Marketing claims depend on product data as well as wording. A price is approved, a campaign launches and the artwork is saved with its review record. Later, the product team changes a fee, an eligibility condition narrows or a promotional period ends. The communication may remain untouched. Its relationship with the offer has changed.

The FTC’s Greenway Auto case concerns a different and more direct allegation: that advertised prices were below the amounts customers actually paid. On 8 October 2026, the agency announced a proposed settlement. The complaint’s analysis of vehicles advertised on CarGurus and Cars.com at the defendant Kia dealerships, covering 1 March 2023 to 6 May 2026, alleges that over 92% of consumers were charged more than advertised. Those consumers paid at least $3,354 more on average, excluding required government charges and optional add-ons. The case is pending, and the stipulated order requires judicial approval to have the force of law.

The case does not establish that Greenway’s alleged conduct resulted from stale data or an innocent update failure. It does expose the distance that can open between a marketing claim and the commercial reality it purports to describe. For other organisations, that distance can become a problem even when the original review was sound.

Wording review cannot establish the offer on its own

An advertised price has meaning because of what a customer can obtain for it. Reviewing the number in a file is therefore incomplete without understanding the charges and conditions attached to the purchase. A perfectly copied figure can still misrepresent the proposition if its source omits a mandatory component.

The FTC alleges that Greenway added mandatory charges to advertised prices and, at times, used conditional rebates available only to some consumers. The proposed order addresses price transparency. Those allegations and proposed remedies belong to a US automotive case; they should not be presented as a new global pricing rule or as findings already made by a court.

An illustrative mortgage campaign raises an analogous data problem under a different framework. A quoted rate may depend on a loan-to-value band, fees and eligibility criteria. A reviewer needs the relevant product version and presentation conditions to understand the claim. If the business later changes one of those facts, it should be able to locate the communications whose meaning depends on it.

This is more than checking for a stale number. A benefit described as available to all customers can become inaccurate when eligibility changes, even if the benefit itself remains unchanged. A claim supported by one piece of evidence can need reassessment when that evidence is withdrawn or qualified. The dependency concerns the proposition expressed, rather than a literal match between a spreadsheet cell and a sentence.

Approval should leave a route back to the source

An approval record can show who accepted the wording and when. To manage later changes, it also needs to identify the facts and evidence on which the acceptance depended. That route back to the source lets the organisation ask which decisions a change affects.

For a price claim, the record might connect the product, relevant fee schedule and eligibility conditions to the assets using them. For an efficacy claim, it might connect the precise assertion to the supporting evidence and the limitations the reviewer considered. These are proposed operating arrangements, not requirements derived from the Greenway order.

The practical benefit appears when the source changes. Instead of relying on someone remembering every campaign that mentions the offer, the organisation can identify the dependent work. A system could flag it for review, stop further generation using the superseded fact or route the issue to the responsible owner. Which action is appropriate depends on the change and the context.

This is an extension of preserving the conditions of an expert decision. The source version is part of those conditions. A historical approval remains useful evidence of what the organisation decided, but it should not be treated as a perpetual permission to repeat the claim.

Live content makes ownership harder

The affected work may sit across a website, paid media accounts, customer emails and partner material. Some assets can be edited immediately; others have already been delivered. A source change therefore needs more than a list of files. It needs a practical account of what is live, who controls it and which response is possible.

An illustrative campaign may have an approved landing page and several advertisements directing to it. Updating the page does not necessarily correct the promise in the ads. Equally, stopping the ads does not remove a claim from an affiliate’s page. The business needs to follow the relationships far enough to understand the exposure and decide the appropriate response.

The related Consumer Duty monitoring argument shows why the response should also consider customers’ experience. Replacing the source data is an operational action. Whether customers have been misled, need an explanation or require other intervention is a further question under the applicable framework.

Ownership should be established before a change creates urgency. Product teams can identify authoritative facts, marketing teams can identify where claims are used and compliance can assess materiality and presentation. A clear relationship between those responsibilities is more useful than expecting one team to discover the whole problem after a complaint.

AI can follow dependencies if the organisation has defined them

An assistant could help locate statements that appear to depend on a changed fact, including paraphrases that a simple text search would miss. It could assemble the live versions and supporting records for a reviewer. That is a valuable use of AI because the task often involves dispersed evidence and repetitive comparison.

It also has limits. A model may miss an implied claim or flag material that is unaffected. A generated summary may lose a condition that determines whether a price applies. The business needs authoritative source records and an accountable assessment of the affected communications, rather than assuming that semantic search has produced a complete inventory.

There is no need to make every update trigger a universal reapproval. A change to an internal label is different from a new mandatory charge. The workflow should distinguish changes that alter customer meaning from those that do not, while preserving enough evidence to explain that judgement.

Greenway is a case about alleged price deception, not a study of marketing systems. Its relevance to operations lies in the question it forces into view: whether the advertised proposition can be reconciled with what a customer is actually offered. A durable approval process keeps that relationship observable, so the truth beneath the claim cannot move unnoticed while the words remain reassuringly familiar.

What teams need to know

Are the FTC’s allegations court findings?

No. The case is pending. Describe the complaint’s account as allegations and the settlement terms as proposed pending judicial approval.

Does the case prove that stale product data caused the problem?

No. The argument about source changes is a wider operational implication, not an explanation established by the Greenway evidence.

What should a claim be connected to?

The authoritative product facts or evidence supporting it, including the relevant version and conditions that affect its meaning.

Should every source update invalidate every asset?

No. Assess which communications depend on the changed fact and whether their customer meaning has materially changed.

How can AI assist with this work?

It can identify possible dependencies and collect affected material. Treat that output as support for an accountable assessment, with its limits visible.