At 7:42am, you do the same thing you’ve done for years: open the weather app, glance at wind and rain, decide whether you’re cycling or taking the train.
It’s muscle memory: two taps, five seconds, done.
Then an update lands.
The Met Office has just rebuilt its app “from the ground up” to make forecasts and warnings “clearer and easier to use”. But overnight, the routine breaks. The data is still there, you just can’t reach it at the speed your brain expects. What used to be automatic becomes conscious: scroll, scan, back, try another screen.
The feeling isn’t “I dislike change.” It’s “I’m suddenly slower.”
You see the same pattern any time a familiar, high-frequency workflow gets redesigned. When the Met Office trialled a refreshed forecast website experience in late 2025, regular users complained that key detail took more scrolling and felt “dumbed down” and some shared ways to switch back to the old layout while they still could.
And here’s the twist: this reaction isn’t really about weather.
It’s about procedural memory: the invisible mental wiring that turns repeated steps into effortless flow.
As users, why are we so precious about these things? Shouldn’t we welcome enhancements?
In theory, yes. In practice, when an interface change breaks an ingrained routine, it forces people to think about something they previously did without thinking. That extra cognitive effort is what users experience as friction.
A good example is XCWeather. By modern UI standards it looks dated, but people who depend on wind and observation detail keep using it because it’s fast, familiar, and information-dense. That’s not brand loyalty, that’s muscle memory choosing the shortest path to the answer.
And it’s not uniquely British. When Australia’s Bureau of Meteorology (BOM) launched a “modern, sleek” redesign in October 2025 - a project that eventually cost A$96.5 million ($62.3m; £48m) - the reaction was fierce. The rollout coincided with extreme weather, and almost immediately social media, emergency services, farmers and commuters were up in arms. Users complained that radar maps were confusing, local town data had disappeared and key observations were buried. The backlash was so strong the BOM reverted radar visuals back to the familiar style within days, issued a public apology, and government ministers initiated enquiries.
The lesson was clear: when a critical public service gets significantly redesigned, unfamiliarity can feel unsafe.
That frustration wasn’t about design. It pointed to something deeper in how our brains work. It’s a reminder of how strongly we rely on familiarity, because it signals safety. It’s part of procedural memory, the mental shortcuts that make familiar tasks effortless.
It’s why you can ride a bike without thinking, or send a contract for approval without pausing. After enough repetition, these workflows become second nature. They free up mental bandwidth for higher-value decisions, not repetitive step-by-step instructions.
An example is when people choose to wear the same outfits every day - think Steve Jobs and Barack Obama - to reduce daily decisions, leaving more cognitive space for more meaningful work.
For lawyers, procedural memory isn’t only convenient. It’s a safeguard.
With ingrained workflows, built on known procedures and tools, risk is reduced and compliance kept tight. Disrupt that rhythm, and what was automatic becomes a conscious effort, taking up space in the already limited bandwidth we have. There’s added friction, and friction costs time.
Change can be good, but adoption should respect familiarity. When the way we work stays recognisable, the cognitive load stays low, and teams get value faster.
AI adoption is becoming increasingly widespread. A McKinsey Global Survey found 62% of respondents said their organisations are at least experimenting with AI agents.
Most organisations already live in Microsoft 365: Outlook, Word, Teams, and SharePoint. That itself is an advantage for M365 Copilot, as it brings AI assistance right into the tools your teams use every day, so you can retain the muscle memory.
Dovetail agents build on that principle. As Copilot extensions, they run natively in your Microsoft environment, governed by your existing policies and controls, so adoption feels like an evolution, not a reset. Data stays in your tenant. There’s no new system to stand up and no external data hosting to introduce.
Familiar tools and minimal training: AI inside Microsoft 365 surfaces in Teams and Outlook and works with SharePoint/OneDrive content. Users start quickly because the interface feels second nature.
Governance stays where it belongs: Authentication, audit trails, retention, sensitivity labels, and access permissions remain under your Microsoft 365 control plane, simplifying support and compliance.
No data leaves your tenant: Processing stays within your M365/Azure footprint, avoiding external hosting or new platforms.
Reduce risk while accelerating adoption: By keeping AI inside familiar tools and under existing governance, you minimise operational and regulatory risk from day one while speeding up adoption.
When AI lives inside Outlook and Teams, the shape of the work stays familiar. Staff don’t have to learn new systems or break habits. It avoids the friction that often comes with change.
AI agents make routine reviews consistent. Every contract, policy, or RFP gets the same criteria applied in the same way, with audit trails that satisfy compliance and clients.
Agents handle the repetitive. Lawyers focus on advice, negotiation, and risk calls, the places where human judgement is essential.
The weather-app backlash wasn’t about the forecast. It was about the break in familiarity.
In legal work, we can avoid that break. When AI adoption respects procedural memory, organisations win twice: lawyers keep their rhythm, and the organisation gains efficiency without upheaval.
Familiar doesn’t mean static. It means building on what already works.
To see what AI agents inside your existing M365 workflow look like, book a session below.