The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Companies are using AI in marketing to generate and test creative, predict which audiences are more likely to convert, personalize customer journeys and offers, and help advertisers manage campaigns. Vendor-published case studies report gains in metrics such as click-through rate, conversions, sales, revenue, and return on ad spend—but those results belong to specific campaigns, not a forecast for another business.
What the case studies report
The figures below are reported by the vendors that published the customer stories. They are not a like-for-like comparison: campaigns, time periods, metrics, and evidence designs differ. A lift in clicks, for example, is not equivalent to a lift in revenue.
As an Amazon Associate I earn from qualifying purchases.
| Company or case | Marketing job | Reported outcome and evidence context |
|---|---|---|
| Oneisall, UK | AI-generated ad creative and testing | Amazon Ads attributes to Oneisall in 2025 more than 50% year-over-year sales growth, a 22% year-over-year reduction in advertising cost of sales (ACOS), and ad recall 12 percentage points above an industry benchmark. The case describes creative variations and A/B testing but does not establish that every other campaign variable was controlled. |
| Dandy Blend and Trellis, US | AI-generated images tested against existing creative | For the September 2024–January 2025 campaign, Amazon Ads reports click-through rate (CTR) rising from 0.6% to 1.1% (described as an 83% lift), conversions increasing from 481 to 1,055, and ACOS moving from 7.0% to 6.8% as ad spend increased. The case describes a comparison of selected AI-generated images with original brand images; it does not establish that other variables were controlled. |
| Blueair, US | Predictive ad targeting | Amazon Ads reports a 176% ROAS lift, 50% lower cost per acquisition (CPA), and 66% year-over-year sales growth for February–December 2024. Amazon says this is a single-advertiser result and is not indicative of future performance. |
| Thorne, US | Predictive ad targeting in an early beta | Amazon Ads reports 1.5× unique reach, 1.7× pageviews, and 1.9× attributed purchases for November–December 2024. These outcomes were advertiser-reported and came from early Brand+ beta adoption. |
| Coca-Cola en tu Hogar (CCETH), Latin America | Real-time cart-abandonment reminders | Adobe’s customer story, dated December 10, 2024, reports increases of 36% in email opens, 21% in click-through, and 8.5% in conversion for this intervention. |
| Coca-Cola Store, US | On-site recommendations and search | In the same Adobe story, recommendations based on behavior and affinities are associated with 117% more clicks and 36% more revenue; “Frequently Bought Together” recommendations had a 17% CTR, and on-site search conversion reached 19%. These are US store results, separate from CCETH’s cart-reminder campaign. |
| Unnamed quick-service restaurant client | Personalized offers and customer journeys | ZS reports more than $100 million in incremental revenue lifetime to date, revenue lift above 6%, more than $4 return per marketing dollar, and 70% higher net revenue per targeted customer. ZS does not name the client or state a publication date on the consulted case-study page. |
| Unnamed e-commerce company | Advertising operations and seller support | Accenture reports year-over-year advertising-spending growth above 30% and says some sellers who had not previously advertised became active. The consulted case does not name the client or state a publication date. |
These are vendor customer stories, not independent evaluations of typical results. Treat them as examples of reported deployments and outcomes; they do not establish that AI alone caused the change or predict what another company will achieve.
How companies are using AI in marketing
Generating and testing advertising creative
Oneisall, a UK smart-pet-product brand, used AI image generation for Sponsored Brands, Sponsored Products, and display advertising. Amazon Ads says the team used brand-guided prompts to produce creative variations, then A/B-tested them and selected assets based on campaign performance. The useful lesson is the workflow: AI supplied options, while testing—not generation by itself—helped determine which creative to use.
In the Dandy Blend campaign, advertising technology company Trellis generated more than 200 images for the US beverage brand. The Amazon Ads case describes comparing selected generated images with the brand’s original imagery. This is a more explicit creative comparison than a simple before-and-after account, although the published description does not establish that other campaign conditions were held constant.
