69 questions, answered exactly.

The shape of the product, before the feature list.

1. What is ArthurAI™ Studio?

   Studio is a course-authoring Digital Employee. You describe a course in plain language and Arthur authors the whole thing — modules, lessons, learning outcomes, worked examples, practice activities, gradable assessments and a delivery schedule — then reviews its own output with a panel of specialist reviewers before handing it to you. You edit anything you want, approve it, and deliver it to the LMS you already run or to a hosted link. It is a product that produces finished courseware, not a text box you assemble courseware in.

2. How is this different from asking a general-purpose model to write a course?

   A prompt returns a draft. Studio returns a course. A team of Arthur agents authors the lessons in parallel on a live build board, grounded in and cited to the material you supplied, aligned to the frameworks you named while it writes rather than checked afterwards. A panel of reviewers then audits alignment, readability and structure, disputes each other’s findings, and a lead reviewer rules on what they cannot settle — and Arthur fixes what it can of its own output before you see it. What arrives is assessed, scheduled, exportable to your LMS, and carries the record of how it was made.

3. Who is Studio for?

   Three groups, and the free plan is genuinely useful to all of them. Independent creators and trainers who need a complete course faster than they can write one. Departments and multi-author teams who need review, approval and consistent house style across several authors. And regulated organisations who need training authored from their own policies and regulations, with credentials an auditor can verify — that last group is what Compliance Autopilot is for.

4. Do I still need an instructional designer?

   Studio does the work an instructional designer would do first: structure, sequencing, learning outcomes, assessment design, and the alignment pass against your frameworks. What it does not do is decide what your organisation should teach, or take responsibility for it. Teams who have an instructional designer use Studio to remove the drafting and leave the judgement; teams who do not have one get a defensible structure they would otherwise not have had.

5. What does "agentic" actually mean here?

   That the work is decomposed and delegated rather than generated in one pass. Lessons are authored concurrently by separate agents, each grounded in the retrieved passages relevant to it. Reviewers run as their own agents and read each other’s findings. The run heals itself: a stalled step retries at lower effort rather than failing the job, and bounded watchdogs mean a run never hangs. You watch all of it happen — the reasoning streams as it goes.

6. What AI powers Studio?

   Eve-Education™ F5/reasoner, part of the Eve-Fusion™ family — a composition of five cooperating reasoning models calibrated for educational reasoning, running on Eve-Grid™ Azure infrastructure through a private endpoint authenticated by managed identity, with no API keys to leak. Reasoning effort is calibrated per step, highest where a mistake would cascade through every lesson beneath it.

7. Is our content used to train the models?

   No. Your content, organisation profile, frameworks and authored courseware stay inside your own tenant and are not used to train the underlying models. The reasoning capability is built on Eve-Genesis™ synthetic training data.

From a brief to a finished, teachable course.

1. What do I have to give it to start?

   A topic and a sentence about what learners should be able to do afterwards. Everything else — hours, weeks, depth, edition, language, assessment threshold — has a sensible default you can change. If you have material of your own, add it; if you do not, Arthur researches the subject and your organisation and proposes a full brief with the reasoning behind every recommendation, then waits for your go before building anything.

2. How long does a course take to author?

   Minutes, not weeks, for a complete multi-module course with assessments — and you watch it happen rather than waiting on a progress bar. The variable is length and depth, not queue time: a longer course means more lessons authored in parallel.

3. Can Arthur write from our own material?

   Yes, from Creator up. Add a link, paste text, or upload a PDF. Studio splits it, embeds it, and retrieves the passages that matter while Arthur writes, so the course teaches from your material rather than from the open internet. Sources stay attached to the course, and what was actually used is recorded in a grounding manifest with honest coverage counts.

4. We have an expert but nothing written down. Can Studio use that?

   Yes, and this is one of the two things Studio does that a prompt box cannot. Arthur interviews your subject-matter expert in a guided conversation, captures the transcript as a grounding source, and authors a complete cited course in that expert’s own substance. Institutional knowledge that lived in one person’s head becomes teachable courseware. Interviewing is available on every plan, including free; saving the transcript as a reusable source needs a paid plan.

