide a click for a day is in far more danger than a firm where someone clicks.
Then Maya told the rest of the office the same thing, out loud, with Jess’s permission. The email had gotten past Google’s filters and past a careful person. It hadn’t gotten past the passkey. Anyone who clicks something odd should walk over, the way Jess did, and nobody would be in trouble for it.
What you tell the bank
Kim was next. Maya called her at the number on the bank’s website, the same one she’d used on Friday. She didn’t use the number in the email.
The FTC’s guidance on phishing says the same thing in one line. Contact the company “using a phone number or website you know is real,” and not the details in the email.
Kim confirmed that Dale’s upload had gone through at 9:05. The bank hadn’t sent anything. She asked Maya to forward the email to the bank’s fraud team, and gave her that address on the phone. A lookalike domain using the bank’s name was the bank’s problem too. Other firms working with Kim might get the same email.
Priya reported it in two more places the FTC names. She forwarded it to the Anti-Phishing Working Group at reportphishing@apwg.org. She also filed a report at ReportFraud.ftc.gov. Then she marked it as phishing in Gmail, which teaches Google’s filters what it looked like.
What you tell the client
Dale was the hardest call, and Maya made it at 12:30.
He’d been patient for two weeks. He’d caught the first fake email himself. Now Maya had to tell him that someone still had a copy of a thread with his name, his bank and his loan in it. She told him plainly what she knew:
- The email went to the firm, not to him, and nobody acted on it.
- His return reached the bank safely on Monday morning, and Kim had confirmed it.
- The details in the fake email came from the thread copied two weeks ago. Nothing new had been taken.
- That copy is out there for good. Anyone holding it knows his lender, his loan officer and his deadline.
Then she told him what it meant for him. Any email about the loan that asks him to click, sign in, pay or send a file is suspect, even if it looks like Kim’s. He should call Kim or Maya at numbers he already has. She also asked him to check his own email account for forwarding rules he hadn’t made. The attacker had his address and knew he was worth targeting.
Dale asked whether he should be worried about his Social Security number. Maya had already checked. The copied thread had a payment schedule and a lease amount. It didn’t have his return, his SSN or an account password.
Whether the law says you have to tell anyone
Dale wasn’t the only client in those eleven days of forwarded mail. On Tuesday morning, Maya and Priya went through every message the rule had copied. There were 23, from nine clients. They wrote down what each one contained.
None held a Social Security number or a full return. Two held invoices with a business’s bank account number on them. The rest were payment amounts, dates and names.
Then Maya checked the three rules that could apply to a tax firm in Ohio.
| Rule | What triggers it | Who must be told, and when | What it meant for Maya’s firm |
|---|---|---|---|
| Ohio Revised Code 1349.19 | Personal information taken, with a material risk of identity theft or fraud | Each affected Ohio resident, within 45 days | Likely not triggered, because no copied message held “personal information” as Ohio defines it |
| FTC Safeguards Rule, 16 CFR 314.4(j) | Unencrypted customer information taken without permission | The FTC, within 30 days, if 500 or more consumers are involved | Nine clients, so no FTC notice |
| IRS guidance for tax professionals | Client data theft | Your local IRS stakeholder liaison, quickly | No taxpayer data was taken, so nothing for the IRS to block |
Ohio’s law turns on its definition of personal information. It means a person’s name linked to an unencrypted Social Security number, a driver’s license or state ID number, or an account number together with the code or password that opens it. A bank account number on an invoice, with no password, falls outside it. Ohio’s definition of a breach adds a second test:
unauthorized access to and acquisition of computerized data that compromises the security or confidentiality of personal information owned or licensed by a person and that causes, reasonably is believed to have caused, or reasonably is believed will cause a material risk of identity theft or other fraud
The FTC’s rule counts more as a notification event. Its definition in 16 CFR 314.2 covers any “acquisition of unencrypted customer information without the authorization of the individual.” The forwarded mail met that. The FTC notice only applies at a larger scale, though:
if the notification event involves the information of at least 500 consumers, you must notify the Federal Trade Commission as soon as possible, and no later than 30 days after discovery of the event.
The IRS sets out what it wants when tax data is stolen. Speed matters, it says, because “the IRS can take steps to block fraudulent returns from being filed using your clients’ personal information.” Without a Social Security number in the copied mail, there was no return anyone could file. Maya still put the liaison’s number on the firm’s one-page record, next to the bank’s fraud line.
Maya isn’t a lawyer, and she didn’t decide this alone. She sent the list of 23 messages and the table to the firm’s attorney on Tuesday at 10:00. The attorney agreed with her reading by the afternoon. If your own copied mail held Social Security numbers or returns, the answer changes. Ohio’s 45 days and the IRS liaison both apply, and the IRS also asks for “an individual letter to all victims.”
The letter that went out anyway
The law didn’t require a letter. Maya sent one to all nine clients on Tuesday afternoon, because Dale’s call had shown her what the copied mail could do. Each of them was now someone a stranger knew a little about.
Every client got the same letter, with one line changed to say what had been copied from their own mail:
Two weeks ago, someone got into my email account for eleven days. During that time, copies of some of our emails about payments were sent to an outside address. For you, that was [two invoices from March and one payment reminder]. They didn’t include your Social Security number, your tax return or any password.
We’ve locked the account down, and we’ve checked that nothing has been taken since. But the person who copied those emails knows your name, that you work with us, and some details of what we’ve discussed.
Please treat any email about payments, documents or bank details with care, even if it looks like it’s from us or from someone you work with. We won’t ever send you a link to upload files or new bank details by email. If you get one, call me at the number on our website.
If you think anyone has your Social Security or bank account number, the FTC’s IdentityTheft.gov walks you through what to do.
The last line came from the FTC’s own phishing guidance, which sends anyone in that position there. Maya added the rest herself. She kept it short, said what was taken, and gave each client one thing to do.
Seven clients replied by the end of the day. Most said thanks. One said it was the clearest email she’d ever had from a professional about a problem.
Tuesday at 4:15
At 4:15, Maya sat back and counted what had held. The passkey held when the password didn’t. Jess walked over instead of hoping nobody would notice. Kim’s real number was in the firm’s records because Maya had called it on Friday. Every protection that worked on Monday had been set up in the week before.
Then she thought about the week itself. It had been the firm’s whole focus. Everyone had dropped something to make room for it. Tax season starts in January, and nobody would have a spare week then. Jess’s four minutes depended on Jess still noticing in March, on her sixtieth return of the month, with the phone ringing.
The firm had locked everything it could find. Keeping it locked was a different job, and nobody had it yet.
On the screen in front of her, the firm’s shared calendar was already filling with January drop-off appointments and filing deadlines. Maya clicked into the first empty Monday and started a habit, not a project.