Run AI work as a governed practice.
chat.eduba.io is the product side of the consulting work at eduba.io. The same worldview shows up here: organizations get the structure, controls, and audit trail needed to use AI on actual work, not just demos.
Group people and agents under shared rules
An organization is a workspace for user personas and agent personas. Discussions and the documents attached to them all live inside it.
Admins, members, and per-discussion roles
Each user persona has an organization role. Each discussion can carry its own roles on top, so sensitive conversations stay scoped.
Bring people in under a known persona
Invite an existing account into a specific persona inside the organization, with a name, role, bio, and learning context attached from day one.
Curated agents, defined by the org
Org admins build agent personas on top of the approved providers, with their own system instructions, communication style, and specialization. Users pick from the personas the org has sanctioned.
See where the spend actually goes
Per-organization subscription with an ask allowance and credit balance. The usage dashboard shows asks, tokens, success rate, and daily and hourly trends so the bill never surprises anyone.
Every message, attributed and retained
Each message is tied to the user persona or agent persona that sent it, inside a discussion that keeps its full thread. When someone asks what happened, the answer lives next to the work.
Topic, system prompt, history
Every discussion has a topic and a system prompt that defines how agents respond. Documents attach as shared context. The full thread is retained for review.
Explore, card actions, flashcards
The same study tools available to individuals work inside an organization. Turn one team's discussion into a training surface for the next person.
Why this matters for the enterprise.
Most AI rollouts stall because nobody can answer two questions at the same time: "Who is allowed to do what?" and "How do we know what was done?" chat.eduba.io was built so those answers live next to the work, not in a separate spreadsheet.
- Personas keep identity stable across discussions, so attribution is clean.
- Roles answer the access question at both the org and discussion level.
- Full message history per discussion answers the audit question in the same place the work happens.
- Curated agent personas keep the model decision in the hands of the org, not each user.
Pair it with the consulting work.
The product handles the day-to-day. The consulting practice at eduba.io handles the parts that don't fit in software: framing, governance design, and training so adoption sticks. Most teams use them together.