#1 Best Overall
Predicting which audiences to reach
Amazon Ads describes Performance+ and Brand+ as using behavioral and first-party signals to predict customers likely to convert. Blueair used Performance+ across display campaigns; Thorne’s example involved Brand+ in an early beta. The reported outcomes are illustrations from those advertisers and periods, not evidence that the products will produce the same results for other brands. In particular, Amazon explicitly cautions that Blueair’s single-advertiser result is not indicative of future performance, while Thorne’s outcomes were reported by the advertiser.
Personalizing journeys and recommendations
The Adobe story covers two distinct Coca-Cola contexts. For Coca-Cola en tu Hogar, the Latin American business unit connected ecommerce behavior, order and profile information, and ERP and CRM data into unified customer profiles using Adobe Real-Time CDP, Journey Optimizer, and Commerce. The team used those profiles and real-time flows to send an email when a shopper had not checked out within an hour, rather than relying on data that could arrive as much as 48 hours later.
The same story describes the separate US Coca-Cola Store using behavioral and affinity signals for one-to-one recommendations, including a “Frequently Bought Together” cross-sell. Keeping the two examples separate matters: CCETH’s intervention was a timely cart-abandonment email, while the US store case concerned on-site recommendations and search.
Selecting offers for restaurant customers
ZS describes an unnamed quick-service restaurant client using Personalize.AI to assign customers to journeys such as churn prevention, upselling, and cross-selling. The client ran more than 100 campaigns annually, according to ZS. The case describes the platform drawing on historical and real-time data and supporting multivariate tests across offer types, messages, product selection, creative, and pricing. Because the client is unnamed and the page does not give a publication date, the reported results cannot be tied to a specific restaurant or period from the consulted material.
Supporting campaign operations and advertisers
Accenture’s unnamed e-commerce case focuses on the people running campaigns rather than on a consumer-facing AI-generated ad. The company used data, AI, and generative AI to give sellers advertising recommendations and real-time campaign insights through an ad portal, alongside human account support. Accenture says some sellers who had not advertised before became active advertisers. This is an example of AI supporting the operation and adoption of advertising, not a disclosed test of campaign creative or consumer response.
How to assess an AI marketing example
Before applying a case study to your own campaign, identify what was measured, how the comparison was made, and what information is missing. The reported metrics answer different questions: CTR reflects clicks relative to impressions; conversion counts or rates concern actions taken; ROAS relates advertising return to spend; CPA measures acquisition cost; and sales or revenue are business outcomes. A strong result on one measure does not automatically mean improvement on the others.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Rank #4
- Check the actual intervention. Was AI used to make creative variations, select audiences, trigger a customer journey, choose offers, or advise advertisers? These are different applications with different requirements.
- Read the comparison carefully. A/B testing provides a comparison, but a published summary may not say whether other variables were controlled. Pre/post changes and reported outcomes without a detailed control design are weaker grounds for attributing the change to AI.
- Keep scope attached to every number. Geography, dates, advertiser, product beta status, and whether results were vendor- or advertiser-reported affect what a figure can reasonably mean.
- Look for the operating inputs. The examples involve brand-guided creative prompts, behavioral or first-party signals, connected customer records, timely event data, or campaign support. The AI component is part of a workflow, not a substitute for the data and decisions that workflow depends on.
- Do not treat vendor examples as benchmarks. None of these cases provides an independent cross-platform comparison that ranks tools or predicts performance for a different business.
What these examples do—and do not—show
Together, the cases show several practical roles for AI in marketing: producing candidate creative for testing, predicting audiences, activating customer data in relevant moments, tailoring offers, and making campaign management more accessible to sellers. They also show why a headline percentage needs its context. The vendor-published stories document particular deployments and reported outcomes; they do not establish a universal return, prove that AI alone produced each result, or identify a best platform across these different jobs.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
