How we think
Why we start with friction.
Because the problem usually isn't people.
The problem
The work that gets in the way of the work.
Every organization has smart people.
Every organization also has spreadsheets no one fully understands, documents no one can find, meetings no one remembers, approvals no one questions, and knowledge that walks out the door when someone leaves.
You've heard the symptoms. You've probably said them.
- "I know the answer exists somewhere."
- "I've explained this five times."
- "I've been staring at this email for an hour."
Those aren't intelligence problems. They aren't effort problems.
They're friction problems.
What we mean by friction
Friction is the unnecessary effort between people and meaningful work.
Searching for things that should be findable. Repeating things that should be remembered. Waiting on things that should be moving. Re-entering data that already exists. Rebuilding something someone already built.
None of it produces anything new. All of it compounds, every day.
The part nobody talks about
Here's the strangest thing we've learned about friction: it rarely announces itself. It feels personal.
We assume we're disorganized when information is scattered. We assume we're forgetful when knowledge was never captured. We assume we're overwhelmed when the work demands constant context switching. Capable people quietly conclude they're the problem, when what they're really doing is pushing a cart with a locked wheel.
No amount of motivation fixes the wheel. You fix the friction.
Good systems don't ask capable people to spend their best energy fighting friction. They let them spend it on work that matters.
Our approach
Good systems remove friction.
Across every kind of work we've touched, we keep seeing the same pattern: collaboration doesn't happen by accident.
The way information moves, the way decisions get made, the way knowledge gets kept or lost. All of that is a system, whether anyone designed it or not. Undesigned systems accumulate friction. Designed ones shed it.
So that's what we do. We look at how people, knowledge, and tools actually work together, find where the unnecessary effort lives, and redesign around it.
In practice, that might mean cleaning up an intake process, turning scattered knowledge into something searchable, connecting tools that don't talk to each other, or giving a team a clearer path from request to decision to done.
Sometimes the fix is AI. Sometimes it's a better document. Sometimes it's a process change, an automation, or deleting a standing meeting nobody needed. The best solution is rarely the one with the most technology. It's the one that leaves people with the least friction.
We don't care which tool wins. We care about removing unnecessary friction.
The tools
Where AI fits.
AI gives us new ways to build systems around people, instead of forcing people to work around systems. Used well, it's scaffolding: temporary support that helps people do more than they could alone.
For one person, that looks like organizing a messy idea, drafting a first version, breaking an overwhelming project into steps, or learning an unfamiliar subject quickly. Not because people aren't capable. Because working memory is finite.
For a team, it looks like meetings that produce usable notes, context that doesn't get lost between handoffs, and fewer conversations that start with "wait, what did we decide?"
For an organization, it looks like documentation people can actually search, recurring problems that get noticed instead of endured, and knowledge that outlives any single person's tenure.
Notice what's missing from that list: AI making the decisions. The scaffolding holds the weight. People still choose the direction. Good scaffolding eventually comes down, and the building stands on its own.
The direction
Technology should adapt to people.
Technology has spent decades asking people to learn its language. Learn the workflow. Memorize the process. Use the right template. Find the right folder.
We're finally reaching a point where technology can learn ours.
Instead of asking people to think more like software, we can ask software to support how people already think.
Before we build
Our test.
Before we build anything, it has to pass three questions.
Does it reduce unnecessary friction?
If not, we don't build it.
Does it preserve human agency?
If not, we redesign it.
Does it genuinely increase capability?
If not, why are we doing it?
A solution has to clear all three. Removing friction by removing people fails the test. Adding capability that creates dependency fails the test. This is how we keep ourselves honest.
Our commitments
What we won't build.
- We won't automate decisions that need human judgment.
- We won't replace expertise with confident guesses.
- We won't recommend technology because it's fashionable.
- We won't create dependency where we could create capability.
- We won't remove people from work that's better with people in it.
If a project requires any of those, we'll tell you, and we'll tell you what we'd do instead.
Start with the friction
We believe most people are more capable than the systems around them allow them to be.
That's why we don't start with AI. We start with friction.
When people spend less time fighting their tools, they have more time to solve problems, care for customers, teach one another, and create things that didn't exist before. We can't eliminate work, and we wouldn't want to. We want to eliminate the work that gets in the way of the work.
So tell us where it lives. The spreadsheet ritual. The search that never finds anything. The handoff that loses context. The explanation you've given five times.
The places that feel hardest are often where people rediscover what they're actually good at, once someone's curious enough to look.