# Flyfish: AI-made Facebook ad concepts

I broke down 16 proven B2B Facebook ads, then used AI to make 14 new Flyfish ad concepts ready to test.

- Industry: B2B SaaS
- Services: Paid media, Marketing Strategy
- My role: Researched, briefed and made the test ads
- Audience: B2B sales teams
- Tools used: ChatGPT, Gemini, Nano Banana 2, Meta Ads
- Case study page: https://dilosan.com/work/fly-fish/

## The brand

Flyfish gives B2B sales teams full account intel in one place, so reps know their buyers before they reach out.

## Why

Flyfish needed fresh Facebook ads to test. Instead of guessing, I started from ads that already work for other B2B brands.

## How

1. I researched ads that already work, using competitor analysis tools, the Facebook Ad Library and the LinkedIn Ad Library, and picked 16 from brands like Cognism, PayFit, GoCardless and Directive.
2. For each one, I wrote down the visual, the story, the core message and the conversion tactic, like loss aversion or peer proof.
3. I turned each pattern into a Flyfish concept: chat screenshots, memes, before and after, and bold claims.
4. I made the ads with AI tools and built them into a test set for Facebook.

## Step by step

1. **Research what already works.** I searched competitor tools, the Facebook Ad Library and the LinkedIn Ad Library, and saved 16 proven B2B ads.
2. **Break each ad down.** For every ad I wrote down the visual, the story, the message and the tactic behind it.
3. **Remake it for Flyfish.** Each pattern became a Flyfish ad, made with AI and checked by hand.
4. **A test set for Facebook.** 14 ads, ready to test against each other.

## What happened

- 16 proven ads broken down
- 14 Flyfish ads ready to test

## What I learned

- Breaking down why an ad works makes the brief better than starting from a blank page.
- AI images still need a careful check. One draft kept another brand’s product name, so every ad gets a review before testing.

Work by Dilosan Ravindran. Contact: hey@dilosan.com