5. Can Arthur search the web while it writes?

   If you ask it to. It is off by default and opt-in per course, from Creator up — because a course grounded in your own material and a course grounded in whatever is currently on the internet are different products, and that should be your decision rather than a default. When it is on, what it found is cited.

6. Can I edit what it produces?

   All of it. Every course opens in an editor: rewrite a lesson, adjust an outcome, reorder or delete modules and lessons, drop in your own media. Or describe the change in plain English — "make this more advanced", "add a real-world scenario", "simplify the language" — scoped to the whole course, one module, or one lesson, and Arthur revises it. Revisions run as a job on the server with the reasoning streamed, so closing the tab does not lose the work.

7. What if a revision breaks the quizzes?

   It does not leave them broken. A quiz whose lesson has been rewritten under it no longer matches what it is testing, so Studio drops it and writes a replacement against the new text — only for the lessons that actually changed, keeping the pass mark already in use. If a replacement cannot be written, you are told exactly which lesson is without one rather than discovering it later.

8. What languages can a course be in?

   Author in the language you brief in, and translate a finished course on one tap into Spanish, French, German, Portuguese, Arabic, Hindi, Simplified Chinese or Japanese — on every plan, including free. Studio produces a new draft course and leaves the original untouched, verifies the structure survived the translation, and where a lesson fails that check it keeps the original text and tells you which one. Right-to-left languages are laid out as such, through the hosted player and into the export.

9. How many courses can I author?

   Unlimited, on every plan. What bounds a plan is the monthly authoring budget — 1.5M tokens on Free, 12M on Creator, 60M on Team, custom above — not a course count. That is the real governor and it is published rather than discovered at the limit.

10. What happens when I reach the monthly budget?

    Authoring pauses until the month rolls over or you move up a plan; nothing you have already made is affected or locked. Usage is visible per workspace as you go — what you have spent, how it sits against the budget, and which courses accounted for it — so the limit is something you can see approaching.

The part most buyers are actually worried about.

1. How do I know the course is any good before I publish it?

   Studio tests it, which is the thing that does not exist elsewhere. Four simulated learners — a complete beginner, an experienced practitioner, a time-pressed skimmer and a non-native English speaker — each read the course as that person and file what tripped them: where they got lost, what it assumed and never taught, which assessment asks about something the content never covered. A structural lint checks for modules with no assessment, lessons with no objectives, and content too thin to teach from. Every finding arrives with a ready-to-run instruction, so the fix is one tap. All of it is on the free plan.

2. What is the reviewer panel?

   A panel of specialist reviewers that audits the course before you see it — alignment, readability and structure — and then argues. Reviewers read each other’s findings and dispute the ones they think are over- or under-called; a lead reviewer rules on what they cannot settle. Arthur then fixes what it can of its own output and verifies the fix actually landed. You can watch the whole deliberation, and you can re-run it on demand at any time.

3. What is the health score, and does it move when I fix things?

   It is computed from the findings that survive the panel, weighted by severity, and yes — it is strictly monotonic, so every finding you clear raises it, at every point on the scale. Re-running the health check records the new verdict and updates the score on the course. A course you have fixed does not keep reporting the score it had when it was written.

4. Courses go out of date. Does Studio notice?

   Yes. A freshness pass looks for claims that have aged — a figure with a year on it, a rule that has since changed, a tool that has been renamed — and for statements that have drifted from current practice. Each finding says what to verify and carries the revision instruction that fixes it. These are things to check rather than confirmed facts: the scan can tell you a claim has probably aged, not what replaced it, and it says so.

5. Does Arthur review my workspace without being asked?

   A sweep runs on a schedule across the courses you own and raises what needs attention: a course whose health has slipped, a draft that has gone quiet, one that is structurally ready to publish. It is deliberately deterministic — rules, not a model guessing — so the same course produces the same notice every time.

