AI in marketing: what small teams can copy from big brands

Big brands have budgets, agencies and in-house studios. Small marketing teams have two or three people and too many channels. Yet some of the most useful AI marketing results of the past year came from teams that were small or stretched, and the methods behind them are easy to copy. Here are five case studies, what each one actually did, and what a Greek or European SME can take from it.

A note on the numbers: most results below come from case studies published by the vendors that sell the tools, and one comes from trade press. They are vendor-reported, not independently audited. Read them as evidence of what is possible, not as a forecast for your business.

1. An AI sales assistant that works weekends: Cabana Life

Cabana Life, a US brand selling sun-protective clothing on Shopify, had a customer service “team of one” and an email-only help desk. Most of its sales happened at the weekend, but weekend questions waited until Monday, and shoppers moved on. The team added an AI sales assistant from REP AI, named it “Jess”, styled it in the brand’s colours and fed it current promotions, launches and the free-shipping threshold.

According to the vendor’s case study, over a 30-day period the assistant generated $4,713.70 in sales on $647 in platform cost, converted 8.71% of chats, and resolved about 90% of conversations without a human. Rollout took about three weeks.

What to copy: start where customers wait. List your ten most common pre-sale questions (sizing, delivery, returns, stock), give the assistant accurate answers, and hand off to a person with the full chat transcript when needed. Then read the chats. Cabana Life used them as a source of product insight, for example learning that its customers wanted longer styles.

2. From 50-page briefs to spreadsheets: Kérastase Taiwan

Kérastase’s Taiwan e-commerce team was two campaign planners and one designer shared with another division. Each monthly campaign needed a 50-page brief that took up to four days to prepare. With creative automation from dipp, the planners started generating product thumbnails, bundles and banners from spreadsheets, using a library of approved layouts. In the vendor’s case study, go-to-market time fell from 20 days to 4, and turnaround on paid-media visuals from 3 days to 2 hours.

What to copy: turn your recurring visuals into templates. Agree a small set of on-brand layouts once, then let non-designers fill them from a product sheet. Your designer’s time goes to new ideas instead of resizing banners.

3. More output from the same team: Intermarché Belgium

Intermarché’s Belgian design team needed more static and video content for ads, social media and newsletters without adding people. Using AdCreative.ai, it turned simple product cutouts into finished compositions and existing static images into short videos. According to the vendor’s press release, the team now produces more than 135 assets a month and cut creation time per batch of visuals by about 40%.

What to copy: reuse what you already have. Before booking a new shoot, see how far your existing product photos can go as on-brand variations and simple video formats.

4. Testing at scale: Brooks Running

Brooks Running had growth targets and a flat budget, and was leaning on a single photo of one shoe in one colourway across audiences and channels. Using Adora, it generated more than 1,500 on-brand ad variations across 40 products and tested specific elements: gym backgrounds versus other settings, shoes shown off-foot, one shoe versus two. The vendor case study reports a 107% year-over-year increase in return on ad spend (ROAS) and a 53% lower cost per acquisition (CPA), on the same budget.

What to copy: test variables, not random variety. A small team does not need 1,500 ads. Five to ten variants per product, each changing one element, will tell you more than one “perfect” ad.

5. Creative as inventory: Jones Road Beauty

This one is reported by trade press rather than a vendor, and the automation is mostly inside the ad platform. D2C Times reports that Jones Road Beauty noticed its best ads wore out in roughly 11 to 14 days. A small in-house team now produces 15–20 new video variants a week, and Meta’s automated creative testing picks the winners within about 72 hours. Losing ads are paused; winning ones get more budget the same week.

What to copy: the rhythm. A weekly creative sprint, a rule for when an ad is “tired”, and a separate testing budget so you can try new ideas without disturbing what already works.

What these cases have in common

• Each started with one narrow bottleneck: weekend questions, slow briefs, too few assets, too little testing.

• Templates, approved information and brand rules came first. The AI worked inside them.

• People stayed in charge of the brand and of what got published.

• Every result was measured against how things worked before.

A 30-day plan for a small marketing team

1. Week 1: pick one bottleneck and write down today’s numbers, such as hours per campaign, assets per month, chat response time or conversion rate.

2. Week 2: set up one tool with your brand templates, approved product information and tone of voice.

3. Week 3: run it for real, with a person approving everything that is published or sent to customers.

4. Week 4: compare with your week-1 numbers and decide: scale, adjust or stop.

Mind the European rules

A chat assistant collects customer conversations, which often include personal data, so GDPR applies: check where the data is stored and what the vendor’s terms allow. The EU AI Act also includes transparency obligations for AI systems that interact with people, so tell customers clearly when they are chatting with an AI, even if you give it a friendly name.

Frequently asked questions

Do these results apply to a 10-person company?

The numbers will differ, but the methods scale down well. Templates, a weekly testing rhythm and an assistant for common questions work at any size.

Where should we start?

Where the delay costs you most. If customers wait for answers, start with an assistant. If campaigns are slow to launch, start with templates and creative automation.

Will AI replace our designer or agency?

In these cases it did not. It took over repetitive production so people could focus on concepts, strategy and brand.

Want to find the marketing use case that will pay off fastest for your team? Take our AI readiness quiz, or book an AI roadmap workshop to turn ideas like these into a 90-day plan.