Abstract

BrochureBot is an AI-driven tool designed to automate the creation of professional brochures. By leveraging advanced language models such as GPT-4, DeepSeek, and Ollama, it extracts and formats relevant content from web sources, enabling businesses, investors, and recruiters to generate structured brochures effortlessly. This document outlines the system's architecture, methodology, experimental setup, and results.

Introduction

Creating brochures manually can be time-consuming and resource-intensive. BrochureBot automates this process by using AI to generate content, retrieve information from websites, and format it into a structured document. It supports multiple LLM providers, flexible API integrations, and robust error handling to ensure seamless operation.

Methodology

The system is composed of several key modules:

ai-brochure-bot/
│-- summarizer/
│ │-- init.py
│ │-- fetcher.py # Web content fetching logic
│ │-- summarizer.py # Main summarization logic
│ │-- brochure.py # Brochure generation logic
│ │-- llm_handler.py # Generic LLM handling logic
│-- utils/
│ │-- init.py
│ │-- config.py # Environment configuration
│-- main.py # Entry point of the app
│-- .env # Environment variables
│-- requirements.txt # Python dependencies
│-- README.md # Project documentation

AI Model Integration

BrochureBot allows users to choose from different AI models:

• GPT-4 (via OpenAI API)

• DeepSeek (via DeepSeek API)

• Ollama (via API or local library)

• Data Extraction

It fetches relevant data from company websites using Python’s requests and BeautifulSoup libraries. Example:

import requests
from bs4 import BeautifulSoup

def fetch_web_content(url):
response = requests.get(url)
if response.status_code == 200:
soup = BeautifulSoup(response.text, 'html.parser')
return soup.get_text()
return None

Experiments

We evaluated BrochureBot’s performance using different LLMs and company websites. Test cases included:

• Generating brochures for tech startups, enterprises, and academic institutions.

• Comparing output coherence across different AI models.

• Measuring execution time and API response latency.

Sample Prompt & Output

User Input:

Enter company name: ABC Corp
Enter website URL: https://abc-corp.com
Choose LLM model: gpt-4
Select provider: openai

Generated Output:

ABC Corp Brochure

Overview

ABC Corp specializes in AI-driven automation solutions for businesses worldwide.

Culture

Innovation, inclusivity, and cutting-edge research define our workplace.

Customers

We serve global enterprises, government agencies, and SMEs.

Career Opportunities

Join us to work on AI, cloud computing, and automation technologies.

Results

Our experiments demonstrated that:

• GPT-4 provided the most detailed and structured content.

• DeepSeek was faster but required additional fine-tuning for coherence.

• Ollama’s local execution improved privacy but had limited model capacity.

• The system successfully extracted data from over 90% of tested websites

Conclusion

BrochureBot effectively streamlines the brochure creation process, making it a valuable tool for businesses and recruiters. Future improvements include:

• Expanding LLM support for multi-language brochures.

• Enhancing content personalization using company-specific data.

• Integrating more advanced AI summarization techniques.