Research Highlights the Repeat-Purchase Question Facing AI Advertisers

A successful sale can still leave a brand with a trust problem when customers learn later that artificial intelligence created the advertisement. New survey results suggest that disclosure timing deserves the same attention as creative performance.

A ContentEngine survey of 446 U.S. adults examined how people might respond when an advertisement used AI to create all of its words and images. Every respondent imagined buying an unplanned product and feeling satisfied with it. The key difference involved disclosure: one group learned about the AI before buying, while the other found out afterward and learned that the advertisement had not disclosed it.

The group that learned later reported lower trust and weaker interest in buying from the advertiser again. About 65% said they would trust the advertiser at least a fair amount, compared with 76.7% among respondents who knew upfront. Repeat-purchase intent showed a similar gap: 67.3% in the later-discovery group said they would be very or somewhat likely to buy again, compared with 79.8% in the upfront group.

Those figures point to a practical concern for advertisers. AI can help produce an ad that converts, while the way the ad was made can still affect the relationship that follows. For brands focused on customer lifetime value, a disclosure shown before the purchase may carry more value than a notice added to a post-purchase page or buried in a policy.

The survey gave every respondent the same positive product outcome. Each person imagined using the product and feeling satisfied. That detail matters because it removed dissatisfaction with product quality as an explanation for the reactions.

Even with that favorable result, the later-discovery group gave weaker answers on the measures most closely tied to future business. Trust was 11.7 percentage points lower, and repeat-purchase intent was 12.6 points lower. Respondents who learned after the purchase also felt more misled: 70% reported feeling misled a lot or some, compared with 57% of those who knew before buying.

The results do not establish that disclosure timing alone caused every difference. The later scenario specifically said the advertisement had failed to disclose its AI use. That wording calls attention to nondisclosure and could have influenced the response to the question about feeling misled. The two groups also came from separate online samples, so unmeasured differences between respondents may have affected the results.

Even with those limits, the pattern gives advertisers a useful warning. Customer satisfaction may protect a brand from some complaints about the product itself. It may not repair a sense that the brand withheld information about how it persuaded someone to buy.

A customer who discovers undisclosed AI involvement may reassess the entire interaction. The product can remain useful. The advertisement can remain effective. The brand may still lose some credibility because the customer now views the purchase decision through a different lens.

A ContentEngine survey of 446 U.S. adults suggests that when consumers learn about an advertisement’s AI-generated words and images can influence what happens after the sale. Respondents who were told upfront reported higher trust and stronger repeat-purchase intent than those who learned afterward that the ad had not disclosed its AI use, even though both groups imagined being satisfied with the product. The findings are hypothetical and do not prove that timing alone caused the differences, but they give advertisers a practical reason to disclose material AI involvement before consumers act, maintain meaningful human review and measure relationship outcomes beyond conversion.

A disclosure shown before an ad does its work gives consumers information while they are deciding how much weight to give the message. A disclosure delivered afterward primarily explains an event that has already happened. The two notices can contain the same words and serve different purposes.

This distinction matters most when an advertisement encourages an unplanned purchase. The consumer has already moved from attention to action. Once the purchase is complete, a later label cannot change the information available at the moment of persuasion.

The survey’s upfront scenario used a clear condition: the ad stated that its words and images were created entirely using AI. That is a stronger disclosure than a label for a small AI-assisted task, such as resizing an image, removing a background or generating alternate headlines. Brands should avoid treating all AI use as one category when setting internal rules.

A useful disclosure should appear where consumers can reasonably see it before acting. Depending on the channel, that could mean an ad label, an expandable disclosure near the creative, a voice statement in an audio spot or a visible notice attached to a sponsored post. Placement, wording and readability all affect whether a disclosure actually informs a person.

The IAB AI Transparency and Disclosure Framework V2 describes a risk-based, materiality-driven approach for deciding when and how AI involvement should be disclosed in advertising and marketing. That approach gives advertisers room to account for the role AI played, the audience, the medium and the possible effect on consumer understanding.

The survey found broad support for human review. Across both groups, 88.3% said it was very or somewhat important to know that a person had reviewed and approved an AI-generated advertisement before publication. The figure was almost identical in each sample: 87.9% among those told upfront and 88.8% among those who learned later.

That result separates two questions that often get combined. Consumers may want to know that AI helped create an advertisement. They may also want to know that a person remains accountable for the final claims, images and targeting decisions.

A human review process should have clear responsibilities. A reviewer can check whether product claims are accurate, whether an image creates a false impression, whether a synthetic person or scene could confuse viewers, and whether the targeting is appropriate for the audience. The reviewer can also confirm that the disclosure is visible and understandable.

A rubber-stamp approval offers little practical protection. If a brand tells consumers that a person reviewed the ad, the company should be able to identify who approved it, when the review happened and what standards were applied. A simple record can help marketing, legal and compliance teams resolve questions later.

The European Commission’s explanation of Article 50 of the EU AI Act also emphasizes the role of the organization using an AI system. Employees working under the direction and control of an advertising company are not treated as separate deployers merely because they operate the tools. For brands serving European audiences, responsibility remains connected to the organization deploying the system and to the way it communicates with the public.

