Inside TEEI Skills Academy: Professional Learning with AI in Every Field

A complete AI-first learning platform: live classes taught by working professionals, self-paced learning on DataCamp and Coursera, AI feedback that has to quote the learner, credentials anyone can verify and add to LinkedIn, and a way for technology companies to teach their own tools. Technology companies bring the tools and the expertise. Organisations bring the people and what they need to learn. TEEI connects them.

A person rebuilding a career after war or displacement needs three things a course on its own rarely gives. They need someone who does the work to look at theirs. They need a room where they can practise with other people. And they need proof that an employer can check.

Skills Academy is where The Educational Equality Institute (TEEI) brings those three together. Working professionals teach live, in five formats, in TEEI’s own video rooms. Participants hand in work and get feedback that quotes what they wrote, first from AI and then from a person who has the last word. Between sessions they study on DataCamp, and on Coursera, where Google Career Certificates reach them as scholarships, and they study together in rooms of their own. When they complete a class, they receive a certificate that anyone can check on TEEI’s own domain and that they can add to LinkedIn, with TEEI as the issuer. Companies send employees to teach. Technology companies teach the skills behind their own tools. Organisations bring their people in and ask for the classes they need. Alongside a class, the same account opens TEEI Language, to practise the language of the same work with a volunteer, and TEEI Mentorship, to work with a mentor from the same field.

AI runs through all of it. AI and data is a field of its own, and in every other field AI is part of how people learn and how their work is read.

From submitted work to verifiable learning. A participant hands in a task and scores their own work first. The model reads it against the class criteria, and every judgement has to quote the participant’s own words. A person on the class team reviews the work, and completion goes into a learning record that can become a TEEI certificate anyone can check. Each step below opens the screen it happens on, and the AI engine chapter has the detail.

1. TaskThe instructor sets the task and the criteria on the class desk. The participant hands in text, a link or files.
2. Self-assessmentThey score their own work first, 1 to 5 on each criterion, before anyone else has.
3. AI first feedbackThe model reads the work against the criteria and must quote it for every score, or the reading is refused.
4. Evidence-linked feedbackEvery comment points to a numbered sentence in the participant’s own work, so they can find the place it is about.
5. Human reviewA person on the class team scores and comments before seeing the participant’s scores, and has the last word.
6. Learning recordAttendance and reviewed work are recorded. The class certificate follows, with a public page anyone can check.

of 6

Previous stepNext step

Close

Your product, taught by the people who build it. For a technology company, the shortest way in is one class. The company brings the people who build its product and the material they teach from. TEEI provides the rest: the class page and the registration, the video room, the task, the attendance, the certificate with the company’s name on it, and the LinkedIn record that each instructor and each participant can show. The whole path, from proposal to published class, is in the companies chapter. Follow the whole journey, step by step, in the company journey.

TEEI is a nonprofit registered in Norway, with US 501(c)(3) status. Its programmes serve people from Ukraine, women from Latin America and women from Afghanistan. Skills Academy is free for learners, for instructors and for the organisations that take part. We keep it free because the people we work with are starting again, often in a new country and a new language, and a fee would decide which of them get to learn.

Skills Academy runs on the same platform as TEEI Language and TEEI Mentorship, with one account for all three. At 11:53 UTC on 2 November 2026, TEEI’s public counter showed 12,930 completed conversations and 9,755 hours of conversation across TEEI Language and TEEI Mentorship, with volunteers and learners from 148 countries. By then TEEI had issued 865 certificates across TEEI Language, TEEI Mentorship and Skills Academy, and each one has a public page that says whether it is valid. The board below reads the live figures when it can reach the site, so its numbers can be newer than this sentence.

Try it. You can open the rooms yourself. TEEI’s test room opens every room in its roles with one press, signed in as one of the demo personas: the sample learners, volunteers and hosts from TEEI’s public tours. It runs in a test environment with its own database, so no participant’s data is in it. Open a Skills Academy class as a participant or as its host, a webinar as a listener, an Expert Q&A, a TEEI Language conversation as the learner or a mentoring session as the mentee. The test room’s own page lists every room, with the code behind each.

The research behind the AI side, and the measure TEEI calls the crutch gap, has an article of its own: Can AI help people learn without making them dependent on it?

AI in every field

The Skills Academy page puts it in one line: “Everything in Skills Academy is professional learning, with AI in every field of it.”

Fifteen fields. The Skills Academy skills taxonomy has fifteen fields and 663 concepts. The fields are career development; HR and recruitment; leadership and management; project and programme management; product and design; AI and data; software and cloud; marketing and communications; finance and business; law and compliance; education and facilitation; languages and intercultural communication; sales and customer service; digital and office tools; and personal effectiveness. The same concepts connect a class to the mentors who work in that skill, so a participant can go from one to the other.

AI as a field. AI and data runs from AI foundations and prompt practice to responsible AI, and it goes in three directions. Each one ends in work a person can look at.

AI in every other field. In the other fourteen fields, AI is part of how people learn and of how their work is read. A participant in finance checks a budget with it, a recruiter drafts a job description and then tests it against the role, a marketer asks it to argue against a campaign brief. Their submitted work gets AI first feedback, and a person on the class team reads it before any score counts. The engine chapter explains how.

