Scroll down

How companies have got to grips with Adaptive Learning technologies

Evan Friburg |

Adaptive learning technologies have been around for a while now. With a first appearance at the end of the 2000s, higher education was the first to benefit from it with very successful use cases. For several years, companies have also got to grips with adaptive learning to individualize learning with a skills-based approach. Let’s take a look back at the evolution of this key technology.

The origins of adaptive learning

Historically, adaptive learning takes us back to an educational method which was already used before the digital era: one-to-one tutoring. Like many psychologists of the 20th century, Bloom[1] promoted the merits of personalized face-to-face learning in his studies as a more effective method than traditional teaching. However, the “one teacher for each learner” approach cannot be applied on a large scale, in mass educational systems. With the digitalization of learning, differentiated learning has emerged: the learner is “categorized” and guided through a predefined path thanks to an expert system.

Adaptive learning as it exists today was born at the end of the 2000s. Compared to differentiated learning, it analyzes learning data more thoroughly, using artificial intelligence algorithms in particular, to provide dynamically changing learning paths and improve the efficiency of learning. To provide individualized learning experiences, adaptive learning focuses on three main categories of data to model each learner (or define his or her learner profile):

  • Information related to content: skills graphs, learning resources, metadata
  • Information related to learners: learning objective, level of proficiency, learning preferences, learning history, etc.
  • Learning events: interactions between content and learners on one or more learning platforms

In terms of functionalities, adaptive learning particularly consists in implementing recommendation engines for learners. Two levels of recommendation are considered:

  • The recommendation of skills suggests items (skill, knowledge, concept, etc.) to be prioritized according to the user’s learning objectives and profile.
  • The recommendation of content suggests learning resources (classroom, e-learning, video, quiz, micro-learning, serious game, immersive learning, etc.) according to the knowledge and skills which the learner must acquire.

Early results in higher education

The first applications of adaptive learning are observed in the United States, in higher education institutions. Arizona State University, a pioneer in this field, saw the number of its students who completed a module increase by 18%, while the dropout rate fell by 47%, according to a report by Tyton Partners.[2]

Another study conducted in six public universities showed that students who took a statistics course with adaptive learning achieved the same results as other students in 25% less time (the others took the same course with a teacher). Those results were observed regardless of the students’ social background.

Adaptive learning has proven its worth in companies

For several years, adaptive learning has penetrated the corporate training market, where HR managers have perceived the impact of this technology on the acquisition of skills, the engagement of employees in their learning path and the return on investment of learning. What made this development possible is the amount of existing data to analyze:

  • The skills repositories allow to know the skills to be acquired for each employee
  • The course catalogs are structured and the learning material is increasingly multimodal
  • The assessment data (to know each employee’s level of proficiency) has become more and more common: positioning test, self-assessment, assessment by the manager, feedback from peers, etc.

For the company, the challenge is to individualize learning with a skills-based approach, while learning plans are traditionally defined with a one-size-fits-all approach or manually personalized by managers: in both cases, the approach is not optimal. The benefits are as follows:

  • Positioning tests are optimized thanks to adaptive testing, reducing the time spent on assessments by 36%*
  • The time spent on “mandatory” learning is also shorter because adaptive learning allows to go through only 63% of the learning content, compared to a traditional path, to acquire the same skills*

As a conclusion, adaptive learning augments companies’ learning environments and improves its impacts if it respects two things. First of all, it must integrates into the environment (learning platforms and catalogs) to collect the data: fortunately, this integration is made easier by interoperability standards such as SCORM, LTI or xAPI. Besides, despite a tendency to automate processes, humans must remain at the heart of the system. This is why L&D teams, managers and HR departments must have visibility on the adaptive learning tool to monitor what artificial intelligence does.

* These results have been achieved as part of experimental protocols carried out with several clients of Domoscio.


[1] Source: B. S. Bloom, “The 2 Sigma Problem: The Search for Methods of Group Instruction as Effective as One-to-One Tutoring”, Educational Researcher, pp. 4-16, Jun. – Jul. 1984

[2] Source: “Learning To Adapt 2.0: The Evolution Of Adaptive Learning In Higher Education”, Tyton Partners, 2016