Ten years ago, corporate training lived in a portal, and that was uncontroversial. Today most vendors in the category deliver into chat in some form. That shift has causes worth understanding, because they determine whether it applies to your situation.
The portal assumed a learner who arrives
Every LMS built on a catalogue model assumes the same thing: the learner comes to the platform. They log in, browse, enrol, and work through material.
That assumption holds when learning is voluntary and self-directed. Someone pursuing a skill will absolutely go and find the course.
It fails completely for mandatory training, which is most of what companies actually assign. Onboarding, policies, compliance, process changes — nobody seeks these out. The model requires the learner to arrive, and for this category of training they don't, so completion depends entirely on chasing.
Chat delivery inverts the assumption. It doesn't ask anyone to arrive.
Remote work removed the scaffolding
Portal-based training worked better in an office than anyone realised, because the office supplied ambient support nobody designed.
A colleague mentions they've done it. Someone complains about it at lunch. A manager notices people clicking through it. There's a poster by the printer. All of it created awareness that the training existed and was currently a thing.
Distributed teams lost all of it simultaneously. The portal became socially invisible — nothing reminds anyone it exists. That's a large part of why programmes that seemed fine in 2019 quietly stopped working, and why the shift accelerated when it did.
The chat tool is now the only shared surface a distributed team reliably has. Training delivered there inherits the visibility the office used to provide.
The vendors followed
You can read the shift directly off the market. Continu's Eddy agent delivers training and cited answers inside Slack, Teams, and SMS. Arist delivers microlearning through SMS, Teams, WhatsApp, Slack, and email with no login. Doozy and Haekka are built Slack-first from the ground up. Litmos, 360Learning, TalentLMS and LearnUpon all ship Slack integrations.
The depth varies enormously — many of those integrations are notifications with a link out rather than training completed in chat — but the direction is unambiguous.
What actually changed for learners
The concrete difference is the number of steps between "assigned" and "done".
The portal model: notice a notification, decide to act, click through, authenticate, land on a portal, find the course, start, complete. Eight steps, most of them unrelated to learning, each losing people.
The chat model: read the message, answer the question. Two steps, both of which are the training.
That's the whole mechanism behind the completion figures these tools report — SC Training's 82% average on five-minute microlessons, Arist's stated 90%-plus. It's mostly the removed journey, not better content.
Where it doesn't apply
The limit is reach, and it's absolute rather than a matter of degree.
Chat delivery reaches people inside your workspace and nobody else. Customers, partners, contractors without accounts, and deskless staff are all out. If those audiences are central to what you do, you need a portal, and the trend doesn't help you.
It also doesn't suit self-directed learning, where a catalogue is genuinely the right interface, or purchased SCORM libraries, which chat-native tools can't ingest.
The likely end state
Not replacement — division. Portals for anything self-directed, external-facing, or purchased. Chat for mandatory internal training where completion is the problem.
Most companies have both kinds and will end up running both, which is a more stable arrangement than trying to force either model to cover everything.
If mandatory internal training is where your programme struggles, Poplearn is built for that half — documents become modules, modules arrive as Slack direct messages, and /progress gives live completion by team. Free under 20 seats.