When AI is allowed. One rule decides it. When using AI well is the skill being learned, the AI tool is allowed in the assessment. When the assessment is of understanding without help, it is not. The rule plays out differently in different kinds of learning. In AI literacy, using the tool well is the skill, so the tool is open throughout. In AI at work, the tool is open for the task, and the question afterwards is what the participant understood of the result they put their name to. In agent development, the agent is the work, and the participant explains what it does and where it fails. In a professional skill such as finance, recruitment or marketing, the instructor decides for each task, because a budget checked with AI and a budget checked without it teach different things. The rule comes from the evidence: open access to a strong model can raise practice scores and lower what people can do once it is gone. How much of a skill survives without the tool is the question our research article sets out, with the studies behind it and the measure TEEI calls the crutch gap.

What is on. On 1 November 2026 the public programme listed four upcoming sessions. One is about working with AI, and three are in other fields:

• Expert Q&A: “From Hot Seat to Conversation: How to Relax and Interview with Confidence”, 4 November 2026, 17:00 UTC, 60 minutes, 20 places.
• Workshop: “Beyond the Numbers: Think Like a Data Analyst”, 5 November 2026, 19:00 UTC, 90 minutes, 12 places.
• Workshop: “Learn Smarter with AI: Gemini Notebook Workshop”, 17 November 2026, 17:30 UTC, 90 minutes, 12 places. Each participant leaves with a personal study notebook they have built.
• Expert Q&A: “Networking for Introverts: Building Career Relationships Without Pretending to Be an Extrovert”, 2 December 2026, 17:00 UTC, 60 minutes, 20 places.

A focus of their own. On their profile, participants choose what they want Skills Academy sessions to help them with. Sessions that match come first and are marked for them.

The AI curriculum

The AI side of Skills Academy is built from two things: the AI training that the providers publish openly or give free, and what TEEI puts around it. The providers publish the courses. TEEI puts four things beside them, and they combine as the learning needs: a live session with a professional, a task in the participant’s own field that a person reads, a mentor who works in that skill, and a study room shared with the others on the same course. One person takes a course alone. Another takes it in the study room, with people who are on it at the same time. A third takes it with a professional, in a class with a task each week. A path for a group combines several of the parts, and an organisation can ask for the one its people need.

What the providers publish, and how it is used.

Two kinds of record. A provider’s badge or certificate is the provider’s. The TEEI certificate is for the class a person attended and the work a person reviewed, and it is the one anyone can check on TEEI’s domain and add to LinkedIn with TEEI as the issuer. The credentials chapter shows the whole chain.

Who the providers are to TEEI. DataCamp’s licences come through DataCamp Donates, and the Google certificates come as Grow with Google scholarships on Coursera. The other resources are published openly or given free, and TEEI uses them as their licences allow, with credit. When the people who build a tool teach it themselves, they come in as instructors, through the companies chapter.

Beside every course. Whatever the course, the same four parts are there to combine. A live session with a professional, in a TEEI room. A task in the participant’s own field, scored by the participant first, read by AI under the quote rule, and reviewed by a person. A mentor from the directory who works in that skill. A study room shared with the others on the same course. Nobody has to take all four: a participant who wants the course on its own takes the course, and a class that wants a task each week gets one. The engine chapter shows how the reading works.

Learning live

A participant’s week in Skills Academy starts on My Academy, the home page that shows what they have booked and what they have been invited to. The menu has five places: Home, Notifications, Discover, Your plan and Messages.

Finding a session. Discover shows the coming days as a strip, with the sessions of the chosen day listed below. Search takes a topic or an outcome. A filter sheet narrows the list by focus, by when the session is, by commitment (one session or a series) and by the language it is spoken in. A session page says who teaches it, in which format and language, for how long, how many places are left, and what participants will leave with.

Five formats. Every format is one live session online, except the weekly class.

The group sizes are the limits an instructor chooses within. A workshop stays small enough for everyone to practise. An expert Q&A can take a larger room, because the work is asking.

Registering. A participant registers with one tap. When a session is full, they can join the waiting list, and when a place opens, the first person on the list gets it and is told. If the instructor moves or cancels a session, everyone registered hears about it, with the instructor’s reason. An hour before a session, each participant gets a reminder. Once a week, a digest lists what is coming.

Your plan. Your plan lists the sessions a participant registered for, the meetings of their classes and the deadlines of their tasks, in date order. A whole class goes into their own calendar with one file.

Before, during and after. Before a session, the participant picks what they want from it, from six goals. The instructor sees how many chose each goal, never who. Afterwards the participant is asked whether it happened: yes, partly or no. Once attendance is counted, the instructor writes a recap for the people who came, with what they covered and one thing to try next, and it appears on the session page under “After the session”. Every session and every class has a group chat in Messages, so the conversation carries on between meetings.

A class from week to week. A class has its own page with every meeting, the agenda for each one, the readings and the task for the week. The participant hands in an answer, a link or files, and scores their own work first against the class criteria. Then come AI first feedback and a review from a person on the class team. The participant completes the class when they have attended and the instructor or an assistant has approved their work. The certificate follows.

