How It Works

One environment, three layers

Each person gets their own assistant, each course gets a shared AI team, and university services connect according to role and permissions.

Layer 1

Personal assistant

One per student and per instructor. It organizes tasks across the spaces each person can use, and it's optional: anyone can go straight into a space.

Layer 2

Shared academic teams

Knowledge and workflows for courses, sections, and working groups. The spaces are shared; each person's sessions and data stay separate.

Layer 3

Institutional services

Library, paperwork, scholarships, guidance, and research, connected according to each person's role and permissions.

The scenario in this demo

School of Social Sciences,
Introduction to Economics

First year

Section A

80 students

Unit 3 covers supply, demand, and market equilibrium in a 90-minute class, with the midterm on Friday.

Sofia

Student

Professor Martínez

Instructor

Try It

workflows

Pick a workflow and click through it. Each one shows the work first, and at the end, which agents made it possible.

Workflow 1, instructor

Preparing the next class

From the syllabus and readings to a class ready for review, without losing pedagogical control.

Section A, Unit 3
The need
Professor Martínez needs to prepare next week's 90-minute class for Unit 3.
How it starts
She opens Section A directly, or asks her assistant: “Prepare the 90-minute class for Unit 3 for next week.”
Where it runs
In the Section A space, which already knows the course syllabus and readings. She confirms topic, goals, and deliverables.
Kolmena Campus · Section A, Unit 3
Class focus
Draft ready for review
Class sequence, 90 minutes
  1. 15 minOpening: why do prices change?
  2. 20 minSupply and demand concepts.
  3. 20 minReading graphs and equilibrium.
  4. 25 minGroup simulation of a market.
  5. 10 minExit question and wrap-up.
Try a change: “Simplify the activity for first-year students and use the price of oranges.”
Generated materials
Class plan (90 min)
Reading guide: supply and demand
Presentation with equilibrium graphs
Group activity: market simulation
Formative assessment questions
Status: draft, pending instructor review
Agents that worked on this
Curriculum designAgent
Identifies objectives, concepts, and prior knowledge.Uses: Course syllabus and authorized readings
Class designAgent
Sequences timing, examples, and activities.Uses: Objectives from the previous step
Materials creationAgent
Prepares the guide, presentation, activity, and questions.Uses: Class plan and approved sources

Usage and cost are illustrative, pending real measurement of generation time, tokens, and agents invoked.

What this shows: the personal assistant coordinates, the section team does the work, the agents produce materials that fit together, and nothing reaches the 80 students until the instructor approves it.

See these workflows with one of your own courses

We define the scope, materials, and evaluation criteria together, in one short conversation.

Design an academic pilot

BG Image

Built Around People

An environment, not a catalog of chats

Students and instructors can go through their assistant or straight into a space. Both paths use the same spaces and follow the same permissions.

Student space

Sofía: one assistant, five course spaces

“My tutor” handles her schedule, exams, study planning, enrollment, and scholarships. Each course space knows its syllabus and AI-use policy. Procedures require an integration and her confirmation.

My Campus
Good morning, SofíaTwo deadlines this week. Want me to plan your study time?
This week
History paperDue ThursdayIn progress
Economics midtermFriday, Unit 3Plan ready
My course spaces5 courses
Library
Scholarships
Enrollment

Instructor space

Professor Martínez: one assistant, three sections

Her academic assistant covers teaching, research, meetings, and paperwork. Each section space builds on the shared course, but keeps its own schedule, assignments, and group-level tracking.

Introduction to Economics
Good morning, Professor MartínezA Unit 3 class draft is waiting for your review.
80students in Section A
64activities submitted
Needs your attention
Unit 3 classDraft ready, 5 materialsReview
Reinforcement activitySuggested from group patterns15 min
Research projectCall closes in 4 weeksOn track
Group pattern, Section A47%
Struggled with equilibrium quantity. Group level only, no individual chats.

Course, section, and person: three separate scopes

The spaces a person sees are not AI processes running all the time, and they're not a license count. Internal agents such as the planner, tutor, or evaluator are called on only when a workflow needs them. Sizing and cost still need to be validated.

Governed by design

Roles, permissions, and human control

The institution sets the common rules. The instructor keeps academic authority over every decision.

Role-based permissions

Each space, whether course, section, or person, defines who can read, write, or query its information.

Use policies

Each activity sets its own limits on AI assistance, visible to instructors and students.

Instructor control

No material or group analysis reaches students until a person has reviewed and approved it.

Review and audit

Every workflow records its sources, agents, steps, and review points, so the institution can audit how it's used.

Only authorized sources, each with a defined access level

Source

Who can see it

Unit syllabus and readings

The instructor and students of the course.

Learning objectives and approved materials

The instructor and authorized students.

Assessment drafts and answer keys

The teaching team only.

AI-use policy

The instructor and students.

Individual schedule and progress

The person themself. Other roles only with a legitimate reason and explicit permission.

Group-level activity patterns

The instructor, with privacy criteria and review.

These are proposed governance principles for product design. They don't claim proven certifications, data residency, or regulatory compliance. Every point is validated with the university before a real deployment.

Measure and connect

Evidence before claims

Usage metrics stay separate from learning analytics, and every connection to university systems is evaluated before it's switched on.

What gets measured

Proposed

Usage by role and workflow

Workflows started, completion rate, and active courses, from real product measurement.

Proposed

Consumption and cost

Compute usage by section or course. Visible spaces aren't the same as simultaneous agents or licenses.

Extra safeguards

Learning analytics

Group-level only, with thresholds, an authorization basis, and human review. Private chats and individual profiles are never exposed.

What it's designed to connect with

LMS and virtual campus

Potential integration

Academic systems

Potential integration

Repositories and library

Potential integration

Calendars and tools

Potential integration

No real figures on adoption, time savings, or learning impact are shown here, and no active integrations with any university exist yet. Metrics will come from a measured pilot, and each connection needs a technical evaluation and institutional authorization.

How to start

Start small, scale with evidence

A bounded, measurable, and reversible pilot, not an institution-wide rollout on day one.

01 Scope

One first-year course

One or two sections, with instructors and students who volunteer.

02 Materials

Only authorized content

Map the available systems and institutional rules before connecting any personal academic information.

03 Workflows

The three you just tried

Preparing a class, preparing for an exam, and turning an activity into a teaching decision.

04 Indicators

Agreed with the university

Usage by role, workflow completion, satisfaction, perceived usefulness, quality of reviewed materials, and consumption and cost.

Any learning indicator needs its own measurement design and safeguards, agreed with the university before the pilot starts.

From the person to the course, from the course to the university

A short conversation to define the scope, authorized materials, and evaluation criteria for a pilot in one specific course.

Design an academic pilot

The platform for collaborative AI swarms.
From single agents to intelligent teams.

Product

Platform

Solutions

Swarms

Marketplace

Pricing

Resources

Documentation

Developers

Campus

Company

Contact

Partnerships

Soon

Brand Assets

© 2026 Kolmena AI Inc. All rights reserved.

Privacy Policy

Terms & Conditions

The platform for collaborative AI swarms.
From single agents to intelligent teams.

© 2026 Kolmena AI Inc. All rights reserved.

Privacy Policy

Terms & Conditions