eLearning Authoring Tools: Which Tier You Need

Someone has handed you a folder. Inside it is a 60-page policy PDF, a deck the product team made for a conference, two recorded calls and a spreadsheet of the questions new hires keep asking. The request is simple to say and painful to do: turn this into a course, and make it work on the company LMS by the end of the month.

So you start looking at elearning authoring tools, and within an hour you are lost. Every product page says the same four words — interactive, responsive, SCORM-compliant, AI-powered — at prices ranging from free to the cost of a junior salary. The comparison articles rank them one to fifteen without ever telling you which of the fifteen were even built for your situation. And none of them answer the question you actually have, which is not "which tool is best" but "which kind of tool am I supposed to be buying at all."

This guide sorts the category into four tiers, shows which tiers produce a SCORM package and which quietly do not, adds up what a seat really costs once you count the hours of the person building, and walks through generating a course from documents you already own — including the five places that goes wrong. Every product detail here was checked against the vendors' own pages in August 2026, because tier names and AI features in this category change faster than review articles get updated.

The four tiers of elearning authoring tools

The phrase covers four genuinely different products. They have different buyers, different failure modes and price gaps of more than twenty times between them. Almost every disappointed purchase in this category is someone who bought one tier while needing another.

Tier 1 — Slide converters

These take a deck you already made and wrap it so an LMS can play and track it. iSpring Suite is the best-known example: it installs as a ribbon inside PowerPoint, so the authoring surface is the software your subject-matter experts already know. You build slides the way you always did, then click a button in the ribbon to add a quiz, record narration over each slide, and publish the result as a package the LMS understands.

What you get is a linear course. Slide one, slide two, slide three, a quiz at the end, a completion signal back to the LMS. What you do not get is branching, drag-and-drop exercises, or software simulations. For compliance refreshers, policy rollouts and product updates, linear is not a limitation — it is the correct shape. Most people who need "interactivity" in this situation actually need "the LMS must record that Trang finished it," and a slide converter does that perfectly well.

The catch is that a slide converter inherits every flaw of the deck it converts. A 90-slide deck full of paragraph-length bullet points becomes a 90-slide course full of paragraph-length bullet points. The tool does not rescue bad source material; it packages it faster.

Tier 2 — Structured course builders

These give you a browser-based editor with a fixed set of blocks: text, image, video, accordion, tabs, sorting activity, knowledge check. You assemble blocks into lessons and lessons into a course. Articulate Rise 360 and Easygenerator are the ones you will meet most often.

The trade is deliberate. You cannot place an element at an arbitrary pixel, so you also cannot produce something that looks broken on a phone. The output is responsive by construction. A subject-matter expert with no design background can produce something that looks professionally made on their first afternoon, which is exactly why this tier has spread through companies that never had an instructional design team.

Where it runs out is anything the block library does not cover. If your training requires a learner to practise a real procedure — click through the actual billing screen, make a judgement call that changes what happens next, get it wrong in a safe place — you will hit the wall of the block list and start doing awkward workarounds. That is the moment to look one tier up, not to fight the tool.

Tier 3 — Timeline and interaction builders

Articulate Storyline 360, Adobe Captivate and Lectora sit here. These are closer to an animation tool than a document editor: a timeline, layers, states, variables, triggers, conditional logic. You can build a branching scenario where a wrong answer at step two produces a different customer reaction at step five. You can record a software simulation and let people practise it. You can build assessment logic no block library will ever ship.

The power is real and so is the cost, and the cost is not mostly the licence. This tier needs someone who knows it. A person who has built in Storyline for two years will produce in three days what a beginner produces in three weeks, and the beginner's version will have bugs that only appear on someone's iPad. If you have that person, this tier is where the interesting work happens. If you are hoping to hand it to a busy operations manager as a side task, it will not get used, and the licence will renew anyway.

Tier 4 — Document-to-course generators

The newest tier, and the one that changed the shape of the category. Instead of starting from a blank editor, you upload the material you already have — the policy PDF, the conference deck, the transcript — and the tool produces a first draft of the course from it. Articulate's AI Assistant does this inside Rise. Easygenerator ships its own version. Orova Training is built around this direction from the start, and generates multiple output formats from the same upload rather than a single course draft.

What makes this tier different is not speed for its own sake. It is where the work starts. In tiers one to three you begin at an empty page and pay in hours to fill it. In tier four you begin at a draft that is roughly 70 per cent right and pay in judgement to fix the 30 per cent that is wrong — which is a completely different kind of work, done by a different person, and usually a much shorter list of hours. Section five covers exactly which 30 per cent tends to be wrong.