Where it leads. After a class, a participant can go straight to the mentor directory, filtered to mentors who work in the skill that was taught. A participant whose next barrier is the language can book a TEEI Language conversation about the same work. Every session has a public page at its own address, with a card for sharing, so a participant can invite a friend or a colleague.

Asking for a class. When the class someone needs does not exist, they can ask for it at Request a class. They need no account: they confirm the request by email, and TEEI finds a volunteer instructor to teach it.

TEEI’s own rooms

Every live session happens in a TEEI room. The rooms run on Amazon Chime SDK, reached through a signed bridge from TEEI’s platform, and they are the same rooms TEEI Language and TEEI Mentorship use for conversations and groups.

Who is in the room. The instructor hosts. Before the room opens, participants wait in a waiting room, where they set up their camera and microphone. Nobody sees or hears them before the host lets them in. Participants join muted and unmute when they want to speak. In a session built around a talk, they listen and can raise a hand, and the host can invite them to speak. A dropped connection reconnects on its own, and when the session is over, the host ends it for everyone.

Attendance from the room. Attendance is counted from time in the room, not from a tick on a list. A participant’s attendance at a session that has just ended shows as being calculated, not as absent, until the room events are settled. An instructor’s teaching time is the measured time in the room, capped at the scheduled length of the session. Anyone in a room can report abuse from inside it.

How much the rooms are used. Between 1 July and 31 October 2026, AWS measured 184,131 participant minutes in TEEI’s rooms across all its programmes. Verified volunteer hours are counted from the same room events: How TEEI Turns Amazon Chime SDK Events into Verified Hours shows how a room event becomes a verified hour, and What TEEI Runs on AWS shows the rest of the infrastructure. How TEEI Runs Its Platforms on Cloudflare Workers shows the platform the rooms are reached from.

Open a room yourself. TEEI’s test room opens every room in its roles with one press, signed in as a demo persona, in a test environment with its own database: a Skills Academy class as a participant, as its host, a webinar or an Expert Q&A.

Learning at your own pace

Live sessions are a few hours a week. The learning between them happens on DataCamp and Coursera, in courses TEEI learners receive free.

DataCamp. Through DataCamp Donates, TEEI learners study data, programming and AI on DataCamp. By 8 October 2026, DataCamp had allocated 2,933 licences to TEEI learners, and 2,300 of them had been started. Counted from DataCamp’s daily data on 1 November 2026, TEEI learners had spent 78,942 hours on the DataCamp platform and completed 271,611 learning units and 16,259 courses, and 2,434 people had recorded learning time. The AI courses include Generative AI Concepts, Understanding ChatGPT and Understanding Machine Learning. A participant requests DataCamp from Learning resources with one tap. How TEEI Connects DataCamp Learning with Live Teaching explains how the courses and the live sessions fit together, and DataCamp in TEEI’s Programmes: A Thank You tells the story of the partnership.

Google Career Certificates on Coursera. Google Career Certificates reach TEEI learners as scholarships through Grow with Google, on Coursera. Google AI Essentials and Google Prompting Essentials are among the courses. By 1 November 2026, 449 people had joined, 284 had started a course and 93 had completed at least one, with 402 courses completed in all. These are course figures, as Coursera counts them.

Study rooms. So that no one has to study alone, Skills Academy has study rooms. They are part of the shared rooms in Messages, beside the community rooms for the social club and for volunteers, and today there are two of them: participants who have requested DataCamp share the DataCamp room, and those who have requested Google Career Certificates share the Coursera room. A room for another course or another learning community is built on the same room. In a room they write in the chat, use Find a time to agree a study session with a poll, or press Study now to be matched with someone who wants to study at the same moment, for 30, 45, 60 or 90 minutes. Planned study sessions appear for everyone in the room. Each room belongs to the members who requested that learning.

Kept apart. A DataCamp course completion and a Google Career Certificate are the provider’s records. A TEEI class certificate is TEEI’s record of a class a person attended and work a person reviewed. Course completion, a provider’s certificate and a piece of work TEEI has reviewed are three separate records, shown side by side. What a Course Certificate Shows, and What It Doesn’t goes into the difference.

Alongside TEEI Language and TEEI Mentorship

Skills Academy is one of three TEEI programmes on one account. In the same weeks as a class, a participant can practise the language of the same work in TEEI Language and work with a mentor from the same field in TEEI Mentorship. She can take one, two or all three, in any order, and each keeps its own record.

Why they belong side by side. A skill gets someone a job when they can show it, explain it in the language of the workplace and reach the people who hire for it. A class builds the skill and the certificate that shows it. The other two programmes carry it the rest of the way.

The language of the work. An analyst who can build the dashboard still has to present it in an interview, in the language of the country she now lives in. In TEEI Language she books a one-to-one conversation with a volunteer about the same work, and writes what she wants to practise when she books: explaining an analysis, answering the questions an interviewer asks.

Between conversations, Practice Companion lets her rehearse a short job interview with a voice on the other end, and a line she keeps can travel into her next conversation. How Learners Get Sentence Feedback in Practice Companion shows how it corrects one sentence at a time, Inside Practice Companion: Rooms, Evidence and Level how it reads a level from her own answers, and AI Writing-Level Estimate: What TEEI Tested and Learned which AI models TEEI tested for that level.

