When Buttons Beat Screens: What Cars Can Teach Us About UX Design

For nearly a decade, the automotive industry has been conducting a massive real-world experiment in interface design, one that’s now revealing lessons for anyone building digital products.

Car manufacturers rushed into a touchscreen-first future, eliminating physical buttons in favor of sleek displays. It looked modern, felt innovative, and promised unprecedented flexibility.

But there was a problem. The experiment was failing, and the consequences were not theoretical. They were measured in crashes, injuries, and lives.

Today, the industry is reversing course, bringing back physical buttons for critical controls. And there’s a message here for every designer, product manager, and business leader building digital experiences, especially in high-variability contexts where attention is constantly fragmented, including markets like India, but by no means limited to them.

 

The Touchscreen Revolution That Went Too Far

For nearly a decade, car manufacturers rushed headlong into a screen-first future. Tesla led the charge, and legacy automakers followed with increasing urgency. Physical buttons disappeared like an endangered species. Knobs, switches, and tactile controls were deemed relics of an analog past. The dashboard became a canvas for ever-larger displays, with some models stretching across much of the vehicle’s width.

From a business perspective, the reasoning seemed sound. Touchscreens offered compelling advantages. They reduced manufacturing complexity by consolidating dozens of physical components into a single interface. They enabled over-the-air updates that could add features post-purchase. And they projected a modern, tech-forward brand image that resonated, particularly in the EV era.

The problem was not the screen itself. The problem was the setting.

Humans still had to use these interfaces while traveling at highway speeds.

 

When Good Design Theory Meets the Real World

Here’s where things get interesting for us as designers. On paper, touchscreens offer virtually unlimited flexibility. You can reorganize menus, add new features, and customize experiences in ways physical buttons never could. From a purely digital UX perspective, they often appear “better.”

But context matters, and context is where good design lives or dies.

In a moving vehicle, the environment is already working against precision. Roads are imperfect. Weather changes visibility. Traffic patterns are unpredictable. Vulnerable road users appear from blind spots. Even a small bump can turn a deliberate tap into an accidental one.

In some markets, like India, these conditions are amplified by density and variability. But the underlying reality is global. Any place with heavy traffic, rough surfaces, aggressive lane changes, snow glare, rain, heat, or unfamiliar roads creates the same usability penalty.

Research commissioned by IAM RoadSmart and the UK’s Transport Research Laboratory in 2020 revealed something startling. Drivers using touchscreen controls while driving experienced reaction times slowed by up to 57% compared to undistracted driving. To put that in perspective, that’s worse than driving at the legal blood-alcohol limit, which slowed reactions by about 12%, or driving under the influence of cannabis, which caused roughly a 21% delay.

Think about what that means in any high-demand driving situation where split-second decisions matter. At 80 km/h, a driver’s eyes leaving the road for 16 seconds while interacting with touchscreen menus translates to more than 350 meters traveled with reduced attention. That’s roughly the length of three football fields, and for cricket fans, about three pitches end to end.

Now imagine that same distraction in dense city traffic, on a rain-slick highway, or on unfamiliar roads where a pedestrian, cyclist, or vehicle can appear from anywhere.

This is not an edge case. It is a common reality in driving contexts around the world.

 

The Muscle Memory Problem

In 2022, Swedish automotive magazine Vi Bilägare ran a deceptively simple test. They compared twelve modern cars against a 2005 Volvo V70, a vehicle with dedicated physical buttons for every major function.

The results were startling:

  • Physical buttons: Tasks completed in about 10 seconds
  • Worst touchscreen-only car: About 45 seconds for the same tasks

That is more than four times slower.

Why?

Because physical controls leverage muscle memory.

With buttons and knobs, drivers build spatial and tactile awareness. Your hand knows where the volume knob is. You recognize the resistance of a dial. You feel the click without looking.

Touchscreens erase this embodied knowledge.