Which tier for which training job

Tiers are easier to choose once you stop thinking about features and start with the job in front of you. Five jobs that come up constantly, and where each one lands:

Annual compliance refresher. Fixed content, large audience, the only real requirement is a defensible record that each person completed it and passed. Tier one, or tier four if the source policy already exists as a document. Anything more elaborate is money spent on an audience that will click through it in one sitting either way.

New-hire onboarding. Long, mixed content, and it changes every few months as the company changes. Tier two earns its keep here because editing is cheap and the output looks the same on a laptop and a phone. Onboarding is also the job where a printable version matters most, because someone's first week involves a lot of standing next to somebody else's desk.

Software rollout. People must practise on screens that look like the real thing. This is the clearest case for tier three, because a recorded simulation where a wrong click produces the real error message teaches something a slide about the same screen cannot.

Sales enablement. Product knowledge, competitor positioning, objection handling. The content changes constantly and the audience is small, senior and impatient. Short modules plus flashcards beat a long course here, which pushes towards tier four with a strong flashcard output — the recall problem is the actual problem.

Certification you sell or hand to a client. The output has to carry your branding, produce a real score against a threshold, and issue a certificate that survives scrutiny. Whichever tier you pick, this job is decided by the assessment and certificate side rather than the authoring side, so test that part first.

Notice that three of those five jobs do not need the most powerful tier, and one of them needs it absolutely. That distribution is why "which is best" is the wrong question and "which of these five am I doing" is the right one.

Do all elearning authoring tools export SCORM?

SCORM is a packaging and messaging standard. A SCORM package is a ZIP file with a manifest that tells an LMS what is inside, plus JavaScript that reports back four things while a learner works through it: that they launched, how far they got, what they scored, and whether they passed. Every corporate LMS on the market can import one. That is the whole reason the standard refuses to die despite being older than the smartphone.

The four standards you will be asked about

SCORM 1.2 is the safe default. It is limited — one score, one completion status, no fine-grained tracking — but it is universally supported. If you are handing a package to a client or a parent company and you do not know what their LMS is, this is what you send.

SCORM 2004 added sequencing and navigation rules, so a course can enforce an order, plus multiple objectives per course. It has been frozen since the 2000s and remains widely deployed in enterprise LMS platforms. Ask before you publish in it, because a handful of older or lighter systems import 1.2 only.

xAPI (sometimes still called Tin Can) is not a package format in the same sense. It is a way of sending statements — "Nam completed the safety walkthrough on the shop floor tablet" — to a Learning Record Store. It captures things that happen outside the LMS, which SCORM structurally cannot. It also requires that Learning Record Store to exist and someone to look after it, which is where most xAPI ambitions quietly stop.

cmi5 is the bridge: xAPI's rich data, wrapped in SCORM-like launch rules so an LMS can start it predictably. It is the standard most often described as the intended successor. In practice, in 2026, you will still be asked for SCORM by nearly every buyer, and cmi5 will be an option you have rather than a requirement you meet.

What "SCORM-compliant" on a pricing page does not tell you

Three things, all of which have burnt someone:

Which version. A page saying "SCORM support" without a number is worth a support ticket before purchase, not after.

Whether export is on your plan. This is the common one. Several products in this category treat SCORM export as a paid upgrade or a feature of the annual tier, while the entry plan publishes to a hosted link only. Easygenerator, to be fair to it, lists SCORM export on all its published plans and adds xAPI from the Team plan up — but you should verify this pattern for whichever product you are looking at, because "compliant" and "included on your plan" are different claims.

Whether the interactive parts survive. A course can import cleanly and still fail to report. Videos hosted on the vendor's servers may not play behind a corporate firewall. Custom interactions built in tier three sometimes report completion but not score. The only reliable test is to publish a two-slide throwaway course, import it into the real target LMS, click through it as a test learner, and confirm the score lands. Do this in week one of the trial, not in week three of the project.

What a licence really costs once you add the build hours

The published price is the smaller number, and it is not close. But start with the published number, because it sets the floor and it is checkable.

Published prices, checked August 2026

On Articulate's own pricing page, Articulate 360 AI is listed at $1,749 USD per user per year for Teams and $1,449 USD per user per year for Personal. Both include Storyline 360, Rise 360, Review 360, the content library and the AI Assistant, and both list SCORM, xAPI, cmi5, AICC, web and PDF as export options. The Teams plan additionally includes a built-in LMS for up to 300 active learners a year, single sign-on and the enterprise security certifications.