Language carries much of the weight. In a study of refugees assigned to Swiss cantons on either side of the language border, speaking the local language more than doubled employment in the first five years (Schmid, American Economic Review: Insights, 2023). TEEI Language runs in twelve languages.

A person in the field. Research on mentoring separates role models, advisers, coaches, sponsors, introductions and practical help with a job search, and finds that the last three drive employment effects. TEEI Mentorship is built around them. A mentor who works in the skill reviews her CV, practises interviews with her, plans the job search and knows what employers in that field look for. In a field experiment in Munich, an NGO that found suitable vacancies and sent refugees’ CVs to employers raised employment after twelve months, most among those with the least formal education (Battisti, Giesing and Laurentsyeva, Labour Economics, 2019).

How the three connect.

• One account, one booking system and the same TEEI rooms, with attendance counted the same way.
• The skills taxonomy connects a class to the mentors who work in that skill. After a class, the mentor directory opens filtered to them.
• Mentors and mentees choose their occupation from ESCO, the European Union’s classification of 3,039 occupations, with names in English, Norwegian, Polish and Ukrainian.
• One credential registry holds the class certificate, the TEEI Language ladder and the TEEI Mentorship ladder, each with a public check page.
• Each programme keeps its own rules and safeguarding. A learner’s practice history stays out of her volunteer’s view.

The people who give their time move too. The same person can teach in Skills Academy and mentor in TEEI Mentorship. A conversation partner can apply to teach from inside TEEI Language. A company’s employees can do all three, and their verified hours reach Evenhall from every programme.

An example. This one is illustrative. Olena takes the class Data storytelling with spreadsheets, which meets weekly for three weeks. In the same weeks she books two conversations in TEEI Language about presenting an analysis in English. After the last meeting she opens the mentor directory filtered to data analysis and books a mentor who reviews her CV and prepares her for an interview. The class gives her a certificate an employer can check. The conversations give her the words for her work. The mentor gives her the route to the people who hire for it.

Teaching

The people who teach in Skills Academy do the work they teach. They are recruiters, analysts, accountants, product managers, engineers and marketers, giving an hour of their own field to people moving into it. Some come on their own. Others come through their company, which sends employees to teach. Those who apply on their own are approved by TEEI before their first class. An AI model reads each application first and returns a verdict, and instructors and mentors always go to a person at TEEI, whatever that verdict says. A colleague a company sends creates and holds classes without manual approval. What they teach is open: their own skills, or how they use them at work.

Joining. An instructor applies through Teach on Skills Academy, or follows a personal invitation link from TEEI. Their profile says what they do, what they teach and which languages they teach in.

Publishing in three steps. The instructor publishes a session themselves, in three steps. First comes what it is: the format, the title, what participants will leave with, what the group will work on and anything to prepare. Then comes when and how many: the date and time, shown in participants’ own time zones, every weekly meeting of a class, and the number of places. Last comes who it is for. A preview shows how participants will see it. An instructor teaching for their organisation chooses it when publishing, and the class then carries the organisation’s name. A second instructor can be added: they get a request in the app, the session is published when they accept, and the time is then reserved in their calendar too. Assistants can help with the class work.

The class desk. Each class has a desk for its team. It shows the participants, attendance for every meeting, the work handed in and who has completed. The instructor sets the agenda and the task for each meeting and adds readings. Once a session is published, its result, shape, day, time and places are locked, so participants can rely on what they registered for. “Change time” moves a session, and cancelling it asks for a reason; everyone registered is told, with the reason. The desk’s page called First readings shows the AI first feedback and the classmates’ reviews beside the instructor’s own scores, with how far each reading sits from theirs. It has no threshold, no verdict and no total: it helps the team see where the AI and the classmates read the work differently.

Teach again. When a session is over, Teach again repeats it with one tap, so an instructor who taught once can teach the same thing next month without starting from the beginning. Two instructors from the same company can teach a class together.

A page that can be shared. Every published session has a public page at its own address, with a card that shows the title, the instructor and the date when the link is shared. The instructor can also tick one box when publishing: “Post it as an event on TEEI’s LinkedIn page”. The credentials chapter explains what happens then.

What the instructor receives. Each instructor receives a Teaching Contribution credential for the sessions they taught, with the meetings taught, their role in each and the recorded teaching time. It is a TEEI credential like the class certificate, and it goes to LinkedIn the same way.

How Corporate Volunteering Works at TEEI and TEEI Pathways: A Better Way to Meet People, Learn and Contribute describe the other ways volunteers contribute.

Credentials, LinkedIn and public profiles

A certificate is only worth what someone else can check. Skills Academy issues its own credentials, puts a public check page behind each of them, and connects them to LinkedIn, where employers and colleagues already look.

Two credentials. TEEI issues two credentials in Skills Academy. A participant receives a class certificate when the class team confirms they completed a class: they attended, and the instructor or an assistant approved their work. An instructor receives a Teaching Contribution credential, “Teaching Contribution · TEEI Skills Academy”, for the meetings they taught. Both are issued by TEEI and appear on the person’s Certification page.

