Artificial Intelligence in Docebo LMS
The latest launch from Docebo offers another transformative step in learning management system development.
Most traditional LMSs are very good at PUSH. Someone in the HR or Training part of the enterprise decides which learners must complete which courses, and pushes the courses towards them. We’ve all seen the ‘You are due to redo your Manual lifting / FCA compliance / Anti-Bribery / GDPR course’ notifications arrive on our screens.
In an ideal world (and especially in the context of an enterprise that seeks to encourage learning) learners need to be able to PULL learning to themselves – to find courses that interest them or that enable them to develop their skills so that they can take on more complex roles or seek promotion.
This should be easy, surely? You have a library of digital courses and other media assets, all of which are comprised of searchable words?
Hmmm. Well, not as easy as you might surmise.
Imagine that you have a new job in HR, and that you will be writing job adverts and filtering out candidates for the hiring manager to review the shortlist.
You want to do your job really well, but you have heard that some ads can unintentionally put off people from applying, who could do the job perfectly well.
Before you start your new job, you decide to brush up your knowledge from the courses within your organisation’s LMS. There is an absolutely perfect course in there for you, so all you need to do is find it.
You sit down at your keyboard and try to find the search facility. What should you search for? Recruitment? Job adverts? Discrimination?
Most learning management systems can only search for metadata. Metadata is the description entered into an LMS when a new learning resource is added to it. Typically this includes; course name, date of upload, possibly course author or owner, and some metatags – or labels – that can be added to describe the course. Search doesn’t look inside courses, which are like published books – it simply reads the cover and the editor’s description.
If the author of the course has decided to be creative with the title, maybe calling it ‘All you need to know about hiring folk’ then searching for the words that might logically come to your mind will not help.
So, you need to think about all of the possible search terms that the person uploading the course has added to it. The person however, was a bit rushed when they did that, and so the only term they added was ‘Advertising’. To get a match, you need to search for exactly the right word.
If you have a long memory, this is pretty much how Google was in the early days. To find something you had to be very specific in the terms that you searched for. And use Boolean searches with double inverted commas around the core phrases, and AND or OR or NOT in the search to narrow things down. Websites like Ask Jeeves then developed the ability for users to type in a question and would then find the answer. And now, we type in whatever words, questions or phrases that enter our heads – and expect an instant and accurate result from search engines.
This is a search based on natural language processing. The search engine has learned, through successive algorithms, to interpret your query correctly, however you spell or say it.
Not only that, but the search takes into its stride the content of the various websites, documents, published papers, images and everything else found online.
With release 7.5, Docebo has introduced a deep search capability powered by artificial intelligence. The system will analyse courses and learning resources added to it (and some courses already within the system) to understand the content, and to identify key words within it. It will then automatically create tags based on those key words. This will make searching for a course immensely quicker and easier. As the search facility looks inside published courses, it will derive many more keywords, and the words chosen will be objective and accurate.
When learners type in a naturally phrased question, for example, ‘Where can I find information on writing non-discriminatory job adverts?’ the deep search facility will be able to search through all of the key words, on all courses, to find the best match(es) for them.
This will mean that pulling courses to them, for their own personal development or personal interests, will be a breeze.
It’s not only learners who will benefit – those administering the system will no longer have to spend time metatagging courses – this will be automatic.
And – those wonderful courses which have languished in the darker corners of the LMS, unloved and forgotten, may find a new lease of life as they are brought to the attention of searchers.
This isn’t all – if a learner finds a great learning resource – whether this is a video, or a pdf, or a podcast, or a course, they can share it with their peers by using ‘Invite-to-watch’ which will automatically generate a list of other learners who have also found similar content useful.
Artificial intelligence isn’t the only new feature in Docebo 7.5. The release also includes updates to Docebo Learn and Docebo Perform:
- ‘My Team’ allows managers to see the learning activities and skills gaps of their team, at a glance.
- Admins and Power Users (with permission granted) can upload screencasts as a formal or informal learning asset.
- You can now set up peer evaluations, so that users in the same role, organisation branch, or with the same manager can evaluate each other.
- Learners can now have multiple managers, meaning that they can have both a functional and direct manager in the platform, which helps you better structure your hierarchy directly in the platform.
- Superadmins can now use the Best Fit search engine to look for the best person in a company’s organisational chart to fit for a position, a promotion, a project, etc.