AI Ad analyser agent

AI-powered ad analysis workflow that uses behavioral science to evaluate and improve visual ad performance with actionable insights.

Playground API Pixelflow

If you're looking for an API, here is a sample code in NodeJS to help you out.

const axios = require('axios');
   
   const api_key = "YOUR API KEY";
   const url = "https://api.segmind.com/workflows/6808fa177cbc8be62975d036-v1";
   const data = {
     Ad_Image: "publicly accessible image link"
   };
    
   axios.post(url, data, {
     headers: {
       'x-api-key': api_key,
       'Content-Type': 'application/json'
     }
   }).then((response) => {
     console.log(response.data);
   });

Response

application/json

{
  "poll_url": "<base_url>/requests/<some_request_id>",
  "request_id": "some_request_id",
  "status": "QUEUED"
}

You can poll the above link to get the status and output of your request.

Response

application/json

{
  "Agent_1": "any user input string",
  "Agent_2": "any user input string"
}

Attributes

Ad_Imageimage*

To keep track of your credit usage, you can inspect the response headers of each API call. The x-remaining-credits property will indicate the number of remaining credits in your account. Ensure you monitor this value to avoid any disruptions in your API usage.

AI Ad Analyser Agent

This workflow is a behavioral ad analysis system designed to evaluate creative assets—especially image-based ads—using psychological and persuasive frameworks. It leverages multiple LLMs (like GPT-4o and Qwen2) to assess how well an ad communicates, resonates emotionally, and aligns with cultural norms. The goal is to provide actionable feedback to improve ad effectiveness based on top-tier consumer behavior insights from experts like William Leach, Rory Sutherland, and Robert Cialdini.

How It Works

Input Components:
- Text Module: Provides the behavioral science criteria used to evaluate the ad. It includes key questions about typography, color psychology, persuasion tactics, and cultural fit.
- Image Input: The ad visual (e.g., “Nothing Sucks Like an Electrolux”) is uploaded for multimodal analysis.
Model Prompts:
- Two advanced language models (GPT-4o and Qwen2) are prompted with the same context and image, each acting as a behavioral science expert. This redundancy allows for diverse perspectives and richer insights.
Model Outputs:
- Each model independently analyzes the ad across the five criteria and suggests “Fixes That Matter”—concrete, behaviorally-informed improvements to enhance the ad’s performance.
Comparative Analysis:
- Outputs from both models are reviewed side by side, offering cross-validation of suggestions and highlighting any nuanced or differing viewpoints on effectiveness and risk areas (e.g., cultural misinterpretation of humor).

How to Customize This Workflow

This workflow is highly modular and can be tailored to different use cases:

Ad Types: Swap in different images—digital banners, social ads, product thumbnails—to test for platform-specific engagement.
Industries: Adjust the evaluation criteria (e.g., add “Fear Appeals” for health ads or “Luxury Cues” for premium brands).
Regions: Use culturally-specific image variants and enable localization models to assess global resonance.
Languages: Update prompts and evaluation criteria for ads in non-English markets to maintain linguistic and cultural relevance.

This is an ideal workflow for creative agencies, performance marketers, or UX copywriters looking to improve visual ads using the rigor of behavioral science and the speed of AI.

Models Used in the Pixelflow

qwen2-vl-72b-instruct

Qwen2-VL-72B-Instruct is a state-of-the-art multimodal model excelling in image and video understanding, with advanced capabilities for text-based interaction.

gpt-4o

GPT-4o (“o” for “omni”) is our most advanced model. It is multimodal (accepting text or image inputs and outputting text), and it has the same high intelligence as GPT-4 Turbo but is much more efficient—it generates text 2x faster and is 50% cheaper. Additionally, GPT-4o has the best vision and performance across non-English languages of any of our models. GPT-4o is available in the OpenAI API to paying customers.