What a certificate shows. A class certificate names the class, the person, the date and the certificate code. When the provider has said yes, it carries the line “Class provided by” with the provider’s name and logo. When the instructors have agreed, it carries “Taught by” with their names. Each instructor decides that for themselves, and the setting says exactly where the name will appear: “When this is on, learners who complete a class you taught see your name, as “Taught by”, in four places: their certificate (in Skills Academy and on its public check page), the printed certificate, the share image, and the suggested text for a LinkedIn post, which they can change before they post it.” Each certificate also carries a QR code that opens its check page.

Three forms. The certificate exists as a badge in three sizes, as a share image sized for LinkedIn at 1200 by 627 pixels, and as an A4 page to print. The participant downloads the PDF or the image from Certification.

Checking it. Every TEEI certificate has a public page at theeducationalequalityinstitute.org/verify/<code>. It opens without an account and answers one of four things: valid, revoked, expired or not found. An employer who sees the certificate on LinkedIn, on paper or in an application can scan the QR code or type the code and see TEEI’s answer. The check page also has a card of its own, so a link to it shows the certificate when shared.

Adding it to LinkedIn. From Certification, the participant presses Add to LinkedIn. LinkedIn opens the Licenses & certifications form on their profile, already filled in: the name of the credential, TEEI’s official LinkedIn page as the issuing organisation, the month and year it was issued, the expiry date when there is one, the credential ID and the address of the check page. They save it on LinkedIn, and the credential appears on their profile with TEEI’s logo. Anyone who opens it lands on TEEI’s check page.

LanguageMentorshipSkills Academy

Languageby TEEI

Certificate of volunteer service

Awarded to

Jonas Brekke

Senior Language Volunteer

In recognition of the time, patience and encouragement given across 14 verified sessions, and the confidence those conversations helped build.

14

Verified conversations

9

Learners in Ukraine

12 October 2026

Issued

Issued by The Educational Equality Institute · CERT-EXMP-L101

Verify at theeducationalequalityinstitute.org/verify/

LinkedIn · Licences & certifications

Senior Language VolunteerThe Educational Equality InstituteIssued Oct 2026Show credential

Posting it. Post on LinkedIn opens a post with the share image and suggested text, which the participant can change before posting. When the instructors have agreed, the text names them. When the provider has agreed, it names the company that provided the class. The certificate also arrives by email.

A statement of hours. Volunteers, instructors among them, can download a statement of their verified hours from Certification. Each statement has a public link, so a university, an employer or a funder can open the same figures the volunteer sees. How TEEI Counts a Verified Volunteer Hour explains what counts, and From a Verified Hour to a Report You Can File follows an hour into a company report.

The public profile. A volunteer can turn on a public profile. It opens without signing in, at its own address, under the heading “Public volunteer profile”. It shows their roles (conversation partner, mentor or Skills Academy instructor), when they started volunteering, the languages they speak, their TEEI certification, and a link inviting others to “Teach on Skills Academy”. The volunteer decides whether their photo appears. The foot of the page says: “Volunteering with TEEI. This page contains the information this volunteer chose to share publicly.” When the profile is on, the settings page tells them: “Anyone with this link can open your public profile without signing in. You can add it to LinkedIn.” Reviews, contact details and bookings stay private.

Sessions on TEEI’s LinkedIn page. When an instructor ticks “Post it as an event on TEEI’s LinkedIn page”, TEEI’s page announces the session. Every five minutes, the platform compares each request with the session’s public page. It creates a LinkedIn event and a post for a new session, updates the event when the session moves, and takes the event down when the session is no longer public. A session starting within five minutes is skipped. A weekly class goes on LinkedIn with its first meeting. A model writes each class post once, from the facts of the session, and the text is then fixed; if the model fails three times, a template writes it instead. Consent is read again before every call to LinkedIn: a post that names a person who has since been erased, or a company that has since said no, is taken down. When a session has a co-host, the event is requested only after the co-host has said yes.

One LinkedIn page. Everything goes through TEEI’s official LinkedIn page. Skills Academy has no separate page. Credentials, events and posts all point to the same organisation, so that a person’s credential, the session they took and the organisation that issued it can be found together.

Why it matters to each person. For a participant, the certificate is evidence for an application: a class with a named provider and a named instructor, and a check page an employer can open. For an instructor, the Teaching Contribution and the “Taught by” line are a public record of teaching they can show their own employer. For a company, “Class provided by” puts its name on the work its people did, in the place where its future employees and customers look.

Companies

A company’s people know things TEEI’s learners need to learn, and teaching them is good work for the people who do it. Skills Academy gives a company four ways to take part. All four are free.

Four ways.

Product taught by builders. TEEI’s own words for this are: “your product, taught by the people who build it. Hands-on use is teaching; a sales pitch is not.” A company that builds a coding agent, a cloud platform or a productivity tool has people who know the craft behind it: how to plan work for an agent and review what it returns, how to go from an idea to a working tool, how to know when not to use AI. Those are skills TEEI’s learners need for work, whatever tool they end up using. The provider is named on the session and on the certificate. The session is about learning, not selling. Participants’ details are not passed to the company that teaches.