Every interaction demands visual confirmation:

  • Find the control
  • Align your finger
  • Verify the action registered

There is no tactile boundary. No haptic certainty. No forgiveness for bumps or movement.

As designers, we often forget this: the best interface is not the most advanced. It is the one that works with human physiology, not against it.

 

The Return of Mechanical Door Handles: Designing for Stress, Not Aesthetics

The comeback of mechanical door handles in modern vehicles reinforces the same lesson.

From a UX and human-factors perspective, the return of mechanical handles is a textbook case of designing for stress, failure, and zero learning time. Electronic or flush handles optimize aesthetics and aerodynamics. But they violate core usability principles for safety-critical interactions: discoverability, affordance, and reliability under degraded conditions.

In crashes, fires, or power loss, software-dependent handles can fail entirely. Hidden manual releases often require recall and fine motor control, both of which degrade sharply under panic, smoke, injury, and low visibility.

Human-factors research consistently shows that in emergencies, people revert to instinctive, force-based actions. They do not explore. They do not problem-solve. They pull harder.

Designs that align with this behavior, visible, mechanical, one-step actions, dramatically reduce escape time and error rates. Investigations in the US have linked electronic handles to fatal delays during post-crash egress.

Regulators are responding in ways that mirror the earlier backlash against touchscreen-only controls. Proposed measures in the US and draft safety requirements in China call for power-independent, clearly labeled mechanical releases that work for occupants and first responders, regardless of vehicle familiarity.

This is not anti-technology. It is redundancy engineering grounded in safety science.

There is also a security dimension. As door latches become cyber-physical systems, they introduce new failure and attack surfaces. From a systems-UX standpoint, relying on a single electronic pathway for a life-critical function is brittle. Mechanical handles provide an inspectable, non-hackable fallback, strengthening both safety and resilience.

The design lesson is clear:

Any action users must perform under stress, without training, and with severe consequences for failure must have a simple, obvious, mechanical-feeling path.

Mechanical handles are not a regression. They are a mature acknowledgment that good UX prioritizes human limits, not ideal conditions.

 

When Regulators Step In

The touchscreen experiment might have continued longer if not for an uncomfortable trend. Distraction-related crashes were rising.

Euro NCAP reported a nearly 20% increase in distraction-linked accidents since 2020.

Their response was unusually direct.

Starting January 2026, vehicles must include physical controls for five critical functions to earn a five-star safety rating:

  • Turn indicators
  • Horn
  • Hazard lights
  • Windshield wipers
  • Emergency call systems

Matthew Avery, Euro NCAP’s Director of Strategic Development, put it bluntly. Over-reliance on touchscreens forces drivers to take their eyes off the road, and that increases risk.

In the digital product world, we rarely see regulators intervene on usability. In automotive UX, safety made it unavoidable.

The Great Button Comeback

What followed was telling.

Major automakers did not just pivot. They admitted mistakes.

  • Volkswagen restored physical controls after backlash against its touchscreen-heavy EVs, calling buttons essential for the five most important functions
  • Mercedes-Benz abandoned haptic steering wheel controls in favor of traditional buttons after widespread frustration
  • Hyundai reintroduced physical climate and audio controls in newer models
  • Porsche brought back tactile climate controls in the Cayenne after customer feedback

These are not budget brands cutting corners. These are premium manufacturers with elite design teams acknowledging a flawed assumption.

 

What This Means for Digital Product Design

If you design apps, platforms, or digital systems, this story is not peripheral. It is directly relevant.

1. Context Determines the Best Interface

An interface that works in calm, focused conditions can fail disastrously in chaotic ones.

Ask yourself: What else is my user doing when they interact with my product?

Are they walking through crowded stations? Ordering something while commuting? Managing a task one-handed with a bag in the other? Trying to make a payment with spotty connectivity? The optimal interface has to account for these realities, not just the ideal scenario of an undistracted user in perfect conditions.

