Tal MorCase study
eCommerceMentorshipFractional CMOFAQ Case studiesBook a call
BrillianteersMeta adsAI creative

AI makes the ads. Meta picks the winners.

Brillianteers is the engagement ring brand I co-founded in 2011. Meta rewards brands that test a lot of creative, and a jewelry brand can't run a photo shoot every week. So I built a loop: AI generates statics and videos, they launch on Meta automatically, and what doesn't sell gets switched off.

Book a free AI audit call →AI automation services
My own brand, on its live ad account.
Tal Mor, Brillianteers AI Meta Ads Case Study
The creative loop
GenerateLaunchMeasureCutRepeat
Static + videogenerated from real product photos
10 anglesfrom lab diamond pricing to proposal emotion, each tested
Built and running. Happy to show it on a call.
10
Ad angles in rotation
Image + video
From several AI models
Always on
New ads in, losers out

The brand

Brillianteers sells engagement rings and fine jewelry online, natural and lab-grown diamonds. I co-founded it in 2011 and grew it to mid seven figures, mostly on Facebook ads and SEO.

Jewelry is a visual, high-consideration purchase. The ads have to look premium, and they wear out fast.

BrandBrillianteers (brillianteers.com)
IndustryEngagement rings and fine jewelry
Store platformMagento
Ad channelMeta (Facebook and Instagram)
Built withKie.ai API, image and video models (Nano Banana, Veo, Kling, Seedance and others), Meta Marketing API, Claude
My roleCo-founder

The problem

  1. Creative fatigue. Meta burns through ads fast, and a brand running the same five ads watches its costs creep up.
  2. Shoots are slow and expensive. A jewelry shoot takes weeks to plan and gives you one look.
  3. Launching by hand eats the day. Uploading, naming and setting up dozens of ads is pure busywork.
  4. Picking by gut. Without a rule, the ads you like stay on longer than the ads that sell.

What I built

  1. A prompt system by angle. Ten proven angles, including the middleman cut, lab diamond pricing, natural diamond value, proposal emotion, social proof, financing and classic designs. Each has its own prompts.
  2. Statics from real products. Image models take the real ring photos and place them in new scenes, so the product stays accurate.
  3. Video from the stills. Video models, run through the Kie API, turn approved stills into short vertical ads.
  4. Automatic launch on Meta. Approved creatives are uploaded and launched through the Meta API into a testing setup, named consistently.
  5. Meta judges, not me. Each ad gets enough spend to show a result. Ads that don't convert are switched off, and winners keep the budget.
  6. The results feed the next batch. What wins shapes the next round of prompts, so every round starts smarter than the last.

How it runs

StepWhat happens
01. BriefPick the angles and products for the batch.
02. GenerateStatics first, then videos from the best stills.
03. ApproveI review the batch before anything goes live.
04. LaunchAds go to Meta through the API, named and structured.
05. EvaluateMeta's results decide: losers off, winners keep spending.
06. RepeatWinning angles shape the next batch.

What changed

  1. A steady flow of new statics and videos without a photo shoot.
  2. Launching a batch takes minutes instead of an afternoon of uploading.
  3. Budget moves to what sells, decided by results instead of taste.
  4. The same system now makes ads for other brands I work with, from apparel to local services.

How an AI project with me runs

01

Free call

Sixty minutes on your business. I tell you what I would automate first and what I would leave alone.

02

Map the work

I learn how your business actually runs, then rank what to automate by hours saved and risk.

03

Build one thing

We start with the system that saves the most time, with a person approving anything that touches money.

04

Hand over or run it

You get documentation and training, or I keep running and improving it monthly.

Tal Mor
Meta wants volume and variety. AI gives me both, and Meta's results decide what stays.

I run my own ecommerce brands, so every system I build has to hold up on a busy Monday, not just in a demo. I build the same kind of system for other businesses: I learn how yours works, decide what to automate, and build it.

20 years operatingAI architect for ecommerceUpwork Top Rated · 100% JSS

Questions about AI ad creative for Meta

Straight answers, from a system that runs every day.

Can AI make Meta ads that actually convert?

Yes, when they start from your real product photos and a clear angle. AI is fast at variations. The angles still come from knowing your customer.

Which AI tools make video ads?

I use several video models, including Veo, Kling and Seedance, through one API, and pick per job. Each has strengths, and the leader changes every few months.

Can ads be launched to Meta automatically?

Yes, through the Meta Marketing API. Naming, placements and testing structure are set once, and every batch follows them.

Do I still need photo shoots?

For hero brand content, sometimes. For the volume of test ads Meta needs, AI built from your product photos covers most of it.

More AI case studies

Same operator behind all of them.

⌕

An agent that reads every search term, every day

Daily search term review for a personal injury law firm, a chiropractic office, a remodeling contractor and a European fashion store. Negatives queued for approval.

Case study →
◆

Every product sorted into winners and losers

Product-level spend checked against real store sales. Winners get their own campaign, losers get excluded, and every call sits next to its numbers.

Case study →
▤

All case studies

AI systems plus Meta, Google and Klaviyo accounts, with real numbers.

See all →

Want a system like this in your business?

Sixty minutes, one-on-one, free. Bring the jobs that eat your week and I'll tell you what I'd automate first.

📅Book the free call
60 minutes, bring your numbers✉Email me
contact@talmor.ca💬WhatsApp / SMS
Prefer to message? Reach me directly
Free account review60 min · no pitch
Book →