Have you seen those videos on YouTube where people show off basic AI agents in n8n and claim it’s a J.A.R.V.I.S.? It’s not initiating dialogue. It doesn’t remember what happened before. It can’t run autonomously or actually help you while you’re sleeping. It’s a piece of C.R.A.P. I’m pretty tired of these videos and want to live in a world where a real JARVIS exists, so I’ve decided to show these underachievers how it’s done. This week, I’m going to build a real JARVIS assistant in public. It will be able to: - Access your database to take notes, record information, and more. - Do research and generate reports using multiple sources, including your own data. - Initiate dialogue to notify you about events, tasks, habits, unanswered emails, and more. - Remember your conversations, take notes, and help you analyze your growth and changes. - Integrate with any kind of MCP to expand its abilities. - ...and of course recognise your voice and answer you in multiple modes (audio, text, reports, graphs and more) If you have any ideas about how it could help you in your daily routine, please leave a comment - I might add it to the bot. In next 10 days, I’ll share a full overview of this agent and you’ll be able to copy and use these assistants for your own needs.

Here is an example of an n8n automation I built that scrapes an RSS Feed for high quality AI content (whitepapers, research papers) and calls AutoContentAPI via HTTP Request and generates editable podcasts and distributes them to me via multiple channels (Gmail, Telegram, and Google Drive) on a weekly basis. In short, audio learning is the most digestible format for me. This automation helps me stay up to date with high quality AI content every week without having to search for it. I was in search of a NotebookLM API for a while now, but it's still unclear as to whether Google will further pursue the project, so that means no API for now. Feel free to DM me for access or join me in my free Skool community at Seamless-AI where I'll be providing the json :)

Today I want to share an incredible story from our community member, @Aleksandr Maznik A month ago, Alexander had zero experience with AI or automation. But he had a powerful reason to learn: his company helps children worldwide receive life-changing cerebral palsy treatment in China. More efficient processes mean more kids getting the help they need. His biggest challenge? Answering hundreds of questions from parents about logistics—transfers, payments, documents, accommodation, and treatment details. He set an ambitious goal: automate these interactions using AI. Here’s what Alexander built in just 1 month, starting from zero: - An AI-powered Telegram agent built with N8N. - The agent accurately answers questions based on a knowledge base of over 100 detailed Q&A entries. - It remembers conversation history, providing a seamless chat experience. - Supports multiple languages, essential for Alexander’s global client base. Alexander launched the agent into production, and it’s already being used daily by multiple managers. Next, he plans to automate document generation and further expand AI support directly to clients. We’re proud our community empowers people like Alexander to create genuine, real-world impact.

Yesterday I watched a video from AI Engineer Summit (if you haven't seen it, there are 3 dozen top AI talks). At one of the presentations, the speaker showed an interesting approach to working with AI agents. The agent gets access to the necessary tools (database, API services, etc.) and processes a user request as input. Using these tools, the agent builds an interface in real time that matches the specific task. An example was shown on the screen - a simple email interface. The user interacts with this interface - for example, clicks on an email and the agent receives data about this action. It then analyzes what the user is likely to want to see next and updates the interface, tailoring it to the next step - in this case displaying the contents of the selected email. Thus, with access to the data and the context of the interaction, the agent is able to shape interfaces on the fly to the user's current need. Of course, this approach has limitations: everything works quite slowly now, the interfaces are simple, and the result can change from request to request. But if we introduce a system of rules and speed up the AI's reaction, then in the future we can get really adaptive interfaces that will automatically adjust to the user's tasks without unnecessary tinsel. Watch the presentation here

https://youtu.be/C-QA4U2egGY?si=dBUEnnyM7l8bUGtv templates here: https://www.skool.com/aiworkshop/about?ref=bd01ae9e365f4c81aff9d57b53bb6d42