Nudge —
The Learning Signal

Richard Thaler and Cass Sunstein's "Nudge" (Yale University Press, 2008) introduced the concept of choice architecture — the idea that the way choices are structured influences which choices people make, often more powerfully than explicit instruction or financial incentive. The canonical example is the canteen that puts fruit at eye level and chips at the bottom — nobody is prevented from choosing the chips, but the fruit is more likely to be taken. Small design choices, applied consistently, shift behaviour without coercing it.

Applied to AI adoption, nudge theory suggests that the single most powerful intervention is often not a training programme or a mandate — it is a change in defaults. If the default when composing an email in Outlook is that Copilot is ready to help, more people will try it than if they have to seek it out. If the default action after a Teams meeting is an automated prompt asking whether you want the Copilot summary, more people will use the summary feature. Friction is the enemy of adoption; removing friction is a design choice that requires no budget and no compulsion.

The nudge architecture for AI adoption has three components. The first is default-on tools — ensuring that AI capabilities are in the natural workflow path rather than requiring active discovery. The second is timely micro-prompts — short, contextually relevant nudges delivered at the moment when they are most likely to be acted on. The third is social proof signals — showing people what their peers are doing with AI, which is consistently more persuasive than top-down encouragement. Microsoft Viva Insights is the primary infrastructure for components two and three in the M365 ecosystem.

Microsoft Viva Insights is a personal productivity and wellbeing tool embedded in Microsoft 365. It provides individuals with personalised insights about their work patterns — meeting load, focus time, collaboration breadth, after-hours work — and delivers nudges designed to improve the quality of their working experience. Importantly, Viva Insights data is private to the individual: the aggregate signals available to managers and administrators are anonymised and cannot be used to monitor individuals. This privacy-by-design architecture is essential to the psychological safety conditions described in Issue 03.

The core nudge mechanisms in Viva Insights are: the focus time booking prompt (suggesting users protect time for deep work based on their calendar patterns); the meeting effectiveness nudges (suggesting shorter meetings, clearer agendas, and more focused attendance lists); and the connection prompts (suggesting reaching out to colleagues with whom contact has lapsed). Each of these is small, timely, and non-coercive — the Thaler and Sunstein criteria exactly.

Extended to AI adoption, Viva Insights provides the infrastructure for usage-based nudges. If a user has not used Copilot for email drafting in the past two weeks, a nudge appears in their Viva Insights digest reminding them of the feature and linking to a 90-second how-to. If a user's meeting load has increased significantly, a nudge suggests they try Copilot's meeting preparation features. These nudges are personalised to the individual's actual usage patterns, which makes them relevantly targeted in a way that broadcast communications cannot be.

The Viva Insights dashboard for managers and HR provides aggregate signals that support the adoption programme without compromising individual privacy. It can show whether AI usage in a team correlates with reduced meeting load or increased focus time — the intermediate outcome metrics that bridge adoption data to productivity outcomes. Combined with the qualitative signals from Viva Engage retrospectives (Issue 04), this gives your programme lead a 360-degree picture of where adoption is genuinely embedding versus where it is nominal.

Ethan Mollick argues in "Co-Intelligence" that the organisations that will benefit most from AI are those that build it into the texture of daily work rather than treating it as a separate capability. Viva Insights, at its best, is exactly that: AI literacy woven into the day's rhythm, arriving as help rather than homework.

The full vision for AI learning in the enterprise connects five elements: the O*NET task map for the individual's role, the ESCO competence framework for the skills required to perform those tasks, the four-zone classification of each task, a gap analysis of Zone 2 and Zone 3 tasks where the individual is not yet using AI, and a personalised learning path that addresses those specific gaps with specific, curated content.

This is not as complex as it sounds once the infrastructure is in place. The O*NET task list for an HR Business Partner (O*NET 13-1071.00) is publicly available at onetonline.org. The ESCO competences aligned to that occupation are available at esco.ec.europa.eu. The zone classification is the exercise described in Issue 01. The gap analysis is the difference between which Zone 2 and Zone 3 tasks the individual is currently using AI for and which they are not. The learning path draws from your organisation's Viva Learning catalogue, Microsoft Learn, LinkedIn Learning, and any internal microlearning content.

The learning path generated from this gap analysis is delivered through Viva Learning — Microsoft's learning experience platform embedded in Teams. Content can be curated from LinkedIn Learning, Microsoft Learn, internal content libraries, and third-party providers. The personal learning plan is a filter on that catalogue: not everything available, but what is specifically relevant to this person's Zone 2 and Zone 3 gaps, surfaced at the right moment via a Viva Insights nudge.

The spacing effect — the finding that learning distributed across multiple short sessions is retained better than the same content delivered in a single longer session — is one of the most robust findings in cognitive science. Yet most enterprise AI training is delivered in formats that violate this principle: a 90-minute onboarding session followed by nothing, or a 45-minute e-learning module completed once and never revisited.

Microlearning units — five minutes or fewer, focused on a single actionable skill — work better for AI literacy development for three reasons. First, they can be delivered at the moment of need (the nudge arrives when the relevant task is due) rather than in a scheduled session disconnected from the task. Second, they are low enough in cost that they can be repeated — a user who saw the email drafting tip three weeks ago can receive a slightly different framing of the same skill this week, building depth through repetition. Third, they can be created much faster than traditional e-learning, which means the learning catalogue can keep pace with the changing frontier.

The LinkedIn "2024 Workplace Learning Report" found that learners who used AI-powered coaching were twice as likely to recommend their employer. This is a proxy measure for the quality of the learning experience — employees who feel their organisation is investing in their development in ways that are genuinely useful respond differently to those who feel they are being processed through mandatory training. Microlearning, delivered via Viva Learning at the moment of relevance, is the difference between those two experiences.

The microlearning unit structure that works best for AI skills follows a four-part Hook/Core/Practice/Carry format. The Hook is a 30-second scenario that makes the relevance of the skill immediately obvious. The Core is the specific skill, demonstrated in 90 seconds of screen recording or animation. The Practice is a single, small, low-stakes task — try this prompt on your next email draft. The Carry is a one-line reminder of the key action to take. Total time: under five minutes. Total cognitive load: one thing, done once.

The nudge cadence is the operational heartbeat of a sustained AI adoption programme. Without it, adoption peaks at launch, plateaus after three months, and slowly erodes as people fall back to habitual workflows. With it, adoption deepens continuously — the programme keeps learning in front of people at the right frequency and in the right form.