Yesterday Articulate announced Frontline, and the pitch is clean: anyone can now build training. Not the L&D team, but the sales enablement lead, the support manager, the IT lead, the subject-matter expert, each of them turning documents they already have into a full course kit through an AI agent called Nova, no authoring seat required. It is a genuinely capable piece of engineering. It also answers a question almost no learning team was actually asking, and it quietly makes the real one harder.
What Frontline actually does
The mechanics are impressive and worth stating plainly. Nova, the agent underneath, is trained on Articulate’s two decades of learning-design patterns. It takes source content and generates a connected kit — guides, quick references, interactive video, scenarios, a slide deck — in minutes, then translates it into eighty-plus languages and cuts an AI avatar video to narrate it. Frontline wraps that in brand governance and analytics, so what a support manager produces comes out on-brand and visible to the L&D team that set the rules. It connects to Google Drive, OneDrive, Thinkific, and the usual AI assistants. As a way to distribute a capability across an organization, it works.
The bottleneck was never the seat
Here is the thing the launch quietly assumes: that the reason more good training does not get made is that too few people can operate the authoring tool. That has not been true for a while. Producing a course was already the cheap, fast part. The constraint was never the seat or the software. It was the judgment about what to build, and whether to build anything at all. Frontline removes a limit that was not the binding one, and hands the button to people who were hired to do a different job.
That judgment is the whole profession. It is knowing that a request for a course on the new refund policy is often a sign the policy is written badly, not a training need. It is deciding what to leave out. It is designing a scenario so it changes what someone does on Monday rather than what they can recognize on a quiz. None of that lives in the source content, which means none of it is what Nova is working from.
What “governance” actually covers
Articulate is not naive about the risk, and the governance story is the answer to it: L&D sets the brand rules, keeps the analytics, retains oversight. That is real, and it is also narrower than it sounds. Brand governance controls whether the training looks right. It does not control whether the training should exist at all, or whether it teaches the correct thing. A support manager can now generate a polished, on-brand, fully governed course for a problem that a two-line edit to a help article would have solved — and it will pass every check Frontline applies, because none of those checks ask the question a professional asks first.
Why this is a problem for the field
Scale is what turns a useful tool into an industry problem. Articulate is talking about a hundred thousand generation credits per subscription and a library it says has already produced a hundred and forty million lessons. Point that capacity at every team in a company and the predictable result is not better training. It is far more training — confident and unvetted — competing for the same finite attention, most of it made by people with no reliable way to tell a performance problem from a training one. The completion dashboard will look busier. Very little about what people can actually do will change.
The deeper cost is what the framing does to the profession. For a decade the better parts of L&D have been trying to move from order-taking content shop to a function that diagnoses performance and owns outcomes. A product whose headline is “anyone can make the training now” pulls hard in the other direction. It tells every executive that the skilled part was the production, and the production is now free — which is the exact misread that ends with a capable L&D team cut because a tool appeared to do its job.
What the professional still does
None of this makes Frontline useless, and pretending it is would be its own kind of dishonesty. For genuinely local, low-stakes knowledge — a support manager documenting a process they own, a product team turning release notes into a quick reference — distributed authoring is a reasonable use, and doing it inside L&D’s brand and analytics beats the unmanaged sprawl those teams already create in slides and wikis. Used that way, with a professional setting the guardrails and reading the analytics, it is a fine addition to the stack.
What it does not do is the work that made the job a profession. Someone still has to decide whether training is the answer, find the one behavior that actually moves the number, design for that behavior instead of for recall, and go back later to see whether anything changed. Frontline generates the artifact. It does not do the thinking that decides whether the artifact should exist, or whether it worked. That thinking is the job, and it is the one thing in the announcement that is not for sale.
Anyone can make training now. That was always the easy part. Making training was never the job — solving the problem the training was meant to solve was, and that still takes a person who knows the difference.