Embers · learning across a lifetime

For the freedom
to become.

Embers builds AI-native learning environments that help people discover what they are capable of, turn curiosity into visible work, and build the judgment to shape a more meaningful life — from childhood through college, career, and leadership.

Capability should compound.

Most education products stop at the course, the grade, the certificate, or the training event. Embers is built around a longer arc: a living record of how a person learns, what they have made, where they are stretching, and which human support helps them grow.

That starts in K-12, where curiosity and confidence are still forming. It carries into college, where knowledge turns into judgment. It moves into career, where learning has to become real output. It reaches leadership, where clearer thinking shapes teams, institutions, and futures.

Ignite

School · Kindling

Discover how you learn, what sparks you, and how to stay with hard things.

Form

College · Kiln

Turn knowledge pressure into research, judgment, craft, and original work.

Move

Career · Mantle

Build real artifacts, new skills, and a portable portfolio of capability.

Illuminate

Leadership · Sol

See patterns, sharpen decisions, and develop the people around you.

Not a course. A formation record with a pulse.

The common layer is not content. It is the relationship between a learner, an adaptive AI system, and the humans who help that learner become more capable, more self-directed, and more able to contribute.

Over time, the system learns which explanations, projects, examples, mentors, collaborators, and challenges actually unlock progress. That is where the ecosystem compounds.

The record belongs to the learner. EU-native and GDPR-grounded from day one — because a record of how a person learns is theirs to carry, not ours to keep.

01 · Profile

How you learn.

Patterns, preferences, strengths, friction points, energy, motivation, confidence, and the kinds of help that actually work.

02 · Portfolio

What you make.

Explanations, projects, essays, tools, analyses, performances, automations, decisions, and visible work that proves capability.

03 · Relationships

Who you build with.

AI helps learners find the right human support, then helps them become that support for others: mentors, peers, experts, coaches, and leaders in both directions.

Four doors into the same promise.

Four products, one promise: what a learner builds at each stage — profile, work, judgment, relationships — carries into the next. Same fire, different rooms.

AI builds the floor. Humans open the ceiling.

The product surfaces change by stage, but the operating model stays recognisable: adaptive AI for daily practice and feedback; human mentors for care, accountability, and taste; peers for the oldest learning loop; experts for depth when a learner is ready.

01 · Floor

Adaptive AI

Patient practice, explanation, synthesis, feedback, retrieval, and scaffolding available when the learner needs it.

02 · Warmth

Human mentors

The people who notice, steady, challenge, encourage, and turn information into human growth.

03 · Mirror

Peers

Learning with and from others: teaching back, collaborating, comparing strategies, and building confidence in public.

04 · Depth

Experts

Domain specialists who reveal what excellence looks like when the learner is ready to go deeper.

One thesis. Four proof stories.

Kindling

One-to-one tutoring produces a two-sigma improvement over group instruction (Bloom, 1984). A 2025 randomised controlled trial found AI tutors outperforming active-learning classrooms on both learning and engagement (Kestin et al., Scientific Reports). Kindling pilots will measure learning velocity, wellbeing, and transfer to independent work — and publish openly, even if results are mixed.

Mantle

Online courses finish at a 3.13% completion rate (Reich & Ruipérez-Valiente, Science, 2019), and what training does happen fades — application drops from roughly 62% to 34% within a year (Blume et al., 2010). Mantle's proof is transfer: working artifacts people build, keep, and use on the job — with EU AI Act Article 4 now making AI literacy a legal obligation for employers.

Kiln

Exploring. The college proof story will centre on research quality, synthesis, original work, advisor feedback, and the formation of judgment.

Sol

Exploring. The leadership proof story will centre on decision quality, sensemaking, team development, and visible growth in the people around a leader.

Built in Lisbon. The first fires are lit.

Embers is built by WE3 Venture Studio in Lisbon, working alongside the operating team behind 42 Lisboa, 42 Porto, and TUMO Coimbra. Two founders, both parents — their own children are the first pilot users.

The platform

Ember and a cast of 41 specialist tutors — built, orchestrating live multi-agent sessions, and in daily testing.

Mantle

Born from AI Dive bootcamps already running in person with paying enterprise clients. The first remote cohort opens June 2026.

Kindling

Pilot conversations under way with school operators in Lisbon, building toward an 18–24 month validation study.

Help us build the learning arc.

Book a conversation

Schools & operators

Bring Kindling into your school.

Partner schools wanted for the validation pilot: you bring students, mentors, and the room — we bring the AI layer, training, and measurement.

[email protected]

Enterprise teams

Put your people in the lab.

Seats are opening for Mantle's closed-beta cohort — practical AI fluency for regulated sectors, built EU-native from day one.

[email protected]

Investors

Read the deeper story.

The data room covers the thesis, the platform, the market, and the evidence plan. Access is granted per person.

Request data-room access →