On iSpring's own pricing page, iSpring Suite is listed at $970 USD per author per year, iSpring Suite AI at $1,290, and iSpring Cloud AI at $720, all per author per year with a 14-day trial.

On Easygenerator's own pricing page, Pro is listed at $1,399 per year for one author, ten courses and one hundred learners; Team at $6,995 per year starting from five authors with unlimited courses and learners; and Enterprise at custom pricing.

Adobe Captivate is sold as a subscription rather than a per-seat suite; Adobe publishes the current figure on its own product page, and it is the one number in this section worth checking yourself on the day you buy, because it has moved more than the others.

The number nobody publishes

Ask any team that has done this how long one finished hour of course content takes to build, and the honest answer is a range with a wide floor. A linear slide-based module, source material already written, someone experienced at the keyboard: a few working days per finished hour. The same finished hour as a branching scenario with custom interactions in a tier-three tool: multiples of that, plus a review cycle, plus the testing pass on real devices.

Now put an hourly rate on the person doing it — an instructional designer, an operations lead, or the subject-matter expert who is not doing their real job while they build. Against that, a $1,749 licence stops looking like the main cost and starts looking like a rounding error. This is the arithmetic that makes tier four interesting: it does not reduce the licence line, it reduces the line underneath it.

The same trap shows up when buying any per-seat business software. We wrote about it in the context of hiring tools in this look at Greenhouse applicant tracking — the sticker price is knowable, the implementation and the internal hours are not, and the second number is usually the larger one.

The same 40-minute course, built three ways

Take one concrete brief: a 40-minute refresher on a revised refund policy, for 220 staff across four sites, with a pass mark and a certificate. Here is how the work distributes in each tier.

In a slide converter. Someone opens the existing policy deck, rewrites the bullet points into something a person can read on a screen, records narration slide by slide, adds ten questions and publishes. Most of the elapsed time goes into the narration and the rewriting, and the rewriting is the part that decides whether the course is any good. There is no design work, because there is no design surface. Output: a linear course that will look identical to last year's, which is fine, because refund policy training is not where you want surprises.

In a structured builder. Someone assembles blocks: an intro, three lessons, an accordion of edge cases, two knowledge checks, a final quiz. It looks noticeably better than the slide version and works properly on a phone, which matters for the two sites where staff have no desk. The extra time over the slide route goes into choosing and arranging blocks rather than into content, and that is the honest trade — you are paying hours for presentation.

In a timeline builder. Someone builds a branching scenario: a customer asks for a refund outside the window, and the learner's choice at step one changes what the customer says at step three. This is genuinely the best version of the three for teaching judgement, and it is the only one that will still be discussed by staff a month later. It also costs several times the build hours, needs a device testing pass, and will need rebuilding rather than editing when the policy changes again next year.

All three are correct answers to different questions. What separates them is not quality of the tool but how much judgement the training has to transfer — and refund edge cases are judgement, while the refund window is a fact. Split the brief and you can often run the facts through the cheap route and reserve the expensive route for the twenty minutes that actually need it.

Four costs that do not appear on any pricing page

Review cycles. Legal wants a wording change on slide 14. In a slide converter that is two minutes. In a branching scenario where slide 14 is referenced by three triggers and a variable, it is an afternoon plus a retest.

Updates. Policy documents change every year. A course built as a monolith gets rebuilt every year. A course built as small modules gets one module swapped. Nobody thinks about this in month one and everybody pays for it in month fourteen.

Translation. If you operate across Vietnam, Thailand, Laos and Cambodia, ask how the tool handles a second language before you buy. Some export a translation file and re-import it cleanly. Others require rebuilding the course. The difference across five languages is enormous.

The seat that stops being used. The most expensive licence in this category is the tier-three seat bought for a person who left, or for a manager who never had time to learn the timeline. It renews silently. Count the seats you will genuinely staff, not the seats you hope to.

When PowerPoint is already enough

Sometimes the correct purchase is nothing. Four situations where a dedicated authoring tool adds cost without adding outcome:

The audience is under thirty people and sits in one place. A deck, a recording, a shared folder and a short quiz will get the job done. The overhead of licensing, learning and maintaining an authoring tool is larger than the problem.