6. What stops it inventing things?

   When a course is grounded on your sources, Arthur writes from retrieved passages and the citations are recorded with the course. In Compliance Autopilot it is stricter: every assertion is cited to the passage it came from, and any quote that cannot be found in that passage is dropped rather than kept. Beyond that, the honest answer is the governance one — a named owner approves every course before a learner sees it, and the software enforces that rather than assuming it.

7. Can I see how a course was made?

   Every course carries "How Arthur authored this" — its own account of why it is sequenced the way it is, the steps that ran, the sources used, and a content hash that fingerprints exactly what was produced. Versions, reviewers and approvals are recorded with actor and timestamp.

Getting it signed off by people who do not have accounts.

1. Can someone outside our team review a course?

   Yes, with no account, no seat and no licence. Mint a reviewer link scoped to one course; your reviewer reads it lesson by lesson and leaves comments tagged to the lesson they are about. Links expire and can be revoked the moment you want them closed. Comments come back into the workspace where you resolve them, or hand the whole set to Arthur to address at once. On every plan, including free.

2. Can I stop a course being published before it is approved?

   It already is stopped. Publishing refuses a course that has not been approved — it is a rule in the software, not a step in a process document. Submit for review, approve, or send back with changes, on every plan. And editing an approved course drops it back to draft, so an approval always attests to the exact content that shipped rather than to something it used to be.

3. What roles are there?

   Owners approve, transfer and delete; contributors create and revise; readers observe. The roles are enforced at the API, not hidden in the interface, so a reader cannot reach a write route by any means. Owners invite by email and change or remove a member’s access.

4. Can I roll back a bad revision?

   Yes. Save a named version before a large change and restore it if the change went somewhere you did not want. Version history is on every plan.

5. Can several people work in the same workspace?

   Yes, from Team up for seats — 10 on Team, unlimited on Business and Compliance Autopilot. Free and Creator are single-seat. Workspaces are separate containers with their own courses, members and settings: 1 on Free, 3 on Creator, 10 on Team, unlimited above.

6. Can I set a house style once, for everyone?

   Yes, on every plan. "Personalize Arthur" holds your voice, terminology, default edition, depth, language and assessment rigor, and it rides on every generation and every revision — so a course authored by a colleague six months from now inherits the same register as one you author today.

Where a finished course can go.

1. Can I deliver without an LMS?

   Yes, on every plan including free. Publish to a hosted player and share a link. Progress and checkpoint results persist for the learner, quizzes are graded on the server rather than in the browser, and an embed mode drops the course into a page you already run. Each course gets a share card generated for it, so the link looks like something when it is pasted.

2. Which LMS formats do you export?

   SCORM 1.2, xAPI (Tin Can) and PDF from Creator up. SCORM 2004 (4th Edition) and a facilitator’s instructor kit in Markdown from Team up. Packages import into Moodle, Canvas, Blackboard or any conformant LMS. The SCORM 2004 package carries its run-time reporting over SCORM 1.2.

3. Do you support LTI 1.3 with grade passback?

   Yes, on Compliance Autopilot. Learners launch straight from your LMS and grades post back automatically over Assignment & Grade Services. It is not available on Free, Creator, Team or Business — the pricing matrix on the Studio page shows the same thing, and it is worth being exact about because it is a common assumption.

4. Can I deliver into an ArthurAI™ edition?

   From Team up, by export. Studio produces a signed provenance envelope carrying a content hash; the receiving edition re-verifies it byte-for-byte on import and rejects it on mismatch. You choose per course whether the edition renders it verbatim as you authored it, or re-expresses it in that edition’s own voice.

5. Is the exported course accessible?

   The export produces a semantic HTML page per lesson and lays out right-to-left languages as such. We do not publish a Studio-specific accessibility conformance claim; the platform’s accessibility posture is documented in the Trust Center, and we would rather point you there than assert a certification Studio does not separately hold.

6. Can a learner take a course on a phone?

   Yes. The hosted player and the authoring surface are both built for phones and tablets, not just narrowed for them.

7. Can I white-label what learners see?

   From Team up. Your product name, author name, accent colour and logo on the learner-facing pages.

Checking that the learning stuck, and proving it.