Classes a company could teach. These are examples of classes a technology company’s people could teach, and an open invitation to each of them.

From proposal to published class. A company starts at Propose a class, under the heading “Propose a session your organisation can teach”. It chooses “I will teach it myself” or “Someone else from your organisation”, and describes the session. Sending needs an account. TEEI invites the instructor, and the page says what happens next: “Nothing is published until the instructor accepts.” The instructor accepts, the session is published with the company named, and when two people from the company teach it, both are on it. After the session, the instructors receive their Teaching Contribution credentials, participants who complete receive class certificates with “Class provided by”, and the session can go out as an event on TEEI’s LinkedIn page. A company can also ask for a private class for its own staff.

1. Company proposes a classA company describes one session its people can teach, at Propose a class, and names who teaches it.
2. Instructor invitedTEEI invites the instructor the company named. An instructor new to TEEI meets the people they will teach, then registers.
3. Instructor acceptsNothing is published until the instructor accepts. A second instructor from the company accepts in the same way.
4. Class publishedThe session is published with the company named on it, and participants register from Discover.
5. Participants attendThe class meets in TEEI’s own video room, with host tools, conversation cards and raised hands.
6. Work reviewedParticipants hand in work between meetings, and the instructor reviews it with a score and a comment.
7. Credentials issuedParticipants who complete receive the class certificate with Class provided by. Instructors receive a Teaching Contribution.
8. Teaching recorded in EvenhallThe hours arrive in the company’s Evenhall workspace overnight as verified hours, by named employee.

of 8

Previous stepNext step

Close

Evenhall. Evenhall is TEEI’s workspace for companies. A company opens a workspace, which TEEI approves, and invites its employees by email, through single sign-on, from its HR system or from a list. Employees have a page of their own in Evenhall, where they find Skills Academy, TEEI Language and TEEI Mentorship and join with their company. The hours they give in TEEI’s rooms arrive in Evenhall overnight as verified hours. A company’s CSR manager sees each named employee’s hours and session records. Employees never see each other’s records. Participants’ names stay out of the company’s reporting, and so does what was said in a session. The company sees the hours in views for ESRS S1, VSME, GRI and IFRS reporting, and AI drafts the text of a report, which a person at the company approves and which locks once approved. Evenhall is free: no licence fee, no setup fee, no limit on employees and no donation required, and a donation unlocks nothing. Why Our CSR Workspace Is Free, and How to Check It sets out what free covers. How Evenhall Drafts Company Reports with the Vercel AI SDK shows how a report is drafted, and the Evenhall demo shows a workspace with example data.

From a class to the company’s report. A colleague teaches a Skills Academy workshop on a Tuesday evening. By the morning, the measured time in the room, capped at the scheduled length, is in the company’s Evenhall workspace as verified hours with the colleague’s name on it. Where nothing was measured, such as a host’s own time in a group session, Evenhall credits no hours rather than estimating them. The CSR manager sees the hours by programme and by day, no participant is named, and hours a volunteer entered by hand are listed apart and stay out of the total. At the end of the quarter the manager asks for a report, reads the draft, approves it, and it locks. The screens below are the demo workspace with sample data.

Your teaching, on record.

Attendance is recorded with each session you teach. Your company’s activity report shows the sessions you teach and the time you spend teaching online. Three things come from that record: your participants’ certificate, your own teaching certificate with the meetings you taught and your measured time, and the report.

Sample data from the company pages: a clinic hosted on 15 October, 75 minutes recorded, and the instructor’s Teaching Contribution for a class of four meetings, 5 hours measured.Demo dataMoreLess

Source: Evenhall demo workspace, as shown on the company pages, 28 October 2026.

What each team gets. Teaching does different work for different parts of a company. Developer relations reaches people who learn its tools properly, in a session with no sales pitch. Learning and development gets a real group, a plan and a credential for each person who teaches. The people who run volunteering get verified hours they can report. Each instructor gets public recognition for teaching: the “Taught by” line, a Teaching Contribution credential and a public profile they control.

Who does what. TEEI runs the programmes, reviews every employee’s profile before anyone can book them, runs the video room, counts the hours and looks after the participants. The company chooses the programmes, decides whom to invite, gives people room in the week, and reviews and approves the report. How Corporate Volunteering Works at TEEI sets out the split.

Organisations

Many of the people Skills Academy serves first hear of it from an organisation they already trust: a community group, a university, a senior community or a local service. TEEI Network Partners is how those organisations work with TEEI’s programmes.

Share. An organisation can share Skills Academy, TEEI Language and TEEI Mentorship with the people it reaches, with no registration needed for sharing. A code in the link records that the organisation made the introduction.

Contribute. It can contribute its own people as conversation partners, mentors, instructors or group hosts. Free opportunities are open to every organisation’s community, whether or not it contributes anyone.

Request a class. It can ask for a private class for its own group, in any of the five formats, and TEEI reviews the request and finds the instructor. This is the route for an organisation that has a need and no one to teach it.

Offer a class. It can offer a class taught by its own staff, through the same proposal as a company. This is the route for an organisation that has someone with something to teach.