Designing for the ideal user is easy. Designing for reality is the real work.

2. Simplicity Comes From Reducing Effort, Not Visuals

The auto industry assumed fewer visible controls meant simplicity.

In practice, it increased cognitive load.

A single visible button for a frequent action is often simpler than three hidden menu layers, no matter how clean the interface looks.

3. Muscle Memory Is an Asset

Consistency enables learning. Learning enables speed. Speed enables trust.

Every time you move a frequently used feature, you erase muscle memory users worked hard to build.

Stability is not stagnation. It is usability.

4. Modern Is Not Synonymous With Better

Touchscreens looked futuristic. Minimalist interfaces look elegant.

But aesthetics do not compensate for friction, confusion, or risk.

Trends should earn their place through outcomes, not visual appeal alone.

5. Complaints About Small Things Are Signals

Early complaints about automotive touchscreens were dismissed as resistance to change.

They were not.

They were early warnings.

When users consistently struggle with something you consider minor, it is not noise. It is data.

 

The Hybrid Future

It’s worth noting that physical buttons aren’t returning completely, nor should they. Touchscreens remain excellent for certain tasks: navigation with visual maps, rear-view camera displays, occasional settings adjustments, and complex configuration tasks done while parked.

The solution isn’t to eliminate touchscreens, it’s to use the right interface for each specific task.

This hybrid approach recognizes that different types of interactions have different optimal input methods. Frequent, safety-critical tasks benefit from physical controls. Infrequent, complex tasks can leverage the flexibility of touchscreens. The best interface isn’t purely physical or purely digital, it’s thoughtfully mixed based on actual use patterns and context.

 

A Call for Evidence-Based Design

Perhaps the most important lesson from the automotive industry’s button comeback is this: we need to let evidence override assumptions, even when those assumptions are held by very smart people with good intentions.

The push toward touchscreens wasn’t driven by malicious designers or incompetent engineers. It came from intelligent people making reasonable-sounding business and design arguments. Those arguments just happened to be wrong when tested against real-world usage.

How often in your own work do you validate assumptions with actual behavioral data rather than theoretical benefits? When was the last time you set up a study to measure task completion time for a frequent workflow? Have you observed users in their actual environment, not just in controlled usability labs?

The automotive industry had to learn through customer complaints, negative reviews, regulatory pressure, and ultimately, quantified safety data showing their approach was measurably dangerous. That’s an expensive way to discover you’ve made the wrong design choice.

 

Bringing It Home

After two decades designing experiences, I’ve watched trends rise and fall, each promising transformation.

Some delivered.

Others quietly created new problems.

The automotive industry’s return to physical buttons isn’t a rejection of innovation or progress. It’s a mature recognition that the best design solution depends on context, human capabilities, and real-world constraints, not which technology is newest or looks most impressive in a keynote presentation.

Next time you’re tempted to consolidate multiple functions behind a hamburger menu, hide controls to achieve visual minimalism, or adopt an interaction pattern simply because it’s trendy, remember those drivers navigating chaotic Indian traffic, hunting through touchscreen menus while trying to avoid potholes, two-wheelers, and sudden lane changes.

Sometimes, the button is better than the screen. And recognizing when that’s true? That’s what separates good designers from great ones.

The Next Era of Roads Is Predictive, Not Reactive

Why Safety and Pavements Need a Software Backbone?

Transport authorities are under pressure from two directions at once: rising public expectations for safer roads, and the operational reality of maintaining aging networks with finite budgets. Add climate stress, growing traffic loads, and the need to justify every investment, and the challenge becomes clear.

The constraint is no longer data.

Most road organisations are already collecting plenty: crash records, pavement condition surveys, citizen feedback, GIS layers, inspections, work history, sensor feeds. In many cases, they are collecting more than their teams can reasonably process.