1. Does Studio write the assessments too?

   Yes, as part of authoring. Multiple-choice checkpoints with a stored answer key, distractors written to be plausible to someone who read carelessly, and an explanation that teaches why the right answer is right. Questions are written only from what the lesson actually states — an item whose correct option is not among its own options is discarded rather than repaired, because a question whose answer is missing grades every learner wrong.

2. Where is a quiz graded?

   On the server, never in the learner’s browser, so the answer key is never sent to the person being tested. The attempt cap is enforced in the database, so a retry cannot be bought with a page refresh.

3. Can I set a pass mark?

   Yes, per course, as part of the brief. Graded checkpoints carry the threshold you set, and it is inherited when a quiz is rewritten after a revision.

4. Do you issue certificates?

   Compliance Autopilot issues a verifiable credential when someone passes a server-graded attestation: the training they completed, an issuer, a score and an expiry, with a public link anyone can check without an account. It is a record confirmable at source rather than a certificate image. Credentials are not available on Free, Creator, Team or Business.

5. Can a credential be revoked?

   Yes, by an owner, and the public verification link reflects it. Credentials also expire on the date they carry, and the dashboard shows who is current, expiring and lapsed.

Regulated workforce training, end to end.

1. What does Compliance Autopilot do that authoring a course does not?

   The tier for organisations whose training has to satisfy a regulator. You define a programme against the roles and jurisdictions you cover; Arthur authors the training directly from your own regulations, policies and SOPs, citing every assertion to the passage it came from. You assign it to people, they are credentialed on passing, and the programme tracks who is current, expiring and lapsed.

2. How is it different from just authoring a course about a regulation?

   Three things a course does not have. Citations are enforced rather than encouraged — a quote Arthur cannot find in the passage it claims is dropped. Every generation, question and credential is recorded in a per-programme audit trail. And the output is a programme with coverage across roles and jurisdictions, not a single artefact.

3. What does the auditor actually get?

   A report, produced on one click, of the programme scope, the requirement coverage, the trainings cited against each requirement, and every credential issued with its expiry — carrying a SHA-256 content checksum so the document can be shown to be the one that was generated. Individual assertions can be traced to the source passage they came from.

4. How does recertification work?

   Every credential carries an expiry date, and the programme dashboard shows who is current, who is expiring and who has lapsed. When a credential lapses the person retakes the training, and the new credential dates from the day they pass rather than from the old expiry.

5. Do employees need accounts?

   No. You can mint a self-service training link per person — they take the training and are credentialed without an account and without an LMS. Or launch the whole roster from the LMS you already run over LTI 1.3.

6. Can I ask questions of my own policy set?

   Yes. A grounded expert answers questions against the programme’s own sources, with the passages it used, so "what does our policy say about X" is answerable without reading the whole document set.

7. How do I buy it?

   It is quoted to your workforce rather than bought self-serve — talk to us. Everything in it sits above the Business plan, and it is the only tier carrying verifiable credentials and LTI 1.3.

What each plan actually does, exactly.

1. Is the free plan a trial?

   No. It does not expire and it does not need a card. On free you author unlimited complete courses, run the simulated-learner, health and freshness scans, translate into eight languages, use seatless reviewer links, approve and govern, publish to a hosted learner link, keep version history, and use two-factor authentication and full data rights. What it does not have is grounding on your own sources, export, and anything above those.

2. What do the paid plans add?

   Creator at $149/mo adds grounding on your own sources, live web research while authoring, and SCORM 1.2 / xAPI / PDF export. Team at $690/mo adds SCORM 2004, the instructor kit, delivery into an ArthurAI™ edition, white-label, the learner-signal improvement loop, and 10 seats. Business is a custom institution license with unlimited seats and workspaces, every edition, and dedicated support. Compliance Autopilot is quoted to your workforce and carries verifiable credentials and LTI 1.3.

3. What are the real limits on a plan?

   Seats, workspaces and the monthly authoring budget. Courses are uncapped on every plan. Free is 1 seat, 1 workspace, 1.5M tokens per month; Creator 1 seat, 3 workspaces, 12M; Team 10 seats, 10 workspaces, 60M; Business and Compliance Autopilot are unlimited seats and workspaces with a custom budget.