Network Workspace. Each partner has a workspace at Evenhall, where it can, in the workspace’s own words, “Share TEEI programmes with your community, follow participation, contribute teachers, mentors and volunteers, and request classes.” It has a getting started page, its programmes, its team, its requests and contributions, its participants with an export, a report it can print and its sharing tools. Network Workspace is part of Evenhall, on the same platform and the same record of verified hours as a company’s workspace.

Network Assistant. The workspace has an assistant that works inside the product. A coordinator tells it what the organisation is trying to make possible, and it opens programmes, prepares class requests and contributions, fills in forms and saves drafts. The model proposes and the application decides: before every action, the application checks identity, organisation, permissions and state, just as it does for a button.

Attribution without exposure. An organisation sees that its introduction led to participation, and in which programme. Names appear only under each participant’s own sharing choices. Registrations, bookings, attendance, completed activities and recorded hours stay distinct, and people and activities are counted separately. The live page says what the arrangement is: “You share the offer. TEEI runs the platforms and looks after the programmes.”

Inside TEEI Network Partners: From Community Reach to Real Participation goes through the programme in full.

A track for organisations

Skills Academy also teaches the organisations around the people it serves: the staff and volunteers of a Network Partner, learning the work of running it. The track starts from eight areas of documented need:

• financial management, budgeting and accounting workflows;
• reporting, monitoring, evaluation and data;
• fundraising and grant applications;
• communications, programme distribution and digital reach;
• AI and practical automation;
• project and operations management;
• HR, teams and volunteer management;
• governance, compliance and partnerships.

Each course is built around the organisation’s own work. It can combine structured learning, a live workshop, a clinic on the organisation’s own case, a mentor, a professional from Evenhall, a template or a workflow put into use, and follow-up once it is in use.

An example. An illustration of how the track works. A coordinator at a small organisation writes to the Network Assistant: “We want to use AI and need a place to start.” The assistant asks what work repeats every week, which tools the organisation already uses and who will own the result. The coordinator answers that three volunteers answer the same questions from families by email. The assistant proposes a route, and the coordinator confirms it. The coordinator joins a workshop on AI at work. A clinic built on the organisation’s own emails follows, with an instructor from a technology company. An AI specialist from Evenhall then sets up one workflow that drafts answers to common questions for a volunteer to check, and a mentor stays with the coordinator for the next two months.

Four records stay apart. The workshop, the clinic and the specialist’s hours are activity. The workflow and its instructions are output. The volunteers using it every week is adoption. The outcome is measured with comparable evidence, such as the time the answers took before and after.

A neutral classroom for AI training. AI companies train nonprofits to use AI, and Skills Academy is built to be a neutral classroom for that training. An AI company can bring its courses into it as short live sessions, with attendance and a certificate, taught by its own people, who come through Evenhall as other employees do. The rule from the companies chapter applies: the provider is named, the session is about learning, and participants’ details are not passed to the company that teaches.

The AI engine

The AI in Skills Academy follows one rule above the others: it may not judge work without showing where in the work its judgement comes from. This chapter is the technical part of the article.

One engine, two programmes. Skills Academy runs on the engine built for Practice Companion, TEEI’s language practice. Both reach Claude on Amazon Bedrock through the same signed bridge and the same function, and both keep the same rule: no score without a quote from the person’s own work, with the model, the prompt version and the rubric stored beside it. The feedback runs on Claude Sonnet 4.6 through Bedrock’s EU cross-Region inference profile. Inside Practice Companion: Rooms, Evidence and Level describes the engine in the language programme, How Learners Get Sentence Feedback in Practice Companion shows a correction from the learner’s side, and AI Writing-Level Estimate: What TEEI Tested and Learned shows how one of its scorers was tested.

For every task the engine runs the same way, in six steps: the task, the participant’s own assessment, AI first feedback, feedback linked to evidence in the work, a person’s review, and the learning record. The rest of this chapter follows that order.

The task. Each meeting of a class has a task, which the instructor sets on the class desk together with the criteria the work will be read against. A participant hands in an answer, a link or files: PDF, Word documents, spreadsheets, slides, text and code, up to ten files and 20 MiB each.

Self-assessment. They score their own work first, on each of the class criteria, from 1 to 5. They can edit the submission until someone has reviewed it.

AI first feedback. The participant can then ask for AI first feedback, which has been on in Skills Academy since 23 October 2026. The function is forced to call one tool, record_first_feedback. For each criterion, the tool’s schema requires a key, a score, a quote and a comment, and a status of feedback or insufficient. The quote must appear in the submission exactly once. If one row fails, the whole reading is refused rather than partly kept. The participant’s text goes to the model inside a guard-content block, so the guardrail judges only their words, and the model reads them as data, not as instructions. A submission under fifteen words gets insufficient without a model call. Each reading is stored with the model ID, the prompt version, a hash of the system prompt, the schema version, the temperature and the rubric version. The participant sees it labelled “AI · not an assessment”.

