How We Dogfood AI

We don’t just organize a conference.
We push AI to the limit.

We don't just put on a conference. We push AI hard in our own work — vibecoding this website, building ConferenceOS, editing videos, and creating ads. Using the technology ourselves is how we learn what helps, where it falls short, and what is worth teaching.

Try it for yourself

Learn by doing

Our work is part of the lesson.

“Dogfooding” is an awkward word for a useful habit: use the things you tell other people are worth using.

Dogfooding AI means using AI in our own day-to-day work and learning from the experience. At All Things AI, that includes building our website and ConferenceOS, editing videos, creating ads, and experimenting with browser assistants.

The useful questions get specific pretty quickly. Where does AI help? What needs another pass? Which decisions still need a person? That experience belongs in the conversation when we teach.

We teach. We train. We take the stage.

We recruit speakers. We also teach, lead training, and speak at events throughout the year. At our own conference, we spend less time on stage and give more of it to other voices.

Four places we use AI

Real work. Something you can see.

01

We vibecoded the website.

The site you are using is one of our projects. We work with AI to turn ideas into pages, navigation, and useful tools, then keep improving what people actually use.

The human job

We set the direction, review the work, test the experience, and decide what is ready to publish.

Explore something we built
02

We built our own conference software.

We used AI to build ConferenceOS, our own software for organizing conferences. It brings programs, speakers, attendees, sponsors, and event operations into one system. Building it makes the questions real: what should the software do, where can AI help, and what makes the experience better for the people using it?

The human job

We define the workflows, review the code, test the experience, and decide what is ready to use. The goal is software that serves the people behind the event and the people in the room.

Explore ConferenceOS
03

We edit videos with AI.

A useful talk should keep helping people after the room empties. AI is part of our video editing workflow as we prepare material for people to watch and share.

The human job

The speaker’s meaning is the standard. A good edit should make an idea easier to follow and preserve the context that matters.

Explore the talks
04

We create ads with AI.

We use AI in our advertising work, too. It is another place to experiment with how we communicate an idea and give people a reason to get involved.

The human job

We are responsible for the message: what we promise, how we represent the community, and what people can expect when they arrive.

See what brings us together

Try the next experiment

Bring your own AI.
Give it something useful to do.

WebMCP lets this website offer specific tools to a compatible browser assistant. Your AI can find a guide, read a useful answer, or get the published conference details with a link you can follow.

It is an experimental browser capability, and support varies. You can use the walkthrough below in any browser. Read Chrome’s WebMCP overview.

All Things AI · WebMCP walkthrough

Interactive preview

1. Choose a task

2. See the request

Use this page’s WebMCP tools to find the published All Things AI conference dates and venue, then show me the official event link so I can check availability or get updates.

The button runs the website’s tool directly and shows its current data. To try WebMCP itself, use the copied prompt with a compatible browser assistant while this page is open.

3. Follow the result

Run an example to see useful information and a link you can follow. You stay in control of what happens next.

Open this page

Use a browser with an assistant that supports WebMCP. Availability depends on both.

Copy a prompt

Ask your assistant to use the tools on this page. Choose a task you actually care about.

Follow the source

Check the answer, open the guide, or continue to the event page. See whether it helps.

Next experiment · Event signup

From “that looks interesting”
to “I’ll be there.”

We want you to be able to use your own AI to find an All Things AI event and help you sign up. That is a useful everyday task to put this technology to work on.

Today, this demo finds the published conference details and gives you the official event link. The 2027 event page currently offers save-the-date updates. Connecting registration to the assistant is the next experiment.

Explore events.allthingsai.org

The experience we want to build

  1. 1

    Find your event

    Tell your AI what you want to learn and when you can attend.

  2. 2

    Review the details

    See the date, location, availability, and any cost before deciding.

  3. 3

    Choose to sign up

    Review the information being shared and confirm the final action.

  4. 4

    Know you’re registered

    Get a clear confirmation from the event system.

Keep your judgment in the work

The responsibility stays with us.

We choose what to build, what to say, and what to publish. We bring the purpose, the context, and the judgment about whether the result belongs in front of our community.

That is a useful habit to bring to your own work: pick a real task, use the tools, examine the result, and decide what deserves to move forward.

Build your own AI skills

A few practical questions

What does dogfooding AI mean?

Dogfooding AI means using AI in our own day-to-day work and learning from the experience. At All Things AI, that includes building our website and ConferenceOS, editing videos, creating ads, and experimenting with browser assistants. It gives us practical experience to bring back to the community.

How does All Things AI share what it learns?

We recruit speakers. We also teach, lead training, and speak at events throughout the year. At our own conference, we spend less time on stage and give more of it to other voices.

What do you mean when you say you vibecoded the website?

We worked with AI to turn our ideas into website code and refine the result. People still set the direction, review the changes, test the experience, and choose what to publish. The working site is the result of that collaboration.

What is ConferenceOS?

ConferenceOS is our own conference operating system, built with AI to bring programs, speakers, attendees, sponsors, communications, and event operations together. It is designed for one organization to manage multiple events, with people and authorized AI agents working together. We are still developing its agent capabilities, and people remain responsible for permissions, approvals, and what is ready to use.

What is WebMCP?

WebMCP is an experimental browser API that lets a website expose specific tools to compatible browser assistants. On this site, those tools can search public content, read speaker FAQs, find coaching downloads, retrieve conference details, and open public pages. Browser and assistant support varies.

Can I try the demo with my own AI?

Yes, if your browser assistant supports this site’s WebMCP tools. Open this page in that browser and ask your assistant to use the tools available on the page. The walkthrough buttons also work without an AI assistant: they run the same website tool functions and show the returned information. They are an interactive preview, not an AI conversation.

Can my AI sign me up for an All Things AI event?

This website’s WebMCP tools can find published conference details and the official event link. They do not complete registration. Check current availability and complete signup on events.allthingsai.org. A connected event-registration experience is a next experiment we want to explore.

Who is responsible for work created with AI?

We are. Using AI does not change our responsibility for what we publish. Our approach is to set a clear purpose, review the result, test it where appropriate, and improve it with feedback. Our experience is something to learn from; it is not a promise that the same approach will work for every team.

Updated . Our uses of AI are described by the All Things AI team. For the browser technology, see the WebMCP specification.