AI Agent for SaaS Content Ideation

01 — Project details

Course: Hugging Face “AI Agents”
Tools: Python, smol-ai/smol-dev, Hugging Face Spaces, Gradio, YAML
Focus: SEO Titles & Campaign Prompts for B2B Marketing

02 - Project overview

As part of Hugging Face’s agent-building course, I developed a functional AI agent capable of generating SEO-optimized blog titles and marketing campaign prompts for SaaS and B2B contexts. This wasn’t just a plug-and-play tool — I actively worked with Python code, modular agent frameworks (smol-dev), prompt templates, and runtime debugging to deploy a working, shareable Hugging Face Space.

03 — Purpose & use case

The agent was built to assist content and marketing teams during the ideation stage, where speed and variation are critical. It supports creative generation of:

  • Blog post headlines optimized for SEO

  • Messaging angles for product campaigns

  • Topic suggestions tailored by tone and audience

For example, when given a topic like “AI in customer support”, the agent can return campaign hooks like:

  • “From Chatbots to Revenue: AI’s New Role in CX”

  • “Smarter Support Starts Here: AI Solutions for Scaling Teams”

  • “Your Agents Deserve an AI Upgrade — Here’s Why”

04 — The result

A custom Python tool using TextTool that dynamically generates 3 marketing prompts based on user input

  • A reusable prompt template supporting parameters for topic, audience, and tone

  • Logic for the agent to run a think→act→observe loop and select the final response

  • A fully functional Hugging Face Space with API integration, agent orchestration, and Gradio UI

  • Runtime fixes (e.g. token handling, imports, YAML loading, indentation errors) and prompt tuning

See the live agent in action