Modern road safety management and road asset management systems are increasingly helping transport teams detect risk earlier, prioritise interventions intelligently, and direct budgets toward decisions that stand up to scrutiny. That speed matters because road safety is still one of the most urgent and persistent public challenges. The World Health Organization estimates around 1.19 million people die on the world’s roads annually.

The real constraint is decision velocity and decision quality, how quickly agencies can translate data into defensible actions, and how consistently those actions deliver measurable outcomes.

This is why software is now the differentiator. Not as a dashboard layer. Not as reporting output. But as the operating system that connects data capture, analysis, performance measurement, prioritisation, and execution into a repeatable decision workflow.

At Experion, our work in transportation software consistently concentrates in two areas that are tightly linked in practice:

While these are often funded and executed as separate initiatives, they are fundamentally solving the same problem: turning complex road data into timely, defensible, high-impact decisions.

 

The Shift: From Fragmented Tools to Decision Platforms

The global shift underway is not simply digitisation. It is the move from fragmented tools and manual reporting towards software platforms that function as decision systems.

Traditional approaches tend to break in familiar ways: crash data sits in one system, pavement condition surveys in another, GIS files in separate workflows, and citizen reports buried in operational queues. Teams spend disproportionate time reconciling data rather than acting on it. Decision-making becomes slow, inconsistent, and difficult to justify, even when the right intent exists.

Modern road organisations are moving toward platforms that unify:

  • structured capture (field, survey, citizen, agency)
  • consistent analysis and performance measurement
  • prioritisation based on evidence, not instinct
  • reporting that holds up under scrutiny
  • governance that makes decisions repeatable and accountable

This is where software stops being an IT initiative and becomes infrastructure strategy.

 

Road Safety: The Problem Is Execution at Scale

Road safety efforts rarely fail at intent. They fail at execution.

Many agencies can collect crash data. Many can generate periodic reports. The gap appears when it’s time to convert evidence into interventions consistently across geographies, and at the speed required to make a measurable difference.

A modern road safety management system is not simply a repository of road collisions. When built correctly, it becomes an operational loop:

  • incident capture that improves completeness and consistency
  • analysis that identifies patterns, severity drivers, and high-risk locations
  • structured countermeasure planning linked to evidence
  • monitoring that evaluates whether interventions worked
  • visibility that aligns multiple agencies and stakeholders

The outcome is not “more reporting.” The outcome is faster, more targeted interventions, backed by governance models that can defend prioritisation through evidence rather than intuition.

In short, road safety becomes scalable only when evidence becomes action, and action becomes repeatable.

 

Road Asset Management: Where Road Decisions Become Financial Decisions

Road asset management is broad, spanning bridges, signage, lighting, drainage, barriers, culverts, and more. Yet in most networks, the centre of gravity remains pavements.

Pavements typically drive:

  • the largest portion of maintenance and rehabilitation spend
  • public satisfaction (ride quality, disruption, perception of service)
  • safety impacts (surface distress, skid risk, work zones)
  • operational disruption and network availability

Because of this, many authorities invest in pavement condition surveys and assessment programmes. Yet the same pattern often repeats: large volumes of pavement data are collected, but decision-making remains slow and reactive.

Pavement datasets, by volume, frequency, and complexity, can create a data-rich, decision-poor environment. The value emerges only when condition data is transformed into:

  • performance indicators that remain consistent over time
  • prioritised maintenance plans tied to service levels
  • optimised work programmes aligned to budgets
  • measurable KPIs that leadership can track and explain publicly

This is exactly where pavement-focused software becomes decisive. Not because it stores pavement data, but because it converts pavement condition into a disciplined operating model.

 

What Actually Differentiates Successful Road Platforms

In our experience, road platforms succeed when they are designed around decision workflows, not around data storage.

Road authorities do not need more static reports. They need systems that help answer the questions they are held accountable for:

  • What and where are my assets?
  • What needs fixing first?
  • Why now?
  • What is the cost and trade-off?
  • What outcome will this decision deliver?
  • How do we prove improvement over time?

