A curated guide and collection of resources for building, prompting, and integrating high-impact Custom GPTs, with a special focus on advertising, marketing, and creative workflows. As a Director of Creative Team working at the intersection of technology and advertising, I created this playbook to bridge the gap between AI's potential and practical, real-world creative application.
- 🌟 Inspiring Custom GPTs
- 🛠️ Prompt Engineering Techniques
- 🔌 Action & API Integration Guides
- 🎯 GPT Strategies for Marketing & Advertising
- 🔧 Tools & Helper Libraries
- 📄 Must-Read Articles & Whitepapers
- 🤝 How to Contribute
- 📜 License
A collection of powerful and creative GPTs that showcase what's possible.
- Educational Bedtime Stories - My Take: A personal project I designed and developed to transform children's bedtime into an imaginative learning adventure. It's an exploration of how Generative AI can be used for positive, ethical, and creative storytelling for families.
- Unique History Explorer - My Take: Another personal project that combines historical storytelling with AI-driven visual creation. This GPT makes history accessible and engaging by sharing lesser-known facts accompanied by hyper-realistic images, showcasing the potential of AI in digital humanities and content creation.
- Image to Text Explainer - My Take: I developed this 'Image to Text' GPT as a versatile utility for digital professionals. Its primary function is to create clear, objective descriptions for crucial tasks like generating SEO-friendly alt-text to improve web accessibility (a11y), or for analytically deconstructing visual content in creative projects. It's a practical demonstration of applying multi-modal AI to enhance both digital inclusivity and workflow efficiency.
- My Agency's Internal GPT - My Take: The core tool I developed to automate our creative briefing process, directly resulting in a 30% reduction in project turnaround time. A full case study is available within my evidence portfolio.
Resources to master the art of talking to machines.
- Awesome Prompting Techniques on GitHub - My Take: A concise reference of prompt patterns, including few-shot, chain-of-thought, and iterative prompting.
- Prompt Engineering Guide by OpenAI - My Take: Official best practices covering system messages, role separation, and context management.
Guides on connecting your GPTs to the outside world to perform real tasks.
- Zapier + OpenAI Integration Guide - My Take: Step-by-step automation for sending GPT-generated outputs into Slack, Gmail, or other apps.
- Building Actions for Custom GPTs by OpenAI - My Take: Detailed walkthrough on defining, hosting, and deploying custom actions.
Actionable strategies and prompt templates tailored for the creative industry. This section is my special contribution.
- Strategy: Persona-Based Ad Copy Generation
- Description: How to instruct a GPT to adopt different customer personas for targeted messaging.
- Prompt Template:
Act as a [Persona Name], a [demographic description]. Your goal is to react to the following product: [Product Description]. Write three pieces of ad copy that would convince you to click.
- Strategy: A/B Creative Variant Generation
- Description: Use a GPT to produce multiple ad design briefs and copy variations for systematic A/B testing.
- Prompt Template:
Given the campaign goal: [Goal], create 4 distinct ad concepts with a headline, body text, and visual mood description for each.
Tools that make the GPT creation process easier.
- LangChain - My Take: Simplifies building multi-step GPT workflows with chains, agents, and memory persistence.
- GPTCache - My Take: Local caching layer to reduce API calls and speed up repetitive queries.
Seminal texts and recent analyses on the state of Generative AI.
- Attention Is All You Need (Google Research) - My Take: The foundational Transformer paper; understanding it is key to grasping GPT capabilities.
- The State of GPT-4 by OpenAI - My Take: OpenAI's overview of GPT-4 strengths, limitations, and future directions.
Contributions are welcome! Please read the contributing guidelines first. Feel free to open issues or submit pull requests to add new GPT examples, resources, and best practices.
This project is licensed under the MIT License.