4. Can I buy without talking to anyone?

   Creator and Team, yes — start free and upgrade in the product. Business and Compliance Autopilot are quoted to the institution, because both are scoped to seats, editions and support commitments rather than sold from a rate card.

5. What is the learner-signal improvement loop?

   The second of the two things a prompt box cannot do. Publish a course to an ArthurAI™ edition, and bring that edition’s aggregate, PII-free learner signal back into Studio as a file — no live link between the systems and no credentials shared. Arthur ranks the weakest lessons, explains why they are weak, and re-authors any one of them on a click. Team and above.

6. Do you charge per learner?

   No. Studio is priced on authoring — seats, workspaces and the monthly authoring budget. How many learners take a published course does not change what Studio costs.

7. What happens to my courses if I downgrade or stop paying?

   Your courses remain yours and remain in your workspace. Capabilities that belong to the higher plan stop being available — export, for instance — but nothing you authored is deleted for being on a smaller plan, and you can export before you move down.

The questions procurement asks first.

1. Is our workspace isolated from other customers?

   Yes, and it is enforced in the database rather than in application code. Every tenant-scoped table carries row-level security that fails closed, so a query without a workspace context returns nothing rather than returning someone else’s rows. Roles are enforced at the API gate as well as in the interface.

2. Where does the AI run?

   On a private Azure AI Foundry endpoint authenticated by managed identity — no API keys in code or configuration to leak. Prompts and responses are not logged, and a third-party model name never appears on any customer surface.

3. Can I turn on two-factor authentication?

   Yes, on any plan, including free. Time-based one-time codes from the authenticator app you already use, with recovery codes issued once and shown once. An account that can author your organisation’s training is worth protecting whatever you pay for it.

4. Can I get all my data out?

   Yes, yourself, without a support ticket. Download a structured export of everything held about you, whenever you want it.

5. Can I delete my account and everything in it?

   Yes. You can see exactly what a deletion would remove before you confirm it, then delete — immediately and irreversibly. There is no grace period to explain away and no queue it sits in.

6. Who at MindHYVE can see our courses?

   Support access is not ambient. Operator access to a customer workspace is a deliberate, recorded impersonation session with a start, an end and an audit record — not a standing back door into your content.

7. Is there an audit trail?

   Yes, per tenant and append-only, with actor, resource and timestamp, exportable to CSV for auditors. Course-level history — versions, reviewers, approvals — is carried on the course itself.

8. What is your compliance posture?

   ArthurAI’s platform-level security and compliance documentation lives in the Trust Center, including data handling, subprocessors, the DPA and incident response. Studio is not separately certified, and we would rather say so than let a platform attestation be read as a Studio one. For procurement documentation under NDA, talk to our team.

What the first hour looks like.

1. How do I start?

   Create an account and author a course. There is no trial to activate, no card to enter and nothing to install — Studio runs in the browser. A first-run panel and a template gallery give you a running start if you would rather not begin from a blank brief.

2. Is there help inside the product?

   Yes — a searchable help library, guided tours of the main surfaces that you can replay whenever you want, and a getting-started checklist that tracks what you have actually done rather than what you have clicked. You can also raise a support request from inside the product rather than finding an email address.

3. Can I talk to Arthur about what to build?

   Yes, on every plan. Arthur Chat plans, searches the live web, reads your course library, opens any outline and searches your own history before answering — with every step shown as it happens. An answer you like becomes a course: it distils the conversation into a brief and prefills the authoring wizard, or adds the material to a course you already have.

4. Can Studio tell me what we should build next?

   It builds a skills graph across everything in your workspace — what your courses collectively teach, where they overlap, and where the gaps are — and proposes courses that fill the gaps, each one a single tap into the authoring wizard with the brief already filled in.

5. What if I have a question that is not here?

   Talk to our team. Institution and Compliance Autopilot conversations start the same way, and a demo is the fastest route to whether Studio fits how you actually work.