Portnoy Production - a company we founded - works in creative production. This means I have spent years working inside the very processes that people eventually decide to automate: briefs in five different formats, files getting lost in chats, approvals depending on one person, working with contractors, tight deadlines, and spreadsheets that quietly become the foundation of the entire infrastructure.

Now I apply this experience to AI and automation. It all starts with analyzing my own actions and those of the team, identifying “time leaks,” and determining which tasks still require a human. Then - naturally, the classic move - a small MVP. Depending on the task, this may involve Make, n8n, APIs, Telegram bots, document processing systems, AI models, or a small piece of custom code - each has its place.

It is important that the system withstands the test of real work. Whether it is used individually or by a team depends on its purpose. This means having logs, failure-handling scenarios, and manual approval stages, particularly where the cost of an error is high. If the pilot works, we scale it. If it does not, it is better to find out at an early stage.

I have experience with, and a particular affection for, document workflows, internal assistants, production pipelines, and other areas that, in an ordinary setting, are so fond of taking time away from creative work

Machines can do a lot. I support sensible collaboration and moderate delegation. Together, we can do more.