Scaling AI inside a team rarely fails where people expect it to. Every agent ships fine. The breakage shows up between them.
The screener flags a candidate as a strong match, and the outreach agent emails them a generic “thanks for applying” template because nobody passed the screener’s note over. The prospecting agent adds a contact the CRM updater added last week, because each one keeps its own list. The hiring criteria you “fixed” last quarter live in five places, and the fix only landed in three.
You haven’t built a system. You’ve built a collection.
Spaces are how that changes.
What a Space Is
A Space is a shared environment inside Theona where every agent that covers a function lives together. Your HR Space has the agent that screens resumes, the one that handles candidate communication, and the one that tracks where everyone is in the pipeline. Your Sales Space has the prospecting agent, the CRM updater, and the outreach writer. They share the same documents. They know about each other. They run as a system.
When the process changes, you update the relevant agent. Everyone using that Space gets the updated version immediately. The process stops living in someone’s head or in a doc nobody reads. It lives in the Space, and it’s always current.
How You Build One
You name the Space (“Hiring”, “Sales Ops”, “Client Onboarding”) and the Architect comes back with a proposal. It draws on your role, the tools you’ve connected, the agents you’ve already built, and the patterns it has seen on your team. You start from its first draft, not from a blank page.
From there, you refine together. Walk through the plan, change scope, cut a step, add something it missed, describe the parts of the function it didn’t get quite right. Only after you confirm does it create the agents.
You don’t configure the graph. It reflects the function as the Architect proposed it and as you refined it.
What Lives Inside
Every Space has its Sources: a shared knowledge base where the function’s context lives. It holds the hiring rubric, the pricing sheet, the company tone of voice, the policies and definitions the team has settled on after months of working through them. The agents read from it, the team reads from it, the Architect reads from it when proposing new agents, and Theona reads from it when routing tasks. Update one source and the change shows up everywhere the function runs.
You can also see the shape of the function. Every Space draws itself as a map: a card for each agent, with the connections between them (parents to sub-agents, hand-offs across the steps of the process) laid out automatically. The team can read the process at a glance.
The Space tracks Last activity. Come back to your HR Space on Monday and the first thing you see is the work that ran while you were away: which candidates moved through screening, which outreach went out, what Theona handled on its own.
When you add a new agent to an existing Space, it already knows what’s there. The Architect won’t propose a role that duplicates something you’ve built before. It builds into the Space, not on top of it.
How You Work in a Space
Once a Space has agents in it, you don’t open them one by one to work. You open the Space’s chat and describe what you need (“look at the three new applicants in Greenhouse and tell me which ones to talk to”). Theona figures out which of the agents already in the Space are right for the job. If the work crosses several of them (screen the candidate, enrich the profile, draft outreach), it lays out the plan first, then runs it once you confirm.
Some work moves without a prompt at all. Spaces run on triggers: a new candidate landing in the ATS, an inbound lead hitting the CRM, a weekly schedule firing on Friday morning. The Space watches for the event, picks up the work, and posts the result back into the chat for you to review.
You’re talking to the function, and Theona handles the rest.
Sharing a Space
A Space is yours until you decide otherwise. You can build and iterate in private. When the process is solid, you share the whole Space in one action, and your team gets access to every agent inside it, fully configured.
Sharing brings the team onto the same process: one workflow, one knowledge base, one history of what’s been decided, one definition of what “done” looks like. Updates to any agent land for everyone immediately. The Space becomes the source of record for how the function runs, and the team works from it instead of from a stale doc.
API keys and credentials don’t transfer when you share. Each person connects their own accounts. The agent configuration is shared. Access to external systems stays personal.
Sharing requires a Team plan. Personal use of Spaces works on any plan.
Who This Is For
A recruiting team running the same hiring process slightly differently across hiring managers consolidates it into one HR Space. The criteria, the scoring rubric, the outreach templates live there, where everyone reads from the same versions. When the hiring bar shifts, it shifts in one place for everyone. New hires onboard by joining the Space, which carries the current process and its history with it.
A company without a dedicated HR hire can run its full recruiting process from a single HR Space: sourcing, screening, scheduling, candidate communication, and pipeline tracking, without adding headcount.
A consultant or fractional operator running work across multiple clients keeps a separate Space for each engagement. Every Space is configured for that client’s tools, criteria, and workflow. Switching between clients means switching between Spaces.
A team lead tired of rebuilding context every time they open Theona builds a Space for their function once. The agents, documents, and history are all there. Next session picks up where the last one ended.
A Space Is Not a Folder
This distinction matters because Theona has always had a way to organize agents. A Space isn’t a renamed folder.
Agents in a folder are independent. They don’t share context or files. Moving agents into a folder doesn’t change how they work; it changes where they appear in your list.
Agents in a Space are built as a system from the start. They share a knowledge base. The Architect understands what’s already in the Space when you add something new. Every change one agent produces is visible to the whole system, not locked inside a single session.
The folder is a label on a list. The Space is the process itself.
Where to Start
Pick one function your team handles manually, or partially automates today. Hiring is a common first choice. So is sales outreach, weekly reporting, or client onboarding.
Create a Space for it. Name it after the function. Tell the Architect what it covers. You’ll have a working map of agents, a shared knowledge base, and a chat tied to the whole system in a few minutes.
That’s the Space. From there, it runs.
Create your first Space →