LMS, LXP, adaptive learning, AI training engine: how to tell them apart

Definition: what is an LMS?

An LMS (Learning Management System) is software that centralises an organisation's training: catalogue, enrolments, delivery of online modules, tracking of progress and results. It plays content produced elsewhere, usually to the SCORM, xAPI or cmi5 standards. Moodle, distributed as open source under the GNU GPL v3 licence, is the best-known example. Features and pricing are detailed in What is an LMS? Definition, features and pricing.

The LMS: managing and delivering

The LMS centralises content, manages enrolments and obligations, delivers modules and reports on activity: logins, time spent, completion rate. Its logic is administrative: getting the right people to take the right training and keeping a record of it. It is suited to regulatory training, catalogues and session management.

Interoperability standards determine what it can record. SCORM, developed by the ADL Initiative, reports a limited set of data (status, score, time). xAPI, standardised by the IEEE in 2023 (IEEE 9274.1.1-2023), records more varied activities, including outside the LMS, in an LRS (Learning Record Store). cmi5 is the xAPI profile designed for modules launched from an LMS.

The LXP: organising access to content

The LXP (Learning Experience Platform) focuses on the learner's experience. It aggregates content from multiple sources, internal and external, and recommends it based on profile, interests or usage, along the lines of a content platform. Degreed, for example, presents its Degreed Learning offering as an LXP. Its logic is discovery and engagement; it does not, as such, verify that a skill has been acquired.

Adaptive learning: adjusting the path

Adaptive learning is a method rather than a category of tool: adjusting the content, difficulty and order of a learning path to the learner's answers. Area9 Lyceum, a specialist vendor, defines it as a delivery method that automatically adjusts to the needs of each learner. Depending on the implementation, adjustment ranges from conditional paths defined in advance to recalculating the path at every step.

Results depend on that implementation. Kulik and Fletcher's 2016 meta-analysis of 50 evaluations of intelligent tutoring systems found a median effect of 0.66 standard deviations on test scores, much stronger on locally developed tests than on standardised tests. The mechanism is detailed in What adaptive learning is and why it works.

The AI training engine: generating and measuring

The AI training engine starts from the gap between target skills and the skills actually acquired. It generates the path that closes this gap for each person, assesses acquisition along the way and recalculates what comes next based on verified learning. It includes adaptation as a mechanism, but its specific function is producing the path and proving acquisition.

The term is recent and has no standard definition: you need to check, tool by tool, what is actually generated and what is actually measured.

Comparison table

Criterion	LMS	LXP	Adaptive learning	AI training engine
Main function	Manage and deliver	Organise access to content	Adjust the path	Generate the path and measure acquisition
Starting point	A catalogue	The learner's interests	The learner's answers	The measured skills gap
Personalisation	By population, rules defined in advance	By recommendation	By adjustment, depending on implementation	By generation, for each person
What is measured	Activity and completion	Engagement	Performance within the path	Acquisition of target skills
Content	Produced upstream	Aggregated	Produced upstream, sequenced	Generated from the skills framework
Typical use	Mandatory training, catalogues	Learning culture, keeping up to date	Structured, assessable skills	Measured upskilling at scale

Why do the categories blur?

Vendors add features from one category to another: recommendations in an LMS, adaptive paths in an LXP, AI content generation almost everywhere. Several platforms now present themselves as combining LMS, LXP and authoring tool (see the overview of alternatives to the LMS). The useful test is about function: does the tool deliver existing content, or produce a path from a measured gap? Does it verify completion, or acquisition?

How to choose?

The choice follows from what the organisation needs to produce. To trace mandatory training, an LMS is enough. To develop a culture of self-directed learning, an LXP is relevant. To upskill hundreds of people on specific objectives and demonstrate it, you need to generate and measure. The questions to ask a vendor are listed in How to choose an AI training system, and the limits of delivery alone are detailed in The LMS is no longer enough.