Guide
[Guide] How to assess mobile learning tools through Telefónica’s data model and Snap’s AR pilot
Telefónica’s four-stage data model and Snap’s 2023 Inspirit pilot show how schools can evaluate learning signals without mistaking engagement for proof of learning.
Mobile technology can help teachers tailor learning, but collecting more information does not automatically produce better teaching. Telefónica described Big Data adoption in education as accelerating in its May 5, 2023 overview. On October 18, 2023, Snap announced a partnership with edtech company Inspirit, advised by Stanford researchers, to bring augmented reality into American classrooms through a pilot with Stride, Inc.
These examples point to a practical rule for schools: define the learning problem first, then decide what data or AR feature gives a teacher a useful next action.
Begin with the learning target
Telefónica presents adaptive learning as a main use of Big Data in education. According to its overview, data can help institutions understand students’ needs and behaviour, then examine their environment, performance, motivation and requirements to shape a more personalised course of study. The intended result is a tailored study path that adapts training and increases motivation.
Turn that idea into a specific test before choosing an app, camera or data stream. Write down the concept or learning difficulty, what the learner needs to understand and what the teacher should be able to see afterward. A signal that answers none of those questions has no clear role in the lesson.
This also gives schools a narrower alternative to broad monitoring. Start with the least intrusive information that can answer the question. Progress or performance may be enough to show what a student has understood; broader behavioural or mood data needs a separate justification because it changes the scope of collection.
Connect each signal to a teacher’s next action
Telefónica divides data use into four stages: descriptive, diagnostic, predictive and prescriptive. The sequence is useful because it separates observing a problem from explaining it and responding to it.
Descriptive: record what happened, such as a learner’s progress or attendance.
Diagnostic: identify why a difficulty may have occurred.
Predictive: use patterns and trends to establish reliable predictions about the next learning decision.
Prescriptive: apply a response, such as changing the learner’s study path.
A dashboard may report that a learner is struggling without explaining the difficulty or indicating which response to apply. In Telefónica’s model, the useful workflow continues from progress monitoring to diagnosis, prediction and a tailored intervention that a teacher can use in real time.
For each proposed input, write the action beside it. Progress or attendance can describe what happened. Diagnostic analysis may clarify the difficulty. A prediction can guide the next decision, while a changed study path supplies the response. If the chain stops at collection, additional signals have not yet demonstrated their value.
Use extra monitoring only for a defined need
For children aged 0 to 12, Telefónica says applications and tools can track attendance while continuously recording school activity, behaviour, mood and performance. Those signals go well beyond ordinary progress tracking. The relevant question is not whether the technology can collect them, but whether each one supports a defined learning decision.
When the difficulty is still unclear, diagnostic analysis may add value. Continuous behavioural or mood recording should be considered only when the school can explain how it changes a specific intervention, who needs access to the information and when collection is no longer necessary for that intervention.
Telefónica also cites AltSchool, where cameras recorded classrooms from different angles to capture children’s facial expressions, speech, vocabulary and gestures. The company notes that classroom cameras and analysis of children’s behaviour have been controversial among parents.
That example exposes the trade-off. Richer observation may provide more signals for adaptation, but continuous monitoring can increase parental concern. Keep the educational purpose separate from the monitoring capability: the learning target should determine the data requirement, not the other way around.
Assess Snap’s AR pilot by the learning evidence
Snap’s Inspirit announcement describes a different route into mobile learning. The programme combined custom Lenses with a mobile application using Camera Kit and a teaching guide. It was designed for use in class or at home and aimed to help students understand STEM concepts.
In the Stride pilot, 85% of students said that integrating augmented reality helped memory and retention. Nearly half showed a notable increase in engagement. These figures support a limited conclusion: participating students reported a perceived memory benefit, and the pilot recorded a notable engagement increase for a substantial group.
They do not establish that every AR lesson improves learning. Engagement can show that an activity captured attention, but it does not by itself prove that students understood or retained the STEM concept. The teacher still needs a way to connect the activity with progress afterward.
Evaluate the AR activity within the same learning workflow as any other mobile tool. First define the STEM concept it must clarify. Then confirm whether the Lens, mobile application and teaching guide fit the intended classroom or home setting. After the activity, examine the progress or performance information linked to that target. Finally, keep reported help with memory or retention separate from the notable engagement increase, and attribute both measures to the pilot rather than treating them as universal results.
The Lens is therefore part of the lesson, not the outcome. Its value depends on whether it clarifies the target concept and leaves the teacher with useful evidence about progress.
A decision sequence for classroom rollout
Set one learning target. State the concept the activity must clarify and the decision a teacher may need to make afterward.
Choose the narrowest useful input. Decide whether progress, performance or attendance is sufficient before considering broader collection.
Map the response. Identify how descriptive information becomes a diagnosis, prediction or tailored study path. If no action follows, the data has not earned a place in the workflow.
Check the AR setting. Confirm that the custom Lens, Camera Kit mobile application and teaching guide match the intended classroom or home use.
Review the result against the target. Record whether students report help with memory or retention and whether engagement shows a notable increase, without presenting the pilot’s measures as a guarantee for every classroom.
Separate learning tools from surveillance. Treat cameras or continuous recordings as a different decision involving behaviour, mood, access and parental concern.
Verify availability before requiring expansion. Snap said it planned to expand the partnership to 25 STEM experiences in more than 50 schools. That was a plan, not confirmation that the expansion became available, so check the programme’s current status before making it part of a rollout.
This sequence gives schools a practical stopping point. If a tool cannot connect its signals or AR activity to a learning target and a teacher response, keep the scope narrower or postpone adoption. If it can, the next question is whether the evidence measures learning itself or only an intermediate signal such as engagement.
What the evidence supports
Telefónica’s model describes how data analysis can support personalised study paths and faster teacher intervention. Snap’s pilot shows how augmented reality can be delivered through mobile software, custom visual experiences and a teaching guide for classroom or home use.
Neither announcement establishes a universal learning gain. Schools can use the reported pilot figures to frame a limited evaluation, but the decision should rest on the match between the collected signal or AR feature, the learning target and the action available to the teacher. That keeps the potential benefit of mobile learning while limiting collection that cannot improve a specific decision.
Official sources
Official source: telefonica.com
Official source: newsroom.snap.com
Related reading
Guides
Reviews
What you need to know
Reviews