The content will be true for three months. A product launch briefing, a campaign playbook, a temporary process during a system migration. Build it fast, deliver it, throw it away. Do not invest in packaging something with a short life.

Nobody is asking for a completion record. The single strongest argument for an authoring tool is that the LMS needs to record who finished. If no auditor, client or regulator will ever ask, you have removed the main reason to package anything.

The material is not written yet. This is the one that catches people. If the policy itself is still being argued about, buying a tool does not move you forward. Authoring tools format decisions; they do not make them. Write the material first.

There is a real cost to skipping the tool, and it is honest to name it: you lose tracking, you lose version control, and six months later there will be four different copies of the deck on four different laptops with nobody sure which one is current. That is a genuine problem — it is just a smaller problem than the one an unused annual licence creates.

Generating a course from documents you already have

This is the workflow most teams are now actually asking about, so it is worth walking through properly, including the failure modes.

How the workflow runs

Step one — gather sources. The policy document, the standard operating procedure, the deck, the recorded call transcript. Upload them together rather than one at a time; the draft is better when the model can see how the same idea is described in two different documents.

Step two — set the shape before generating. Who is this for, how long should it be, what should someone be able to do afterwards. Most tools now let you specify target length; Articulate's AI Assistant in Rise, for example, lets you choose how long the course should be when you prompt for a draft. Skipping this step is the single most common reason the output feels generic.

Step three — generate and read the outline first. Not the slides. The outline. If the structure is wrong, fixing slides is wasted work.

Step four — cut, hard. The first draft is almost always too long. Generated content is fluent, which makes it easy to keep, which is exactly the trap.

Step five — check every fact against the source document. Then check the assessment questions separately, because they fail differently.

Step six — publish, import into the real LMS, click through as a learner. Every time. Not once at the start of the project.

The five places it goes wrong

1. The draft is fluent and hollow. Generated course text tends to explain the concept beautifully and skip the company-specific detail that was the only reason to build the course. The source says approvals over 50 million dong go to the finance director; the draft says "follow your organisation's approval process." Everything readable, nothing usable. The fix is mechanical: for each section, ask what a new hire could now do that they could not before. If the answer is nothing, that section is decoration.

2. Confident invention. The model fills gaps rather than leaving them empty. Where the source is thin, the draft becomes most authoritative. Invert your reading: spend your checking time on the sections that read most smoothly and cite nothing. We covered this pattern in more depth in this piece on making AI output pass the human test, and it applies word for word to course drafts.

3. Quiz questions that test recall of the slide, not the job. Generated questions cluster around definitions, because definitions are what the source text contains. "What does the acronym stand for" tests nothing. The question you want is a situation: here is a customer message, here is what the policy says, what do you do. Expect to rewrite most generated questions and keep the good ones — there is a fuller method in this guide to choosing an online quiz maker for staff training.

4. Tone drift across sections. Section two sounds like a friendly onboarding guide and section five sounds like a legal notice. Nobody notices while editing section by section; everybody notices reading it end to end. Read the whole draft in one sitting before publishing.

5. Nobody owns it after publication. Fast generation makes courses cheap to create, which quietly makes a library of forty half-maintained courses possible in a way it never was before. Put a named owner and a review date on every course at the moment it is published, or the speed advantage turns into a maintenance debt within a year.

Four formats from one document, and who uses which

The most useful thing about the newest tier is not that it drafts faster. It is that the same upload can produce several different outputs, and different people in the same company genuinely need different ones. Orova Training generates printable slides, an audio version, a video and AI flashcards from a single set of uploaded documents, then lets you attach a quiz and a certificate at the end.

That matters because a single format always leaves someone out.

Printable slides

The format people underestimate. Slides that print cleanly are what a supervisor takes onto the factory floor, what gets stuck on the wall next to the machine, what a trainer holds during an in-person session, and what an auditor asks to see. In workplaces where phones are locked away or connectivity is poor — which is most of manufacturing, logistics, food service and retail across the region — the printable version is not a fallback. It is the primary channel.

Audio

An audio version of the same material is for the commute, the delivery route and the shift break. It works because it demands nothing of the eyes and nothing of the hands. It suits narrative material — the reasoning behind a policy, a customer story, an explanation of why the process changed — and suits procedures badly, because nobody memorises step four of nine while riding a motorbike.

Video

Video is the default request and the most misapplied format. It genuinely wins for anything visual and sequential: what the screen looks like, where the button is, how the machine is supposed to sound when it is running correctly. It is a poor choice for reference material somebody needs to check quickly, because scrubbing a timeline to find one detail is slower than scanning a page.

