Beyond the LMS: understanding and choosing your training tools

What is an LMS and what is it for?

The LMS (Learning Management System) is the platform where an organisation stores its courses, assigns who must take them and checks that they have been completed. It plays content produced elsewhere, to the SCORM, xAPI or cmi5 standards, and produces the reports needed for management, audits and funding bodies. Some are open source, such as Moodle (GNU GPL v3 licence, no licence fees), and others are SaaS, billed per registered or active user.

Features, standards and public price lists are detailed in What is an LMS? Definition, features and pricing.

Which families of tools should be distinguished?

Market vocabulary is confusing, because vendors add functions from one category to another. The useful criterion is function:

Family	What it does	What it measures
LMS	Manage and deliver	Activity, completion, scores
LXP	Aggregate and recommend content	Engagement
AI authoring tool	Produce modules faster	Nothing of its own: it feeds the LMS
Adaptive learning	Adjust the path to the learner's answers	Performance within the path
AI training system	Generate the path from the skills gap	Acquisition of target skills

The full distinction is set out in LMS, LXP, adaptive learning, AI training engine: how to tell them apart. The overview of vendors by category, as of 10 October 2026, is in Alternatives to the LMS for training with AI: an overview by category.

Why is delivery not enough?

Three limits stem from the category of tool, not its configuration. Personalisation relies on rules defined in advance. Completion attests exposure to content, not ability. Content ages at the pace of its subject.

Research in the learning sciences sheds light on these limits. Forgetting is rapid after a one-off learning episode (replication of Ebbinghaus's forgetting curve by Murre and Dros, 2015). Spacing sequences and self-testing improve retention (Cepeda et al., 2006; Roediger and Karpicke, 2006). A module delivered once, without review or spaced assessment, makes little use of these mechanisms. The reasoning is developed in The LMS is no longer enough: what an AI training system adds.

What are completion rates worth?

The only large-scale public data concerns MOOCs: a median of 12.6% across 221 courses (Jordan, 2015), and 3.13% for participants in MIT and Harvard MOOCs in 2017-2018, compared with 46% for those who had paid for a certificate (Reich and Ruipérez-Valiente, 2019). These figures do not transfer to companies, for which no comparable public statistic exists. The evidence-based levers and their limits are presented in E-learning completion rates: what the research really says.

A high rate remains a usage signal. To know whether training has produced a skill, you have to measure it: see Measuring skills acquisition beyond the completion rate.

What can be expected from adaptation and AI?

The strongest evidence concerns intelligent tutoring systems: Kulik and Fletcher's 2016 meta-analysis reports a median effect of 0.66 standard deviations on test scores, weaker on standardised tests. The principle is explained in What adaptive learning is and why it works. These results support the approach, not any particular product. AI does not replace human support or the working conditions that allow people to apply what they have learned: see What AI cannot do in training.

How to choose without mistaking the category?

1. Start from the need: tracing, recommending, producing, adapting or generating.
2. Check what the tool measures: completion or acquisition.
3. Compare pricing models (per registered user, per active user, per tier, on quote) and the three-year total cost: see How much does an AI training system cost a company.
4. Compare vendors on the same grid: see Alternatives to 360Learning: comparing training platforms.
5. Test on your own content and profiles, using the questions in How to choose an AI training system: the right questions to ask.

Replacing the LMS is not always necessary: most components (authoring tool, adaptive learning, AI training system) can connect to it, with the LMS continuing to provide access and traceability.