Agentic Inspiration Ads (At Scale)

Feed it competitor ads you love. Get back hundreds of on-brand variations in minutes.

a/gentic Inspiration Ads Agent

a/gentic Inspiration Ads Agent

a/gentic Inspiration Ads Agent

a/gentic Inspiration Ads Agent

Your performance marketer finds a competitor ad that's crushing it. Screenshots it. Sends to Slack. "Can we make something like this?"

Two weeks later, your designer finally delivers 10 variations.

By then the trend moved on.

How Brands Create Inspiration Ads Today

Most designers already use AI. That's not the problem.

Here's what happens:

  1. Spot winning ad in Meta Ad Library

  2. Open ChatGPT or Gemini Nano Banana

  3. Upload the reference image

  4. Write a prompt trying to describe the style

  5. Generate. Not quite right.

  6. Tweak the prompt. Generate again.

  7. Still not right. Add more detail.

  8. Finally get something decent

  9. Now manually swap in your product photo

  10. Repeat for next product

  11. Repeat for next aspect ratio

One ad at a time. One chat session at a time. One product at a time.

The AI does the generation. But everything around it is still manual. Finding the right product photos. Writing prompts from scratch. Uploading references over and over. Copying URLs. Organizing outputs.

Your designer spends more time managing the AI workflow than doing creative work.

And here's the real issue: none of it connects. Every generation is a standalone session. No memory of your brand assets or pipeline or scale.

The brands winning on paid social test 100+ creatives per week. You can't do that one ChatGPT conversation at a time.

What We Built

The Inspiration Ads Agent.

Give it competitor ads you want to emulate. It finds the right products from your library. Creates production-ready ad images.

The Google Sheets Workflow

Here's what makes this usable for creative teams. Just a spreadsheet.

Your team adds a row to a Google Sheet:

Inspiration URL

Product Type

Prompt

Status

Generated Ad

[competitor ad link]

Vitamin C Serum

Clean, modern. Show transformation.

Pending

That's it. That's the whole workflow.

Behind the scenes:

  1. Our agent watches for new rows

  2. Pulls the inspiration image from the URL

  3. Queries your vectorized product assets to find the right product photos

  4. Triggers the Inspiration Ads Agent via dtc.sh MCP server

  5. When done, drops the generated ad URL back into the sheet

Your creative team comes back to a filled-in spreadsheet. Review. Approve. Send to Meta.

No one needs to learn new software. No one needs to understand MCP or APIs or webhooks. They add rows. They get ads.

The Vectorized Asset Library

This part is important.

Remember the Scene Prep Agent from my earlier articles? We've been helping brands vectorize their entire media libraries. Every video scene-by-scene, frame-by-frame, every product photo, lifestyle shot, before/after image gets embedded and stored in a vector database.

So when your creative team types "Vitamin C Serum" in the Google Sheet, the agent doesn't need a direct image URL. It searches your vectorized assets. Finds the best product photos. Pulls them automatically.

Your team can write things like:

  • "Summer collection hero product"

  • "Best-selling moisturizer"

  • "Before/after for anti-aging line"

The agent figures out which assets to use. Your creative team focuses on creative direction, not file management.

How The Agent Works

The agent runs on our dtc.sh MCP server. Triggered by the Google Sheet automation.

Step 1: Fetch inspiration images

Grabs the competitor ad from the URL. Studies the composition, colors, typography, lighting, overall vibe.

These are style references. The agent copies the aesthetic, not the content. Unless you want the content style also matched. Customizable through the agent’s system prompt.

Step 2: Find your brand assets

Searches your vectorized product library. Pulls the right product photos, lifestyle shots, whatever matches what the creative team requested.

Step 3: Generate the ad

Creates a new original ad that matches your inspiration aesthetic but features your actual products.

Step 4: Return results to the sheet

Generation runs async on Modal. When done, uploads to your S3 bucket. Drops the URL back into the Google Sheet row.

What You Need

  • Your product assets vectorized in Pinecone (we help with this)

  • Access to dtc.sh MCP server

  • S3 bucket for generated assets

  • The Google Sheet template connected to our agent

We set up the automation. Your team just uses the spreadsheet.

Want access to a/gentic's Inspiration Ads Agent? Reply to this email or DM Bora on LinkedIn.