Software becomes the mechanism that makes these answers available quickly, defensibly, and consistently.

This is why both safety and pavement management are increasingly converging into a common transformation pattern: from fragmented tools to integrated decision systems, where operational choices and long-term capital planning are driven by the same evidence discipline.

 

Where Experion’s Product Engineering Fits

The differentiator in transport platforms is rarely feature presence. It is whether those capabilities are engineered into a system that road organisations will actually use at scale, across agencies, roles, and regions.

That requires product engineering depth in:

  • building resilient data pipelines (survey, crash, GIS, condition, citizen data, connected and autonomous vehicle data)
  • designing analytics and KPI layers that support performance reporting
  • engineering workflows aligned to real operating models
  • delivering UI/UX that works for both technical and non-technical users
  • ensuring scalability, reliability, and security for national deployment contexts

Our focus is software that turns transport data into operational leverage, making safety programmes more repeatable and pavement decisions more predictable, defensible, and outcome-driven.

 

The Next Era of Roads Is Software-Led

The transport sector’s next step is not more digitisation. It is building systems of decision-making, platforms that connect condition, safety, budgets, and interventions into measurable operating models.

Road safety improves when evidence becomes action. Pavement performance improves when condition becomes strategy. In both cases, software is increasingly the mechanism that makes those outcomes repeatable, scalable, and accountable.

And that is why, going forward, road agencies will not compete on who collects the most data, but on who can make the best decisions, fastest.

AI Innovations in Ecommerce – Part 2

At the recently concluded NRF 2026, one message came through clearly:
The future of ecommerce is agent-driven.

Retail is entering a phase where AI agents don’t just assist shoppers, they orchestrate entire commerce journeys. Speaking at NRF, Sundar Pichai, CEO of Google and Alphabet, highlighted why the retail industry urgently needs a universal, scalable approach to AI-powered commerce, one that understands the nuances of complex ecommerce workflows.

 

Universal Commerce Protocol: A New Foundation for AI in Ecommerce

One of the most significant announcements was the introduction of Universal Commerce Protocol (UCP), an open, platform-agnostic standard designed to create a common language between AI agents and ecommerce services.

Developed in collaboration with partners like Walmart, Target, and Shopify, UCP enables seamless ecommerce experiences from product discovery to checkout, while allowing retailers to retain ownership of their customer relationships.

This is a major step toward interoperable, agent-powered ecommerce solutions at global scale.

Know more at https://developers.googleblog.com/under-the-hood-universal-commerce-protocol-ucp/

 

From Keyword Search to Agentic Retail Experiences

Retail is rapidly shifting from keyword-based search to natural, goal-driven conversations.

Instead of typing queries, customers can now express intent:
“Plan breakfast for five kids”, and AI agents handle the rest.

Powered by agentic AI in retail, these agents understand context, preferences, and constraints to deliver personalized shopping journeys across platforms like Google Gemini, fundamentally changing ecommerce customer experience.

 

Google’s Full-Stack AI Strategy for Retail

Pichai also emphasized Google’s full-stack AI approach, spanning:

  • Custom TPUs and high-performance infrastructure
  • Advanced models like Gemini
  • Developer-friendly AI APIs

This strategy is enabling massive scale. Retailers processed 8.3 trillion tokens using Google AI APIs in 2024, and are now processing over 90 trillion tokens, an 11x year-over-year increase. This growth highlights how deeply AI-driven workflow automation for ecommerce is being adopted across the industry.

 

Drone Delivery Moves Closer to Mainstream

Another notable innovation discussed was the expansion of Wing, Alphabet’s drone delivery service. Through an extended partnership with Walmart, Wing is now scaling to 270 locations across the U.S., enabling ultra-fast, last-mile delivery for millions of customers.

Drone delivery is no longer experimental, it’s becoming a practical component of modern ecommerce solutions.

