A Fraud Model Froze a Caterer’s $8,740 for 11 Days

In June 2024, a caterer we will call Lena opened the spreadsheet she used to track every event. One line showed the money available for the next week: $312.

The number should have been $9,052. Most of that money, $8,740, sat in her business account, but she could not send it to a supplier, pay a worker or move it to another account. A notice on the dashboard said the account had been restricted after unusual activity. It did not identify a transaction she could explain or a document she could correct.

June was her busiest month. The spreadsheet held deposits for a graduation dinner, a retirement party and two company lunches, each entered beside the food cost and the amount still due. Lena had spent five years building the business from weekend trays prepared in a rented kitchen into regular work that supported her and four part-time workers.

The account freeze began after she received several customer payments close together and bought more food than she did in an ordinary week. Those were normal events for a caterer entering graduation season. To the system reviewing the account, they appeared to be something else.

She could not see the model’s inputs or the threshold she had crossed. The notice did not say whether the concern involved the deposits, her purchases or some combination of activity. It said the company would review the account.

The line that changed the month

The first event on Lena’s spreadsheet required $2,180 in food and kitchen costs. The customer had already paid a deposit, but that money was inside the restricted account. Canceling would have meant returning money she could not reach, losing the remaining payment and leaving a family without food for 86 guests.

She used a personal credit card.

The next event pushed another $1,460 onto a second card. She moved two workers’ payments through a separate payment service linked to her personal checking account, which left too little there for her apartment payment. Her brother transferred her $2,000. She entered that amount in the spreadsheet under available cash, though it was not business income and she did not want to treat it as such.

By the fourth day, she had contacted the support inbox and used the number on the account page. The replies confirmed that a review was underway, but they did not identify a person who could reverse the restriction. She submitted identification and business records through the account portal, then sent copies of customer invoices after support requested more information.

The documents showed what the transactions were. They did not produce a conversation with the person making the decision, and Lena could not tell whether any person had looked at them.

That distinction mattered to her. She understood why a company might stop a suspicious transfer. What she could not understand was how the same company could hold ordinary business revenue for days while giving her no way to explain the pattern to someone who could compare the account with her calendar.

Her spreadsheet became the appeal the platform did not provide. It connected each deposit to an event and each large purchase to a menu, with amounts that could be checked against invoices. Lena kept updating it even though there was nowhere to upload the file as a whole and nobody asking what the month looked like from her side.

No person to appeal to

A fraud model does not need to close an account to change a business. A temporary restriction can be enough, particularly when a company depends on customer deposits to buy food before collecting the final balance.

Lena had no reserve equal to 11 days of expenses during June. She had cash moving through the account, but little cash outside it, which meant the platform’s decision turned a liquidity problem into personal debt without issuing her a loan or making any assessment of whether she could repay one.

She kept taking orders because June revenue covered slower months. Turning customers away would protect the remaining limits on her cards, but it would also remove income she expected to use for rent, taxes and summer kitchen fees. She accepted one smaller lunch and declined a larger weekend event that required a $1,900 food order.

There was no clean calculation. The declined event might have produced about $1,100 after food, labor and kitchen costs, according to her spreadsheet, but only if the customer paid the balance and nothing went wrong. Lena left that projected profit in the file and marked the event lost.

On the seventh day, one card was close to its limit. Her combined personal card balances had risen from $3,260 to $8,440, while her total limits remained $10,800. She had not missed a payment. The amount of available credit had changed anyway.

Credit scores can respond to the share of revolving credit in use after card issuers report balances. The score displayed through one of Lena’s card accounts later fell 31 points, from 704 to 673. The display did not assign a single cause, though the higher balances were the largest change she could see during that period.

The drop did not stop the catering work. It narrowed her margin for the next problem. When she later checked the terms for replacing a failing refrigerator, the financing offer carried a higher rate than the offer she had seen that spring. She postponed the purchase and paid for a repair.

Meanwhile, support continued to say that the account was under review. Lena received no explanation of what evidence would settle the concern and no estimate she could put into her spreadsheet. She began each morning by checking whether the $8,740 had moved from restricted to available.

It had not.

The account reopened, but the costs stayed

On the 11th day, Lena signed in and found that she could use the account again. The company did not identify the transaction that had triggered the restriction. It did not say whether her documents cleared the concern or whether the model had reconsidered the activity after more time passed.

There was no apology from a person because she had never reached one with authority over the review. The restriction was removed through the same dashboard where it had appeared.

Lena paid $3,600 toward the personal cards that week. She paid her brother back the next month. The remaining card balance took four months to bring near its earlier level, and her statements recorded $286 in interest charges during that period.

Those charges were not the largest cost in the spreadsheet. The lost event still showed $1,100 in projected profit. Worker payments had gone out, customers had been fed and the company had released every dollar it held. From the account’s point of view, the incident had ended.

From Lena’s side, the release came after the decision had already moved through the rest of her finances. A restriction on a business account became a higher personal credit balance, then a lower displayed score, while the reason for the original flag remained unavailable to her.

Eight months later, Lena still used the platform. Changing services would have meant moving customer payment links during another booking season, and the alternatives she reviewed also reserved the right to restrict accounts during fraud checks.

She changed what she could see. The spreadsheet gained a separate line for money held outside the main account. By the following June, that line showed $2,400, less than one busy week of costs but more than the $312 she had been left with.

Questions people ask

Can an automated fraud flag freeze a business account?

In Lena’s case, unusual activity detected by the platform led to an immediate restriction while the company reviewed the account. She could receive information about the review through support, but she could not use $8,740 for 11 days, even though the account remained open and customers had paid for real events.

Was there a human appeal?

Lena sent identification, invoices and business records through the channels the platform provided. She never reached a person who explained the flag or had a live discussion with her about reversing it, so she could not tell whether a human reviewed the evidence before the account became available again.

How did the freeze affect her credit?

She charged $5,180 in business costs to personal cards because the operating money was restricted. Her balances rose from $3,260 to $8,440, and the credit score displayed through one card account later fell 31 points. The display did not state one cause, but the balance increase was the main financial change she recorded.

Did the company cover what the freeze cost her?

Lena received access to all $8,740, but she was not reimbursed for $286 in card interest or the event she declined. Her spreadsheet kept those amounts separate. The last line tied to the freeze was the $1,100 in projected profit from work she never booked.