Every team should shape its own software.
Malleable started from one belief: software can now fit how each team works.
Why we started
Eric Pelz spent ten years at Asana, where he led its workflow group. He worked on making one tool fit every kind of team, at companies from ten people to hundreds of thousands. Those teams were not asking for a platform to build a solution to their problem. They were asking for a solution to their problem.
So he set out to build it. With Malleable, a team describes how their work should run, in plain language. The software carries it out, shows each person exactly what they need, and suggests improvements from real runs for the team to approve. It fits how your team works, and it changes when you do.
What we kept seeing.
Six things Eric saw again and again in ten years at Asana. Researchers had spotted most of them decades earlier.
1Every team works a little differently.
Software is built once, for an average customer who doesn't exist. Then every team bends around it.
So in Malleable, you start from how your team works, not from a template.
2The real process lives in the workarounds.
The spreadsheet on the side. The reminder in someone's calendar. The rule only one person knows. That's where the work actually happens, and no tool can see it.
Lucy Suchman saw this at Xerox PARC in 1987: people use a plan as a guide, and adapt it to the case in front of them.
So in Malleable, those steps go into the plan, and the AI uses judgment on the case in front of it.
3Know-how lives in people, not in code.
When the person who understood a process leaves, it keeps running, but nobody can safely change it.
Peter Naur made the same point about programming in 1985: the real product isn't the code. It's the team's shared understanding.
So in Malleable, the plan is written down where everyone can read it, and it's the thing that runs.
4The people who do the work can build their tools.
They know the edge cases and feel the pain first. They need tools built around their task.
Bonnie Nardi showed it with spreadsheets in 1993, the most successful tool people ever used to program for themselves.
So in Malleable, you build a workflow by describing how the work goes, so the person who knows the edge cases writes them into the plan.
5Repeatable work needs structure to be done well.
We saw it again and again at Asana. Work that repeats goes well when the steps, the owners and the checks are clear.
Atul Gawande found the same in surgery in 2009: a checklist with three pause points cut deaths almost in half, without replacing anyone's judgment.
So in Malleable, AI gets the same structure. You choose what's fixed and what's left to judgment, and the run stops when a person needs to decide.
6Maintenance is where the time goes.
Rules change, people change, the business changes. A tool that can't keep up becomes the next workaround.
Meir Lehman made it a law of software in 1980: a program used in the real world has to keep changing, or it becomes steadily less useful.
So in Malleable, say what changed. Your team reviews the draft, and it goes live when they publish it.
Flexible doesn't have to mean chaos.
The worry is that everyone ends up working their own way. A shared plan is the opposite: one way, written down, that the whole team changes together.
Your team doesn't need an AI employee. It needs a plan it owns.
Why every company has a process nobody will touch, and what a 1985 essay about programming says about keeping one alive.
The team
We're a small team that has built workflow software together, with a combined three decades at Asana.
Eric Pelz is Malleable's founder and CEO. He bakes (breads), cooks (complicated recipes), and pretends his cat is a dog.
Rahul Bachal spent 10 years at Meta building WhatsApp. He knew Eric at college, and plays guitar in a band.
Aaron Gutierrez worked with Eric at Asana for 9 years, leading software integrations. He is in the backcountry skiing as much as possible.
Alex Davies spent over 10 years building Asana with Eric. He has two kids and enjoys swimming and walking his dog.
Our office is in downtown San Francisco, near Powell station. Come by and talk workflow.