 

Responsible, Collaborative AI Innovation

Beyond technology, Pichai stressed the importance of being bold and responsible with AI. Google’s approach focuses on embedding safety controls, watermarking, and governance, while ensuring retailers remain the merchant of record and the primary owners of customer trust.

The message was clear: AI should enhance retailer–customer relationships, not replace them.

AI in ecommerce is no longer about incremental upgrades.
It’s about rebuilding commerce around intelligence, intent, and agents.

References:

https://nrf.com/

https://developers.google.com/merchant/ucp

AI Innovations in Ecommerce – Part 1

AI in retail is no longer just about automation. It’s about creating smarter, more personalized, and seamlessly connected shopping experiences, ones that fundamentally redefine how customers discover, decide, and buy.

As Doug Herrington, CEO of Worldwide Amazon Stores, aptly puts it:

“AI is becoming transformative for our business, and we really haven’t had a technology revolution as large as this since the start of the internet.”

 

From Features to Intelligence-Led Commerce

Ecommerce technology has evolved rapidly. We’ve moved from early innovations like chatbots, AR/VR trials, and Wi-Fi–powered devices such as Amazon Dash buttons, to a far more sophisticated era, one driven by artificial intelligence in ecommerce.

Today’s online stores are powered by AI agentsdrone delivery serviceshyper-personalizationcashier-less retail, and data-driven insights that influence real-time decision-making. The modern ecommerce landscape is no longer defined by isolated features, but by AI-first, intelligence-led experiences.

 

How Retail Leaders Are Using AI in Ecommerce

Top retailers like Walmart and Amazon have already embedded AI assistants directly into their ecommerce websites and apps, Sparky and Rufus, respectively. In October 2025, Walmart further reinforced its AI strategy by partnering with OpenAI, signaling a clear shift toward more intelligent, conversational shopping experiences.

This evolution isn’t limited to retail giants alone.

AI is now a core layer across popular COTS ecommerce platforms. IBM was an early pioneer, integrating Watson with IBM WebSphere Commerce as early as 2015. Since then, the ecosystem has expanded rapidly with platform-native AI solutions such as:

  • Einstein (Salesforce)
  • Sensei (Adobe)
  • Loomi (Bloomreach)
  • Joule (SAP)

These platforms embed AI deeply into search, personalization, merchandising, ecommerce virtual assistants, and AI-driven workflow automation for ecommerce.

 

The Rise of Agentic AI in Retail

Over the past year, Agentic AI in retail has fundamentally changed how online shopping works, for both customers and retailers.

For shoppers, AI agents can:

  • Compare prices across sites
  • Apply the best available coupons
  • Automatically purchase recurring needs, such as a weekly grocery list

For retailers, AI agents enable:

  • Dynamic pricing based on demand and competition
  • Smart restocking driven by trend analysis
  • Instant resolution of customer support tickets

Customer journeys are also shifting, from keywords to natural conversations.

Instead of searching for “Milk” or “Diapers”, shoppers now express goals like:
“I’m hosting a themed birthday party for my 7-year-old this Saturday. I have a $100 budget. Can you handle decorations and goody bags and deliver by Friday?”

The shift is clear: from browsing lists of products to expecting solutions, recipes, bundles, carts, and instructions, built by AI agents for ecommerce, rather than manually clicking “Add to Cart.”

 

Google’s Open Standard for Agent-Powered Commerce

Google recently announced a community-driven open standard, Universal Control Protocol (UCP), designed to integrate retailer checkouts directly with Google Gemini.

With UCP:

  • AI agents can guide customers across any brand’s ecommerce website
  • The entire ecommerce flow, from product discovery to cart management and checkout, becomes agent-assisted
  • Shopping shifts from navigation-heavy journeys to conversational, intent-led ecommerce experiences

More on UCP, along with key AI innovations in ecommerce unveiled at the recently concluded NRF (National Retail Federation) event, coming up in Part 2.