Jonathan Archibald looks at the new adaptive technology entering the corporate learning space and discusses the challenges and opportunities that lie ahead.
In the consumer digital space personalisation and adaptive technologies have improved significantly over the last few years, moving beyond a mechanism to make you buy more from Amazon to something which can genuinely enhance all kinds of digital experience. As a result the perception of adaptive personalisation technology has changed from something invasive and externally mandated into something learners, end-users and consumers of all kinds actively want in their online life.
Using myself as an example, Spotify's Discover playlist has become a key part of my listening habits – widening my musical tastes and introducing me to many new artists that I would not have found by myself. Netflix's recommendation engine helps me get the most out of the service, ensuring I see the things which I will like and am interested in. These two examples enhance my experience significantly. It's difficult to imagine navigating the vast amounts of quality content these brands supply if not for the personalised selection and curation services they also offer.
Adaptive technology for learning
As a learning technologist I am interested in how adaptive technology can be applied in the workplace to help personalise learning to a user's needs and experiences, with the aim of improving the effectiveness of time spent learning and ultimately the user's performance at their job.
I am not alone in thinking adaptive technology opens up important new horizons for L&D. In the last few years adaptive tech has been hot property in the US education market, with companies such as Knewton ploughing investment into creating a sophisticated adaptive engine to power better tools for teaching key subjects to children. Knewton are not alone, joined in this space with many others such as Amplify, CogBooks and DreamBox.
We have also seen a few toes dipped into the water of workplace learning; a good example of this is Filtered. Using a mix of up-front and ongoing surveys and assessments, the Filtered system works out how skilled the user is and what they need to learn for their individual professional development. This is then used to 'filter' out the learning that the learner does not need to do, focusing their efforts into the most effective direction and letting them cut through the noise to the learning content and resources which will yield the most benefit to them personally.
All of these systems are impressive and work in completely different ways, but they do have one very important thing in common: they own both the content ecosystem and the tech, which are delivered as one application and cannot be separated. The content is supplied or it has to be authored using their software. It is structured into tiny blocks and precisely tagged so that it fits into the best structure for their respective adaptive engines. The content is also accompanied by question banks that are used to assess the end user's competency in the subject. This is what allows these systems to work so well, but it is also what's stopping adaptive technologies breaking into the corporate learning mainstream.
Just for a minute, consider a typical workplace and the types of training resources on offer. There will be content libraries covering generic training. There will be custom content created by third party development companies (like Brightwave) for specific needs, and there will be content produced by internal L&D teams. This content will be built to conform to a mixture of standards, SCORM, AICC, xAPI etc., or even have no standard at all. And there are the types of social and informal learning that are more difficult to track and measure – from face-to-face training to virtual classrooms, internet search and so on.
So the typical workplace asset mix is a very different proposition from the hand-crafted and prepared content required to power the current learning focused adaptive systems described above.
Generally speaking, adaptive systems are pre-loaded with competencies and attainment levels, and for each competency and level there are small chunks of content and accompanying targeted assessment questions. The system then uses the questions to determine the user's level against the required competency framework and delivers the matching content to improve the learner's individual level. More questions are then delivered to see if the learning content was effective and so on. The more precise and binary the competency and the content chunks, the better for the tech.
On the one hand you have great tech that’s improving every day but only works with certain subjects and content that is hand-crafted for it. On the other, you have businesses with real problems and a mountain of existing quality content that is not viable to remake just to fit the adaptive tech.
To solve this problem we are going to have to step back and look at why we think that adaptive technology is needed in the first place. Is it just because it's the latest shiny tech trend? Or is it because we want a better way to teach us all how to do maths or use technical software better? In my view, it's needed because I think it will help improve the performance of employees and the organisation as whole.
Therefore, we need to move beyond static premade content and only measuring success by the learner's ability to answer questions straight after they have been taught something - which is only ever a test of their short term memory.
Instead we need the tech to be able to use everything at its disposal (elearning, social, and internet resources, virtual classrooms, etc.) It needs to blend this with human interaction and intervention. Success needs to be measured over time from real world performance, and the tech needs to learn which content and interactions lead to improve business outcomes, and which are irrelevant and unnecessary.
If anyone is using adaptive tech in the corporate space or exploring any of these issues it would be great to hear from you.
Follow Jonathan on Twitter.