← HomeAsana · Trust systems

Unblocked sensitive work at 300,000-seat scale

I led the design of field-level permissions that reached 99% adoption among eligible enterprise organizations within three months. The system let customers control sensitive work and was designed to preserve existing behavior across millions of tasks, rules, and AI-powered workflows.

99%Eligible enterprise adoption

TeamDesign lead · cross-functional team of 6

When2024

Problem: Enterprise deals were stalling because task access exposed every field, so customers kept budgets, salaries, and strategy out of Asana.
My contribution: Led customer research, the permission model, field-level visibility UX, executive alignment, and rollout with a six-person cross-functional team.
Key decision: Put visibility controls in the field library and traded launch completeness for a mental model customers understood without training.
Downstream effect: Customers could bring sensitive work into Asana; the same model now governs what AI teammates and automation rules can access.

This case study is private

Some details are under NDA. Enter the password or email me for access.

A permissions gap was costing deals

Asana’s task model made collaboration easy, but it treated every field the same. Share a task and its budgets, salaries, and strategic metrics came with it. Enterprise prospects chose competitors with finer controls; existing customers kept their most valuable work outside Asana.

A table of work with the Budget column visible to all team members — Budget values, visible to anyone on the project — the gap that was costing deals.

I led the work end to end: customer research, usage analysis across every enterprise account, the permission model, and the field-level visibility experience. The system had to work for teams of ten and 300,000 — and across custom fields, rules, and AI teammates.

My work on access controls affects all of Asana's features — Access controls touch nearly everything in Asana's feature surface.

Field-library controls beat task-by-task permissions

Custom fields carry the business-critical data, but Asana's whole model is built on sharing tasks across projects. Without finer-grained permissions, a field couldn't be both private and collaborative — it had to pick one.

Four constraints became our decision framework: scale across thousands of fields, compatibility with rules and AI teammates, zero regressions across existing work, and trust — no permission could change without making the consequence clear.

A rule template that says 'When Trigger, check if Condition, then do Action' — Flexibility, tested: what happens to a rule when the field its condition depends on disappears from view?

I explored three directions: a low-friction reveal, a full-screen surface with room to explain tradeoffs, and a split view that exposed every field at once. Each made a different bet on speed, context, and the risk of an accidental change.

Direction A — reveal-on-click, to discourage flipping the setting by accident.

A full screen modal with multiple tabs: Overview, field definition, Usage dashboard, and Advanced settings — Direction B — a full-screen surface: more explanation, more friction.

A large modal with a split view showing all available fields and the selected field detail — Direction C — a split view: everything visible, more context per field.

Across 20 customer and internal usability sessions, three signals held: plain language beat permission jargon, an eye communicated visibility better than a padlock, and admins expected to manage access from the field library — not task by task.

Security, Platform, AI, and Product leaders used the same four constraints to evaluate the finalists. I turned their competing requirements into a decision framework and recommended clarity over launch completeness.

We shipped the low-friction reveal in the field library: admins opened visibility controls only when needed, saw the consequence before confirming, and then returned to work. Compared with the full-screen and split-view concepts, it gave up some feature breadth but preserved the clearest mental model.

Decision matrix plotting design complexity against user clarity — The framework we used to choose: clarity over completeness.

The shipped flow — restrict a field from the column menu, confirm once, done.

99% adoption — and a foundation for AI

Metric	Goal	Beta	3 months post-launch
Adoption rate	30%+	40%	99%
CSAT	80%+	80%	90%
Unblocked enterprise use cases	✓	✓	✓
Scalable permissions foundation	✓	partial	✓

99% adoption within three months — on a foundation that now underwrites Asana's AI teammates and automation rules too.

Simple defaults beat more explanation

What worked

Simple defaults did more work than explanation. Iterating with customers and internal teams produced a permission model people could use without training.

Opportunities

I onboarded teams one by one. Documentation and a cross-team workshop up front would have gotten us to full adoption faster.