The opportunityWhat I builtThe experienceThe impact

Winning Ads Hunter

AI creative research
2026

My ad research method, built into an app. I turn Meta and TikTok research into full scripts, avatar casting and creative briefs ready to produce.

The finished hunt: 321 unique ads, 20 winners and the research flow behind them.

#The opportunity

My best creative decisions came from precise, hands-on research: finding the right competitors, watching their videos, comparing the message and delivery, then deciding what to test. Broad searches brought back too much irrelevant material. I wanted research tailored to my niche, market and campaign, following the same method I used manually, with the capacity to examine a whole market in one run.

#What I built

  • I built Winning Ads Hunter around my own research process, from the search strategy to the creative decisions at the end.
  • It gathers ads and UGC from Meta and TikTok, removes duplicates, checks relevance and brings the strongest creative into one workspace.
  • It identifies winning ads and the patterns behind them, so I can see which creative deserves a closer look and what to test next.
  • The Creative Playbook turns the research into full ad scripts, opening hooks, messaging angles, structures, calls to action and briefs ready for production.
  • The Best Avatar section recommends who to cast, how they should speak and where to film, with avatar prompts and voice direction I can use in AI UGC production.

#The experience

I start with the niche, target market and language. The app searches for the competitors and creative that actually fit the campaign, then works through the collection in batches.

Explore the results and identify winners by score, format and market.

It examines the message, video structure, offer, presenter and delivery. Comparing the strongest ads with the wider set helps me see which creative choices deserve a test. The recommendations stay connected to the ads behind them.

Winning hooks, linked to the ads and delivery patterns behind them.

For life insurance, the playbook gave me a clear first casting direction: an everyday young parent explaining the offer plainly from a couch. It also proposed a French-market alternative, comparing options at a kitchen table with a laptop. Each direction came with an avatar prompt and a voice brief.

Choose the best avatar from presenter traits and the winning ads that use them.

I use the research myself. The playbook gives me the full script, the person delivering it, the production brief and the next tests to run. I take that direction into production and put the findings to work in my campaigns.

#The impact

8 h → 90 min8 h → 90 minFrom manual market research to a creative playbook ready for production
1,6681,668Ads collected in one auto insurance hunt, reduced to 556 unique videos for analysis
−73%−73%Cost per lead (also reflects AI agents, better landing pages and the other projects in this portfolio)
Project details
Year
20262026
Category
AI creative researchAI creative research
Dimensions
Characters
2,7922,792
Tags
Ad research, Creative strategy, AI analysisAd research, Creative strategy, AI analysis
Role
Research strategy, product design, development and campaign applicationResearch strategy, product design, development and campaign application
Stack
Python, FastAPI, Vite, Apify, ScrapeCreatorsPython, FastAPI, Vite, Apify, ScrapeCreators
Links
HomeHomeContactContact
EmailEmailLinkedInLinkedIn
AI sales intelligence
Next projectSales Coach2026
Ismail Ismail Sossi AlaouiSossi Alaoui
ContactContactTHEME[D]THEME[D]SOUND[|]SOUND[|]
--:--GMT MA --:--0720 X 0450 Y

[T] Back to top · [B] Go to end

[L] Light · [D] Dark · [A] System

[S] Toggle music