Guardrails and receipts. Two guardrails are in use, each pinned to a numbered version, and the infrastructure refuses a draft guardrail. The content filters run on the standard tier because TEEI’s learners write in Ukrainian, Polish and Norwegian. In Practice Companion, the level scorer cannot be deployed without a green evaluation receipt that fingerprints the code, the rubric, the model, the temperature and the guardrail. The functions log outcomes, counts and token numbers rather than text, and each one has a per-person hourly limit.

Feedback linked to evidence. Every comment the participant sees stands on a quote from their own work, so they can go to the place the comment is about. The participant sees the comments and quotes at once, and the scores after a person on the class team has read the reading.

Human review. A quote shows what a judgement rests on, and a person on the class team decides whether it is right. A reviewer does not see the participant’s own scores until they have saved their own, so the participant’s view does not steer theirs. The instructor’s scores then appear beside the participant’s. There is no total, no grade and no pass or fail. The AI’s reading keeps its label, “AI · not an assessment”, with the date the class team saw it, and the team’s review is the one that counts.

Classmates, if the instructor wants it. Reviews by classmates are off unless the instructor turns them on for the class. Then a participant can ask a classmate for a review, anonymous both ways and never a requirement. Only a review by the instructor or an assistant counts towards the class certificate. AI readings and classmate reviews never do.

First readings. On the class desk, the team’s First readings page puts the AI’s readings and the classmates’ reviews beside the team’s own scores, with the average difference and how often the scores are the same. It is there so the class team can see how the AI reads their kind of work before they rely on it.

The learning record. A participant completes a class when they have attended and the instructor or an assistant has approved their work. Registration, attendance, instructor hours and reviewed work are recorded. Enrolment, logins and videos watched are operating data, not learning. The class certificate in the credentials chapter rests on that record.

The AI trainer. The next part of the engine is an AI trainer for practice between sessions, designed and not yet built. It will give hints and never the answer, and the rule will live in the tool, not only in the prompt: a field for a hint, anchored in the participant’s own text and kept short, a field for the next move, and no field for a solution. A task set with known answers will catch any solution that leaks. A day, a week and a month later, the participant will get a new task of the same kind without hints. The difference between what someone does with help and without it is the measure TEEI will watch most closely, and calls the crutch gap. Our research article, Can AI help people learn without making them dependent on it?, sets out the design of that measure and invites AI companies and researchers to build it with us.

What it offers each team

The same platform does different work for different readers. This is where each of them starts.

How it was built

TEEI’s platforms are built by an organisation of coding agents. Claude Code and Codex work in the same repositories under one shared set of instructions and check each other’s work. In TEEI’s 18 measured repositories, 25,302 commits were recorded, and 91.5 percent of them carry a Claude or Codex marker. The marker shows which agent made the commit, not who wrote each line. Between 28 September and 27 October 2026 there were 212 production deploys to 12 TEEI Workers, and 188 of them were recorded as run by Claude Code.

That is why the rules that matter in Skills Academy are written as code and tested like code: quote or refuse, a person with the last word, no field for a solution, consent read again before every post. How TEEI Works: One Platform, Many Programmes shows how the programmes share one platform. In front of the platforms sits Cloudflare: between 22 September and 16 October 2026 its Managed Ruleset blocked about 1,771 requests at Skills Academy’s address, and the Worker that answers Skills Academy and TEEI Mentorship handled about 239,000 requests in the 30 days to 16 October, as How Project Galileo Protects TEEI’s Platforms shows.

TEEI’s homepage shows what a verified hour costs to deliver: in September 2026, USD 0.78 per verified impact hour, from USD 343.14 in AWS video and platform costs before credits across 439.2 verified hours in TEEI Language and TEEI Mentorship. Skills Academy hours are not part of that figure.

Start here

Skills Academy grows one session at a time, and every session starts with one person who decides to teach it, take it or ask for it.

• Learn: create an account and find a session on Discover.
• Teach: apply to teach. Those who apply on their own are approved by TEEI before their first class, and what you teach is open.
• Propose a class from your company: Propose a class.
• Ask for a class your group needs: Request a class, with no account needed.
• Bring your organisation: become a Network Partner.
• See Evenhall: the demo workspace.
• Build with TEEI as a technology company: Build with TEEI.
• Follow the sessions: TEEI on LinkedIn.

If your company builds the tools people now need for work, we would like your people to teach the skills behind them. One session is enough to start, and TEEI does the rest: the page, the registration, the room, the attendance, the certificates and the record of the hours your people gave.

Verify this

• Programme totals: https://language.theeducationalequalityinstitute.org/api/lc/public-stats (conversations, hours and countries; no login).
• DataCamp figures: https://theeducationalequalityinstitute.org/api/datacamp/partner-stats (counted from DataCamp’s daily data).
• Any certificate: https://theeducationalequalityinstitute.org/verify/<code> answers valid, revoked, expired or not found.
• Skills Academy and its upcoming sessions: theeducationalequalityinstitute.org/programmes/education/skills-academy/
• The AI research behind Skills Academy: Can AI help people learn without making them dependent on it?

Figures were read on 1 November 2026, between 17:42 and 17:45 UTC, unless another date is given. DataCamp licences are from DataCamp’s dashboard on 8 October 2026, and room minutes are AWS usage from 1 July to 31 October 2026.