Flashcards

Flashcards are the only one of the four formats aimed at what happens after the course. Everything else transfers information once; flashcards fight forgetting on a schedule. They are the right tool for terminology, product ranges, safety limits, error codes and policy thresholds — anything that must be recalled cold, weeks later, without a search. Pair them with a low-stakes practice test rather than a single final exam; the reasoning behind that is in this guide to using a practice test maker for certification prep.

The quiz and the certificate at the end

Two things turn material into training that counts. A quiz produces a score, which is what turns "we sent it round" into "eleven people scored below the threshold and need a second session." A certificate produces a record, which is what an auditor, a client or a regulator actually asks for. Whichever tier you buy, check that both exist and that the certificate can carry your own branding — a certificate with somebody else's logo on it is not one you can hand to a client.

An eight-point checklist before you buy

Run these in order. Any answer that stops you should stop you before the trial, not after the invoice.

1. Where does the finished course have to run? Your own LMS, a client's LMS, a parent company's LMS, or a link you send. Answer this first, because it decides whether SCORM export is a requirement or an irrelevance, and that single answer eliminates more products than any other question.

2. Who is physically going to build it? Name the person. If they are a subject-matter expert with no design background, tier three will not get used no matter how good it is. If they are an experienced instructional designer, tier two will frustrate them within a month. The buying mistake is almost never the tool; it is the mismatch between tool and builder.

3. What do you already have written? If you have a shelf of documents, the document-to-course tier removes the largest block of work. If you have nothing written, that tier removes nothing, because generation from an empty folder produces generic content.

4. How often does the content change? Annually or more often means favour tools where editing is cheap and modular. Once every three years means the build cost dominates and editing convenience matters less.

5. How many languages? Ask specifically how a second language is produced and whether translated versions stay linked to the original when it is updated. This question separates products more sharply than any feature list, and it is almost never on a comparison chart.

6. Do you need a score, a completion record, or both? Completion alone is a lower bar and widens your options. A defensible score, with a pass threshold and a certificate, narrows them, and should be tested on a trial account rather than taken from a feature list.

7. What does it cost in year two? Multiply seats by the annual price, then add the seats you will add, then check whether the price shown is the introductory one. Also check what happens to your published courses if you stop paying — whether the SCORM packages you already exported keep working is a different question from whether you can still edit them.

8. Publish one throwaway course during the trial. Two slides, one question. Export it, import it into the real target LMS, complete it as a test learner, and confirm the score arrives in the report. This test takes under an hour and catches more problems than a week of reading comparison pages.

Common questions

Is an authoring tool the same as an LMS?

No, and buying one expecting the other is the most common confusion in this category. An authoring tool builds the course. An LMS delivers it to people, records who completed it and reports on that. Some products now bundle a light version of the other side — Articulate's Teams plan, for instance, includes a built-in LMS for a capped number of active learners a year — but a bundled learner-delivery feature and a full LMS are not the same thing, and the cap is where the difference shows up.

Can I move my courses to a different tool later?

Partly. A published SCORM package keeps working in any LMS regardless of what happens to your subscription, because it is a self-contained ZIP. The editable source file is another matter: it is proprietary in every tool worth using, and there is no meaningful import path between vendors. Plan on the packages being portable and the source not being.

Do I need a separate tool for video?

Usually not any more. Tiers two, three and four all handle uploading and embedding video, and the newest tier generates a narrated video version directly from the source document. A dedicated screen-recording and editing tool still earns its place if you produce software walkthroughs regularly, where frame-level editing and callouts matter.

Is free ever the right answer?

For a small internal audience with no tracking requirement, yes — see the PowerPoint section above. The moment somebody outside your team needs proof that training happened, free stops being viable, because the proof is the product.

What to do this week

Answer questions one, two and three from the checklist honestly and write the answers down. Those three — where it must run, who will build it, what you already have written — narrow this category from fifteen products to two or three, and they do it before you spend a single hour in a trial.

Then run the throwaway-course test on the survivors. Not the full evaluation, not the feature matrix. One two-slide course, exported, imported into the LMS that will actually host it, completed as a test learner, with the score confirmed in the report. Everything that goes wrong in this category goes wrong at that boundary, and an hour spent there is worth a fortnight of reading.

And if your folder of documents is already full — the policy PDF, the deck, the transcripts — start from the tier that begins there. The blank editor is not a feature. It is a bill.