Málaga ·--:-- Open to remote roles
Back home

Turned IKEA's compliance tool into a learning platform and increased adoption

Role
Senior Product Designer
Product
Internal tool where store coworkers share improvements
Platform
Desktop app
When
2023–25, via tretton37
  • 181%Quality submissions↑
  • 38,4%Validation time↓
  • 27,6%Adoption↑

See GitHub prototype

01the problem/

Sharing felt like admin.

Store coworkers saw Vetlanda as one more compliance task. Behind it, validation ran in Excel, the UI ignored the IKEA design system and the backend was mid-migration.

02my call/

Foundation before features.

Stakeholders wanted AI and gamification, but there was no clean data to build them on. I put validation first and made the trade-off explicit: a quarter with no new coworker feature.

03validation/

Out of Excel, into the product.

I moved a shared spreadsheet into one workflow for three reviewer roles, inside the existing platform instead of a third-party tool.

−38,4%validation time in three months, about 6,5 working days

04submission quality/

70% of submissions came back.

Mostly for missing fields. Instead of the planned form rebuild, I redesigned the Help Centre: cheaper to ship, and it fixed the cause.

+181%high-quality submissions, year on year

05adoption/

Make it personal.

84% of the time, coworkers only wanted ideas from their own department. So: topic onboarding, a profile, better search, and one question, “Have you implemented this?”

+27,6%implementations across stores in seven months

06ai in the loop/

An AI copilot, built on clean data.

I prototyped it with Claude. It reads an uploaded PDF, flags gaps and guides people before they submit. Year one's validated submissions are its reference data.

07looking back/

What worked, what didn't.

  • what worked/Fixing validation first. Clean data is what made the AI copilot possible.
  • what didn't/Working off the IKEA design system, on a backend mid-migration, slowed every release.
  • what I'd change/Enforce the design system from day one, run stakeholder discovery in month one, and treat the roadmap as a design problem.

ikea in 30 seconds/

Fixed the foundation, then adoption followed.

The product
Vetlanda, IKEA's desktop tool where store coworkers share improvements so other stores can copy them.
The problem
Coworkers saw it as admin. Validation ran in Excel, and there was no clean data for the AI stakeholders wanted.
My call
Foundation before features. I put validation first and was open about the cost: one quarter with no new coworker feature.
What I shipped
An in-product validation workflow, a Help Centre beside the form, a personal feed, and an AI copilot prototype built with Claude.
What it shows
I find the root cause, put the work in the right order and make trade-offs visible to stakeholders.

See GitHub prototypeBook a 30-min call

if you're a hiring manager/

Contact me for the full case study

In 30 minutes I’ll walk you through the research, what didn’t work, and where AI saved us time.

CET · I reply within a day

info@draganadekic.com
or see more of my work/