Witnium logo
WITNIUM. Chain.
← News

Sector · Retail and e-commerce

Cases · 2026-11-05 · 3 min read

Toast allegedly misled subscribers about *customer service strength*, FTC says

The US Federal Trade Commission has sued Toast, a payment software provider for restaurants, for allegedly hiding fees and making cancellations difficult. The case turns on what the company promised, and what subscribers could prove it said.

EliasBy Elias

Share
Illustration showing a conversation that exists in two different versions
Illustration: AI-generated

What Toast allegedly did

Toast provides payment-processing software to restaurants. According to a complaint filed by the US Federal Trade Commission in January 2025, the company allegedly charged hidden fees, made subscription cancellations difficult and misled subscribers about the strength of its customer service [1].

The allegations include that Toast marketed 24/7 customer support but often failed to deliver it, and that subscribers found it harder to reach help than the company had suggested [1].

It's like booking a hotel that promises a 24-hour concierge and finding a note saying they're on a break.

The complaint doesn't allege that Toast used AI in its customer service. What makes the case interesting is the evidence question it raises: when a company promises something and a customer disputes whether the promise was kept, who has the record of what was actually said?

The evidence question

In disputes about service quality, the records often tell different stories. A company's internal notes say one thing, and a customer's screenshots or recollections say another.

When a customer contacts support through a chat, an app or a chatbot, they rarely keep a sealed record of the conversation. The company usually does keep one, but it's a record the company itself controls and can edit. That makes it easy to challenge.

The FTC's case against Toast doesn't appear to turn on chat logs, but it highlights a wider problem: when what was promised matters, the side without its own checkable record starts on the back foot.

Since 2 August 2026, the AI Act has required people to be told when they're talking to an AI, unless it's obvious [2]. That helps with transparency, but it doesn't answer the evidence question. If a chatbot makes a promise and the customer disputes it six months later, who can show what was actually said?

How a digital witness fits here

The thing to seal is each conversation or each answer the moment it's sent: what the chatbot said, when it said it and that the customer was told they were talking to an AI.

Sealing takes a fingerprint of the record, not the record itself, so the conversation never leaves your systems. You get a witness ID back. If the exchange is questioned later, anyone you share the record with can calculate the fingerprint again and check that it matches the sealed one.

What it shows is that the record of what the chatbot said hasn't changed since it was sealed. What it doesn't show is that the answer was correct, or settle whether the company is bound by what the chatbot promised. Those are separate questions.

A sealed record won't stop every dispute, but it does mean both sides are looking at the same unchanged record when they try to sort one out.

How it would work

  1. Your chat system sends a fingerprint of each answer to WitniumChain through the REST API, the SDK or the MCP server.
  2. You get a witness ID back and store it with the conversation.
  3. Later, anyone you share the record with can check that the fingerprint still matches.

The first 2,000 witnesses a month are free.

What this means for you

If you run a chatbot or any AI that talks to customers, here are three questions worth asking:

  • Could you show what your AI said to a specific customer on a specific day?
  • Do you have a record that the customer can't dismiss as something you could have quietly edited?
  • When you tell customers they're talking to an AI, as Article 50 requires, do you keep a sealed record that you did?

The FTC's case against Toast is about service promises in general, not AI. But the principle is the same: when what was said matters, the side with the checkable record is in a stronger position.

What to watch

The Toast case is ongoing. No date has been set for a hearing [1].

Article 50 of the AI Act has applied since 2 August 2026 [2]. Companies using AI to communicate with customers should already have a way of showing that people were told when they were talking to an AI.

For high-risk AI systems under Annex III, the logging requirements in Article 12 apply from 2 December 2027. Those logs are what you'll need if a decision is questioned, and proving they haven't been altered is part of making them hold up as evidence.

Sources

  1. [1]Toast misled subscribers about strength of customer service, FTC alleges, Compliance Week, 2025-01-09
  2. [2]Article 50, Regulation (EU) 2024/1689 (AI Act), European Commission AI Act Service Desk
Pass it on

Know someone who should read this?

Share it with a colleague, your compliance team or your network.

Email

Sealed with WitniumChain

69b4518f33b8ce552a119d5843097acd308e2472a241e91ae4d568ab3adbedf8 · 2026-11-05 13:27 UTC

Written with AI assistance and reviewed by Elias. Every version of this article is sealed with WitniumChain.

Related