Advertisers should treat AI disclosure as part of the customer experience rather than as a compliance detail added after production. The survey does not establish a universal causal effect, but its pattern is commercially relevant: a hidden production choice can change how customers interpret an otherwise successful purchase. Clear, pre-action disclosure paired with accountable human review is the most defensible operating standard, especially when AI creates the core message or imagery. Brands should test how that transparency affects both immediate response and longer-term trust instead of assuming that a short-term conversion win represents durable value.

Upfront disclosure produced better results on trust, repeat purchase and feeling misled. It did not remove every negative reaction. In that group, 57% said they felt misled a lot or some, and 62.3% said they felt manipulated a lot or some.

The wording of the scenario likely matters. Respondents imagined an AI-created advertisement persuading them to make a purchase they had not planned to make. The sense of manipulation may have come from the persuasive technique, the unplanned purchase, the use of AI or a combination of those factors.

The survey did not ask respondents to explain their answers, so the findings cannot identify the exact source of that discomfort. Advertisers should therefore resist a simple formula in which adding a disclosure resolves every concern. A visible label can improve transparency while the ad itself still feels too aggressive, too personal or too difficult to evaluate.

Creative teams may need to review the whole experience. The questions include whether the ad makes a claim that a consumer can verify, whether synthetic imagery creates expectations the product cannot meet, whether personalization feels intrusive and whether the call to action gives people enough time to consider the purchase.

Disclosure works best alongside sound advertising judgment. It cannot compensate for exaggerated claims or a creative concept that makes people feel pressured.

Other research points to a growing need for clear signals around AI-made marketing. A 2026 Bentley University-Gallup Business in Society survey reported that 79% of Americans had seen advertisements in the previous 30 days that appeared to be made using AI. Canva reported that 70% of consumers believe it will eventually become impossible to identify AI-generated advertising without disclosure, with 56% expecting that situation within two to five years.

Those findings describe a market where visual and written cues may become less reliable. Consumers may increasingly depend on labels, platform signals and brand explanations to understand how an ad was produced.

Research published in Internet Technology Letters has also examined how disclosure mechanisms relate to trust, credibility and purchase intention in AI-generated advertising. The broader research direction is clear: transparency is becoming part of the consumer experience, not a detail limited to internal production teams.

Brands that wait for consumers to ask may create an avoidable information gap. The more common AI-generated creative becomes, the more likely audiences are to develop expectations about when companies should disclose it and how specific that disclosure should be.

Conversion rate remains useful, but it answers only one part of the business question. The ContentEngine findings suggest that advertisers should monitor what happens after the initial action.

A campaign review can include measures such as:

A controlled test can compare different disclosure placements and wording while keeping the product, audience and offer consistent. The test should track both short-term performance and later indicators of confidence. A label that slightly reduces immediate clicks may still support stronger retention if it helps customers understand the interaction.

The survey itself cannot answer how consumers will behave in a live campaign. It measured reactions to hypothetical scenarios, and respondents did not view an advertisement, spend money or use a real product. Brands can add practical evidence through user research, field tests and post-purchase feedback.

A practical process can begin when a team decides how AI will be used. The record should identify whether AI created the full advertisement, generated particular elements, edited existing material or helped produce variations. That distinction can guide the level of disclosure and review.

Before publication, a human reviewer can check four areas:

Teams can then keep an approval record containing the creative version, disclosure text, reviewer, date and supporting claim evidence. This process creates a practical line between using an automated tool and delegating responsibility to that tool.

The same record can help a brand respond when a customer asks how an advertisement was made. A clear answer is easier when the company has documented the production process and can explain the human checks behind the final version.

The survey’s central business question is simple: will a customer want to buy from the brand again after learning how the advertisement was made? In the study, upfront disclosure was associated with a 12.6-point advantage on that measure.

The result should be read with care. The study used separate nonprobability samples of 223 respondents, did not weight the results and did not randomly assign people to the two scenarios. It cannot represent all U.S. adults or prove that timing produced the differences. The later scenario also included an explicit statement that the ad had failed to disclose its AI use.

The findings still give marketers a concrete reason to examine disclosure before launch. Immediate performance shows whether an ad can prompt action. Repeat purchase, trust and perceived honesty show how the interaction may shape the next decision.

For brands using AI to create advertising, the safest operating assumption is straightforward: tell consumers before the ad influences the purchase, make the information easy to find and keep a person responsible for the final message. That approach will not remove every concern, but it gives customers the information at the point when it can still matter.

The survey’s central business question is simple: will a customer want to buy from the brand again after learning how the advertisement was made? In the study, upfront disclosure was associated with a 12.6-point advantage on that measure.

The result should be read with care. The study used separate nonprobability samples of 223 respondents, did not weight the results and did not randomly assign people to the two scenarios. It cannot represent all U.S. adults or prove that timing produced the differences. The later scenario also included an explicit statement that the ad had failed to disclose its AI use.

The findings still give marketers a concrete reason to examine disclosure before launch. Immediate performance shows whether an ad can prompt action. Repeat purchase, trust and perceived honesty show how the interaction may shape the next decision.

For brands using AI to create advertising, the safest operating assumption is straightforward: tell consumers before the ad influences the purchase, make the information easy to find and keep a person responsible for the final message. That approach will not remove every concern, but it gives customers the information at the point when it can still matter.