Skip to content
AI in practice

Why Don’t You Have a Custom AI Yet? — Short Version

Most people are still using AI like a search box. I think they’re missing the most interesting part.

By Brett Dyess

Most people are still using AI like a chatbot.

Ask a question. Get an answer. Close the window.

But what happens when you stop starting over?

Over time, I’ve been building a customized AI assistant I call Jarvina.

And “building” may actually be the wrong word.

Some of it was deliberately configured. But much of it developed through use.

Jarvina learned how I write. How I work. My businesses and projects. My shorthand and terrible typing. My sense of humor. The tools I use. The things I’m likely to forget. Even the completely unrelated things I care about outside of work.

Essentially, you’re creating a little mini-me—except one that can research, analyze, organize, remember context and help execute.

The interesting part is that you can’t completely pre-program that.

You can pre-program the foundation. You can’t completely pre-program the relationship.

I’ve started thinking about AI customization as seven levels:

Level 1 — Chatbot
Ask questions and generate content.

Level 2 — Personalized AI
It begins understanding your preferences, language and context.

Level 3 — Working Assistant
It helps manage recurring work, projects and workflows.

Level 4 — Connected Assistant
It connects to the applications and information you actually use.

Level 5 — Operational AI
It begins helping execute work across those systems.

Level 6 — Collaborative AI
It becomes an ongoing partner in planning, meetings, decisions and problem-solving.

Level 7 — Embodied AI
The intelligence starts moving beyond the chat window into dedicated hardware, voice, cameras, workstations and mobile devices.

Jarvina now has a dedicated workstation, and we’re experimenting with voice, hardware, visual indicators and eventually a mobile form.

The bigger lesson isn't that everyone needs to build an AI workstation.

It's that AI becomes dramatically more useful when it stops being a tool you occasionally visit and starts becoming an intelligence that develops alongside the way you actually work.

The question may no longer be:

“Which AI should I use?”

It may be:

“Why am I still using someone else’s generic AI?”

I’m apparently on Level 7.

We’re currently designing the hardware.

Things have escalated.