AI agent index: /llms.txtFull content index for AI agents: /llms-full.txt
Product & Development

Reducing friction at scale

Human-centered design, at scale, is about reducing the gap between what users are trying to do and how the system allows them to do it, not aesthetics but alignment between human intent and system behavior.

— Category
Product & Development
— Reading
5 minutes
01 — Definition

What human-centered design really is

In enterprise environments, human-centered design is often misunderstood as empathy workshops and interviews, but at scale it's about one thing: reducing the gap between what users are trying to do and how the system lets them. When that gap is small, systems feel intuitive and adoption becomes natural, the difference being not aesthetics but alignment between human intent and system behavior.

02 — The problem

The problem it actually solves

Enterprise products don't fail because they lack features; they fail because they're built around systems, not situations, defined by requirements and flows without real-world context.

  • Too many decisions at the wrong time
  • Critical actions buried under secondary options
  • Interfaces that assume knowledge users don't have

So users compensate by guessing, repeating steps, and creating workarounds, which human-centered design fixes by designing around real behavior, not ideal flows.

03 — Why it matters

Why leaders actually care

Faster completion

A workflow of five steps instead of nine raises throughput and helps operations scale.

Fewer errors

Most enterprise errors are user mistakes from bad design; good systems prevent them early.

Less training dependency

Well-designed systems explain themselves and reduce reliance on documentation.

Higher adoption

People adopt tools that make their job easier immediately, not because they're told to.

04 — What defines it

What it looks like in practice

  • Prioritizes what matters now, surfacing only what's relevant in the moment, not everything
  • Obvious actions, users should act without scanning or deciphering
  • Context-aware interface, the same user in a different situation sees different needs
  • Timely information, information appears when needed, not before
  • Errors prevented, the best error message is the one that never appears

It doesn't show up as personas or journey maps in Figma, but as specific decisions about what appears when.

05 — Best practice

Five practical shifts

  1. 01Design for real environments. Assume users are in a hurry, multitasking, or under stress, and design for that.
  2. 02Remove decisions, don't just improve them. Every extra choice slows users down, so cut the choice where you can.
  3. 03Collapse multi-step thinking into single actions. Don't make users plan; let them act.
  4. 04Design for first-time use. If the first interaction fails, adoption never happens.
  5. 05Treat confusion as a design bug. If users hesitate, something is wrong with the system, not with them.
06 — In practice

Human-Centered Design (HCD) in Action: Regional Rapid Transit (RRTS) Commuter App

A regional rapid transit system launched a commuter app for ticket booking, live tracking, station navigation, and service updates. It had strong functionality but low usage, because it was designed like a feature platform, not a commuter tool, with a booking flow that demanded route, station codes, and ticket types upfront and live tracking hidden under secondary navigation.

The approach

Instead of redesigning screens, the team redesigned around situational usage. They reframed the app around three real scenarios, before travel, during travel, and disruption, introduced a context-aware home screen that adapted to time of day, recent routes, and active tickets, simplified booking to a suggested default flow ("confirm > pay"), made live tracking primary and persistent, and turned generic alerts into actionable guidance ("Train delayed. Next available train in 6 mins on Platform 2").

The results

Task completion sped up, booking drop-offs fell, and navigation support queries dropped, the key insight being that they added no features but changed when and how things appear based on real context.

Start here

Some of our best projects started with a two-line email.

Most of our work starts with a conversation. No deck required.

Start yours