Personalised learning paths in workplace training: myth or reality in 2026?

What is a personalised training path?

Personalising a path means giving up the single path that is identical for everyone. Adjustments can concern the entry point (exemption from what is already mastered), content and examples, difficulty, the order of sequences, pace and delivery mode (classroom, remote, on the job).

French law allows for it. Article L6313-2 of the Labour Code (Code du travail) defines a training action as "a learning path enabling a professional objective to be reached", which may be delivered wholly or partly remotely or in a work situation. For remote training, Article D6313-3-1 requires appropriate technical and teaching support, information on the activities to be completed and their estimated duration, and assessments throughout or at the end of the programme.

Which levels of personalisation actually exist?

Level	What is adjusted	On what basis
1. Catalogue filtering	The modules offered	Profile, job, stated preferences
2. Conditional paths	The branch followed	Rules set in advance by a designer
3. Continuous adaptation	Difficulty, order, reviews, help	Measurement of learning with every answer
4. Adaptive generation	The content itself (explanations, exercises)	Measurement of learning and competency framework

The first two levels are common in training platforms and remain useful. They cannot, however, offer content that does not exist in the catalogue, or respond to a gap nobody anticipated. Levels 3 and 4 correspond to adaptive learning, explained in What is adaptive learning and How AI personalises and adapts a training path.

Do learners expect personalisation?

Yes, according to the available data. In the survey for LinkedIn's Workplace Learning Report 2024, covering 1,063 learners and 1,636 learning professionals in several countries including France, the third reason given for spending more time learning is training personalised to one's interests and career goals, after career progress and the need to stay up to date. Details are in What learners expect from an AI learning platform in 2026.

What actually works in companies?

What is established. The mechanisms behind adaptation are among the best validated in cognitive psychology. The review by Dunlosky et al. (2013) rates practice testing and distributed practice among the most useful learning techniques. Intelligent tutoring systems achieve a median effect of 0.66 standard deviations across 50 controlled evaluations (Kulik and Fletcher, 2016).

What research qualifies. The same work shows that the effect is small when the system is poorly implemented, and larger on locally developed tests than on standardised tests. In a large-scale deployment, Pane et al. (2014) observed no effect in the first year and a moderate gain in the second. Most studies concern schools and universities; controlled trials in companies remain rare.

What drives transfer to the job. The meta-analysis by Blume et al. (2010), covering 89 studies, links the transfer of training to the job with motivation, cognitive ability and a supportive work environment. These factors weigh more for "open" skills (such as leadership) than for "closed" skills (such as using software). A perfectly adapted path therefore remains ineffective without dedicated time, opportunities to apply and manager involvement.

The specific risk of generative AI. In the trial by Bastani et al. (PNAS, 2025), unrestricted access to GPT-4 during practice improved practice scores but reduced results on the exam taken without AI by 17%. Personalisation that does the work for the learner harms learning.

How can you recognise genuine personalisation?

Four questions separate a real mechanism from a façade:

1. Is the path recalculated from the learner's answers, or only filtered by profile?
2. Is learning measured throughout the path and after a delay, or by a single end-of-module quiz?
3. Does the system rely on a structured competency framework with explicit prerequisites?
4. Do assessments require applying a skill, or only recognising a correct answer?

A system that answers yes to all four acts on learning. A system that only adjusts display or pace is a sales argument.

What technology does not change

Personalisation does not solve a lack of time, a lack of perceived relevance, or a work environment offering no chance to apply what was learned. It needs an active role from managers, practical exercises and regular feedback. For training providers and apprentice training centres in France, adapting learning paths is also a requirement of the national quality certification, covered in Individualised learning paths in training providers and apprentice training centres.