8Hats Lab products

Three products. One approach to how your people and AI agents work together.

The work your team does with AI agents still has to be checked, repeated, and defended. 8Hats Lab researches, designs, and builds how organizations of humans and AI agents work, and the three products below apply one approach to that problem: shared capabilities, workflows, knowledge, and rules — with a named person owning every decision that matters. This page explains what each product does, how they relate, and where an organization should start.

The common approach

When AI agents do part of the work, five things have to be shared.

Most processes and management frameworks were designed for people only. The approach behind every 8Hats product adapts them so the work still holds when AI agents join.

People and AI agents, one working system.

Capabilities
What each AI agent can do is stated and checkable — not assumed from fluent output.
Shared workflows
One current way the work runs, with review gates people can see.
Decisions
A named person approves and releases. AI agent agreement is not approval.
Organizational knowledge
Accepted facts, decisions, and their sources stay available to the next task instead of dying in a private chat.
Company rules
How work should be performed, reviewed, and escalated — written so people and AI agents both work from it.

The three products

What each one contributes.

Each product works on different parts of the approach. The availability line repeats each product's own published status, unchanged.

  1. Pilot-stage product

    Agents Coworking

    Where a team and its AI agents do the work.

    A persistent room where a team and its personal AI agents work from shared context without hiding who contributed or who decides.

    Works on
    • Shared workflows
    • Decisions

    Not claimed: No outside team rollout is presented as completed proof.

    Open Agents Coworking
  2. Lab product · in development

    Organization Model

    How the organization keeps what it learns.

    A shared representation of intent, decisions, knowledge, and workflow skills that is updated when people correct or approve AI agent work.

    Works on
    • Organizational knowledge
    • Company rules

    Not claimed: The page shows the current learning cycle and the human decision point.

    See how it works
  3. Alignment Test available · education in development

    The Agents University

    How you examine an AI agent before you rely on it.

    A place to examine AI agent alignment now and, as the curriculum is developed, build learning records for AI agents and the people who work with them.

    Works on
    • Capabilities

    Not claimed: Current availability is stated separately from future education.

    Open The Agents University

How they fit together

Research connects these three. It does not make them interchangeable.

The Organization Model is the shared representation. Agents Coworking is where a team works. The Agents University examines and develops how AI agents learn.

They are three separate offers, not modules of one platform: adopting one does not require the others, and there is no shipped integration between them today. Where two of them make sense together, it is because they work on the same approach, not because data flows between them. A claim made for one of them is not evidence for another.

Where each product works on the approach

Capabilities
The Agents Universityexamines what an AI agent is aligned to before you rely on it
Shared workflows
Agents Coworkingone shared brief, visible contributions, named review gates
Decisions
All threea named person approves and releases — in the room, in the learning cycle, and before you rely on an AI agent
Organizational knowledge
Organization Modelaccepted facts, decisions, and their provenance, updated when people correct or approve AI agent work
Company rules
Organization Modelhow work should be performed, reviewed, and escalated

Illustrative example

One decision paper, from draft to release.

An everyday piece of work, and how the approach holds it together. This scenario is illustrative: it is assembled from the workflow designs published on the use-case pages. It is not shipped software and not a customer result.

  • HumanA person decides or reviews
  • AI agentsAI agents prepare or check
  • SharedPeople and AI agents both carry the step
  1. Human

    The decision owner fixes the brief.

    Priya, the decision owner, states the goal, the source material, what counts as done, and the deadline. That exact version becomes the baseline every later change refers to.

  2. AI agents

    AI agents prepare the first draft.

    AI agents draft the sections from the approved material and cite what they used. One of them flags where the material contradicts itself instead of averaging the contradiction away.

  3. Human

    Reviewers comment on one version.

    Comments are collected independently before they are reconciled, so agreement is not manufactured by whoever commented first.

  4. Shared

    Every comment ends in a stated outcome.

    Integrated, not integrated with a reason, superseded, or escalated to a decision — a comment cannot quietly disappear.

  5. Human

    Priya approves one specific version.

    The release decision comes back to her with every unresolved comment attached. Agreement between AI agents is not approval, and neither is silence.

  6. Shared

    The lesson becomes reusable.

    Her corrections and her approval are recorded where the next document starts from them, instead of dying in a private chat.

Agents Coworking gives this work its room: one shared brief, visible contributions, a human release. The Organization Model is how the last step becomes shared knowledge and updated company rules. The Agents University is how the AI agents were examined before the work was entrusted to them.

Where to start

Start from the problem you already have.

You do not need all three. One of these problems is probably already costing you:

Your team works with AI agents, and the work is hard to inspect.

A room where every contribution is visible and a named human owns the release.

Open Agents Coworking

The same corrections get made again.

See how a correction becomes a reusable lesson instead of a private fix.

See how it works

You are about to depend on an AI agent.

Examine what it is aligned to before you rely on it.

Open The Agents University
Book a 20-minute callPrefer email? di@8hats.ai

Not sure which one answers your problem? Twenty minutes is usually enough to tell.