Eidos Tabletop

People around a table

You sit with a group. In front of you is a large shared surface — glass, touchscreen. You discuss ideas. That surface is the product: not a private laptop per person, not a pile of side chats that die when the call ends.

Same idea on a website that feels like a multiplayer board (think Miro), on a TV in the room, or later on a real table. Start digital; Zoom and a computer are enough to use it.

While you talk

AI records the conversation and puts it back on the surface live — captions, and who is speaking. Mics on the table (and voice print) figure out who has the floor. No distracting “light up the person” gimmick; you just need speaker ID that works.

Cameras can cut like multicam TV: stay on the face of whoever is talking, jump cleanly when someone else takes the floor — not a spinning GoPro. That stream can go to people who are not in the room.

Everything said goes into a permanent record. Audio, video, photos, and the work items that come out of the talk all belong to the same meeting.

The board helps the meeting

In front of people, the system can suggest research or live data when the conversation warrants it — markets, a conflict, a government need, whatever is on the table. A human meeting manager (chief of staff, not necessarily the CEO) approves what actually gets put up so AI is not hijacking the room.

AI can answer a question if asked, but in-meeting it mostly suggests in the background and tries to keep the session moving. Live charts and “where are we going next” assessments belong here when they help.

Talk becomes work

What you discussed becomes work product: things people need to do over the next day, week, or month. You do not self-organize by typing CRUD against a task list. Changes are suggested and queued. The system groups them. People drag and reorder on the shared surface — multiplayer, like a Miro board, so one person’s drag shows for everyone.

One person can take control and maximize what they are working with; everyone else still sees the whole tabletop. After the break, each person gets a clear list of what they own and when — inbox or email, something that lands as responsibility, not a forgotten transcript.

Pen and paper stay allowed. Photograph the page; if it is taken during the meeting window it joins that person’s record for that meeting.

After the meeting

You end up with a company-wide task surface: everything people said they want to do and how they think they get there. Someone is accountable for meting work out to the right owners. Even when AI does most of the labor, a human is still responsible for each slice and uses AI to guide and watch that slice.

In the short pitch version: speak → cards appear on the digital table → people move and prioritize them → the system groups them into jobs → humans deliberate → drag cards into buckets for the computer to execute → it goes and builds what you asked for.

Why bother

In a world where AI does most of the work, people still want to lead, discuss, and debate. That is the meeting. The other side: people say things in rooms — sometimes true, sometimes not — so confidently that everyone goes along. Tabletop is meant to leave digital accountability and keep work moving, not to replace the argument.

How rooms connect

Tables (or digital rooms) can link across cities and countries. Meeting managers in another place can drive the shared surface. Pairing at the start can be as simple as a badge on a box. Surfaces upgrade over years — better screens, better AI — without throwing away the process. Ten-to-twenty-year furniture idea; same primitives.

Website for remote and Zoom-shaped use. Cast to a Roku or a touchscreen TV. Mobile and voice devices pair in. Security matters because the record is the company memory.

Who it is for

Planning and execution for a small org first; also the kind of surface a commercial real estate firm or a Fortune 500 planning cycle could run on — commercial or nonprofit, same pattern: decompose work, assign humans, let AI help execute under those humans.