Why do users keep working the old way weeks after Microsoft rolls out new Copilot capabilities?

The Direct Answer

Because a capability being available is not the same as users knowing it exists, trusting it, and rebuilding a habit around it. Without prompting and reinforcement at the moment of work, people default to the familiar workflow that already gets the job done. Awareness and habit, not access, are the real bottlenecks to Copilot adoption.

Deeper Explanation

Adoption stalls because turning on a feature does not change behavior. When Microsoft rolls out a new Copilot capability, it appears silently for most users, who are mid-task, under deadline, and reaching for the workflow they already trust. A new pane or prompt they never noticed does not compete with an ingrained habit, so they keep doing what works. This is why organizational change is so hard to sustain: McKinsey’s research finds that fewer than one-third of transformations succeed at both improving performance and sustaining it, largely because new ways of working are not reinforced until they stick. A rollout email or a one-time demo cannot overcome the gravity of an existing routine.

The second reason is that whatever users did learn fades before the habit forms. Memory research replicated by Murre and Dros in PLOS ONE shows most newly learned procedural detail is lost within days without reinforcement, so a capability introduced once in week one is effectively gone by week three. Microsoft’s own adoption team addresses this by reintroducing features as they mature and continuously nudging usage rather than announcing once. The practical fix is to meet users inside the app: surface the new capability where and when the relevant task happens, prompt them to try it in their real work, and reinforce it until it becomes the default. Behavior changes when the new way is easier to find than the old one, which only happens when guidance and prompts live in the flow of work.

The Research

• McKinsey finds fewer than one-third of organizational transformations sustain their gains, because new behaviors are rarely reinforced long enough to replace old ones.
• Memory research replicated by Murre and Dros (PLOS ONE, 2015) shows most newly learned procedure is forgotten within days absent reinforcement at the point of use.
• Microsoft’s adoption team reports it drives Copilot usage by continuously reintroducing and nudging features rather than relying on a single rollout announcement.

Strategy and Actionable Steps

• Make the new capability visible in context. Use in-app prompts and walkthroughs that appear in the relevant app so users encounter the new Copilot workflow while doing the task it improves.
• Announce inside the app, targeted by role. A capability that helps finance is noise to sales; in-app banners scoped to audience and app get the right prompt to the right people.
• Reinforce, don’t announce once. Schedule repeated nudges over several weeks so the new way is surfaced until a habit forms, matching how memory actually consolidates.
• Anchor the new habit to a real workflow. Introduce the capability inside a task users already do daily so adoption rides an existing routine rather than competing with it.
• Coach the change, don’t just enable it. A coached Copilot adoption program built on real work moves users from occasional use to embedded, repeatable workflows.
• Measure usage, not licenses. Track whether the new capability is actually being used by app and role, so you can re-nudge where adoption lags instead of assuming rollout equals adoption.

FAQ

Isn’t turning on the feature enough if it’s genuinely better?

No. “Better” only matters once users notice it, trust it, and build it into a habit. A superior capability that is never surfaced at the moment of work loses to the familiar routine every time, regardless of its merits.

How long does it take a new Copilot habit to stick?

Longer than a single announcement. Because procedural memory fades within days without reinforcement, durable adoption typically requires repeated, in-context nudges over several weeks until the new way becomes the default reflex.

Why do rollout emails fail to change behavior?

An email arrives outside the app and outside the task, so it is read (if at all) far from the moment the capability would be used. By the time the relevant work happens, the email is forgotten and the old habit takes over.

How do we tell whether users adopted a capability or just saw it announced?

By measuring actual in-app usage by feature, app, and role rather than counting license assignments or email opens. Usage analytics reveal where adoption is real and where a capability was announced but never used.

What actually moves users off the old way?

Surfacing the new way where the work happens, reinforcing it over time, and coaching it against real tasks. Approaches like VisualSP’s in-app Copilot enablement make the new workflow easier to find than the old one, which is what finally shifts the habit.

Does more upfront training solve adoption inertia?

Rarely. Upfront training decays and happens away from the task, so it cannot compete with a daily habit weeks later. Reinforcement in the flow of work, not a bigger training event, is what sustains the change.