There is one operation that costs more time every year than anyone imagines: pasting a web address into a course. It looks harmless. Multiplied by the number of courses, intakes and trainers, it produces dead links, activities pointing at the wrong content, and learners landing on a scenario meant for another group.
Deep Linking removes that operation altogether. It is one of the three LTI 1.3 services, and probably the one trainers notice first. The principle: the trainer stays inside the LMS, clicks to add an activity, picks a scenario in a selection window, and the activity is created with the right content already attached. No address to handle.
We will look at exactly what happens at that moment, why it changes the scale a team can work at, and how NRPS completes the picture by turning LMS groups into cohorts on the tool side. If the protocol itself is still hazy, start with LTI 1.3 explained for training teams, then come back here; it all sits within our guide to integrating Face Up with an LMS.
A hand-pasted address is a breaking point. It ages badly: rename a piece of content, reorganise a catalogue, and the link leads nowhere. It copies badly: one character lost at the end, one parameter fragment forgotten, and the learner reaches an error page or, worse, a scenario that is not theirs.
Above all, it duplicates beautifully. A trainer copies last year's course for the new intake. The link comes along. It still points at the old configuration, the one that no longer matches the level being taught. Nobody notices until the first session.
This cost appears in no budget. It is counted in support messages, delayed sessions and lost confidence. A trainer who has been misrouted twice in front of a class stops offering the activity at all. That is the real loss: not an hour of fixing, but a practice that never takes hold.
At the scale of a whole institution, the arithmetic turns harsh. H3 Campus takes in between 3,000 and 4,000 learners a year, work-study students across a broad range of fields — management, sales, HR and property. Each field has its own situations, and therefore its own scenarios. Leaving the routing to copied URLs would be unmanageable by the second intake.
Ask a trainer: they are not asking for an integration, they are asking not to leave their tool. They want to add an activity the way they add an assignment or a quiz, from the same menu, with the same gestures. Anything that forces them to open a second tab, hunt for a reference or ask a colleague for an address will be done once and then dropped.
Deep Linking answers exactly that. It does not make the external tool richer: it makes it available in the right place, inside a gesture the trainer already performs.
The starting point is unremarkable. The trainer opens their course, switches to editing mode and clicks “add an activity or resource”. In the list, Face Up appears like any other activity type, because the tool was registered once at platform level. Nothing has to be configured again at course level.
That detail matters more than it looks: the technical registration is done once by the administrator, and every trainer inherits it. Nobody needs to know an identifier, a key or an address.
The trainer selects Face Up, and a window opens inside the LMS. It shows the catalogue of scenarios available to their institution: handling an unhappy customer, a performance conversation, taking a brief, turning down a request, depending on what has been made available. They browse, they choose.
That window is served by the tool but displayed inside the LMS. The trainer has not logged in anywhere else: the same platform-to-platform trust mechanism authenticates them, carrying their trainer role — which is why they see a catalogue where a learner would go straight into the practice session.
The trainer confirms. The tool returns to the LMS not an address but a full description of the activity: its title, the chosen scenario, and the technical link pointing to it. The LMS then creates the activity in the course with those elements.
The result fits in one line: every course points at the right content, because nobody had to designate that content by hand. The trainer can then adjust the displayed title, the opening date, the position in the section — all LMS settings they already know.
Deep Linking creates a clean, healthy separation. On one side, a catalogue of scenarios maintained in one place. On the other, courses that draw from it. Improve a scenario, adjust its difficulty, and every course referencing it benefits, with no rework inside the LMS.
This is the opposite of a frozen package exported and dropped into each course, where every copy lives its own life and ages separately. We develop the distinction between the two logics in SCORM, LTI or xAPI for conversational training.
A customer relations scenario written for front-line staff dealing with passengers has no business in a management course. As long as routing rests on copied addresses, the mistake stays possible — and it surfaces in front of the learners.
With a selection made from a named, organised catalogue, the question disappears. The trainer sees what they are choosing, in plain words, at the moment they choose it. It is free quality control, exercised by the person best placed to do it.
CCI Hauts-de-France and CCI de Grenoble — two chambers of commerce and industry — work with two audiences: professionals on short courses and students through their apprenticeship training centre. They use conversational practice before the session to prepare, during it to put things into practice, and after it to embed the reflexes.
That sequencing requires three separate activities in the same course, each pointing at a different scenario, or at the same scenario at a different moment. It becomes trivial when each activity takes three clicks, and painful when three addresses have to be tracked down. The protocol does not create the teaching, but it makes it workable.
NRPS, or Names and Role Provisioning Services, is the second LTI 1.3 service training teams benefit from immediately. It allows the tool to ask the LMS for the list of people enrolled on the course, together with their role: learner, trainer, administrator.
The request comes from the tool, the answer comes from the LMS, and the LMS remains the source of truth. That is simply sound data governance: you do not copy a list, you consult it.
On the Face Up side, that list takes the shape of a cohort. The LMS group “Business Management, year one” becomes a tracked group in the tool, with the same people and the same roles. Enrol a work-study student in October and they join the cohort with nobody asking you for anything. Remove someone and they drop out of it.
It also settles permissions. A trainer recognised as a trainer sees their group's results; a learner sees only their own. Roles are not retyped, they are inherited.
In practice, NRPS removes the enrolment file sent before each session, with its misaligned columns, its duplicates, its stale email addresses and the processing delay that comes with them. It also removes the question that poisons large deployments: “who has the right version of the list?”.
At Eurotunnel, around 1,000 learners practise customer relations scenarios from Rise Up — the French LMS, to be distinguished from the authoring tool Articulate Rise 360, whose name regularly causes confusion. HR tracking there covers skills progression on their new approach to customer relations and team engagement, encouraged through gamification and inter-team challenges. A ranking between teams only means something if the make-up of those teams is accurate: that is precisely what NRPS guarantees.
The first decision is not technical. Who assembles the catalogue trainers can draw from? A central instructional team, to guarantee consistency, or each programme lead, to stay close to the profession? Both work. What does not work is failing to decide: a catalogue with no owner becomes unreadable within a semester.
The selection window is only as good as the titles it displays. A scenario name should state the situation, not the abstract skill: “Telling a rushed passenger about a delay” is chosen with confidence, “Difficult communication, level 2” is chosen at random.
Add the intended audience when two scenarios look alike, and keep a stable naming convention over time. It is five minutes of work per scenario, and it decides whether trainers find the right content first time.
Before opening this up to trainers, run a full test on a pilot course: adding an activity, selecting from the catalogue, opening it as a learner, and the result coming back. Check enrolment syncing on a real group, with one trainer and two learners. The step-by-step details are in our dedicated tutorials, integrating Face Up with Moodle and integrating Face Up with Rise Up.
The test takes an hour. It prevents the one situation you cannot recover from: an entire intake discovering a fault on the day of the first session. To frame the place of each tool more broadly, see also integrating customer service training with your LMS and our overview of the training stack.
Deep Linking does not show up in a demo: it is measured by what stops happening. No more copied addresses, no more activities pointing at the wrong scenario, no more trainers giving up because they were caught out twice in front of a class.
Combined with NRPS, it moves the administrative load to the right place: the LMS stays the single source of courses and enrolments, the tool provides the practice, and the join is made through an explicit choice rather than a fragile link.
Which leaves the next question, the one every training manager asks once the activities are in place: what comes back into the gradebook, and how do you read a score given to a people skill? That is the subject of our article on AGS and sending results back.
