Daily Automations with ChatGPT Projects and Custom GPTs

How I use AI every day to work smarter and faster Since I started working with artificial intelligence, I stopped seeing it as just a tool. Today, ChatGPT is my digital colleague: it executes tasks, proposes solutions, writes, replies, analyzes. But not alone. It works within an ecosystem I built using two powerful tools: Projects and Custom GPTs.

ChatGPT Projects: my automated dashboard for recurring tasks

What Projects are and why I use them ChatGPT Projects (available in the paid plan) are organized spaces where I collect documents, prompts, instructions, and chat history for each specific task. They’re ideal for ongoing and repetitive activities. More details in the official OpenAI guide. How I organize my workflows Each Project has a clear name, description, and consistent structure. Examples of my workspaces include:

SEO & Copywriting: article generation, title suggestions, readability analysis
Customer Communication: email replies, FAQs, automated templates
AI Training: lesson scripts, slides, quizzes, real-world examples
Technical Projects: Laravel logic, AWS deployment, VPN and network management

Real-life automation examples

Weekly generation of personalized article ideas for Medium
SEO analysis of drafts and improvement suggestions
Automated generation of formatted emails for clients and suppliers
Editorial schedule creation and marketing plans

Custom GPTs: tailored assistants for each business area How I create a custom GPT Just click on “Explore GPTs” and create a new assistant with a well-built system prompt. I customize it with detailed instructions, tone of voice, examples, and scope limitations. Key instructions I set

Clear role (e.g., “You are an expert in Teltonika RUT955”)
Response style (formal, concise, technical)
Preferred output format (bullet points, code, examples)
Usage context (Laravel dev, Ubuntu admin, SME consulting, etc.)

Everyday practical examples

Maintenance GPT: generates optimized routes based on schedule, priority, and location
AI Training GPT: provides prompts, use cases, and commented slides
Tech GPT: assists with Teltonika, OpenVPN, Livewire, Laravel, AWS Lambda
Marketing GPT: analyzes copy tone, CTA strength, UX wording, and funnels

Strategies and best practices

Consistent naming: each GPT has a prefix tied to its function (e.g., “GPT-Tech”, “GPT-SEO”)
Solid system prompts: better one long prompt upfront than many adjustments later
Project vs GPT: Projects = memory + flow; GPT = skilled executor
Sharing and testing: I share GPTs with clients so they can test and use them independently

Automations with APIs and external tools AI on its own is powerful, but when integrated it’s transformative. Here are some connections I use:

Google Sheets API to send data to GPT and collect results
Zapier to trigger GPT prompts on incoming emails
Python scripts on AWS Lambda to read RSS feeds and send them to GPT
Amazon Polly to convert GPT outputs into MP3s
AzuraCast + SFTP to broadcast those audios on webradio

Limits, risks, and opportunities

I always test GPTs internally before production use
I monitor results for 2 weeks before full delegation
I use a checklist to assess every GPT’s weaknesses

Conclusion: working with AI, not replaced by it ChatGPT isn’t here to replace us — it’s here to amplify what we can do. When we train it well, it becomes a reliable and daily ally. You don’t need 10 GPTs or 5 Projects to start. Just one. But make it useful, well-designed, and actually used every day. And you? Are you already using custom GPTs or Projects in your workflow? Want some suggestions to improve your productivity? Let me know in the comments.

Read the full article here: https://medium.com/@andreatadioli/daily-automations-with-chatgpt-projects-and-custom-gpts-7da7243a2bc8