01 — Uncorporate Scout
An AI-supported workspace for reviewing thousands of employee ideas
Problem
Every few months, TSC asks employees across Deutsche Telekom where their work gets stuck. One campaign brings in three to four thousand responses, each a paragraph of free text.
Every response needs four judgments: how bad it is, what it’s about, who can fix it, and whether it’s worth fixing. All of it ran in one Excel workbook.
Understanding the problem
I was given one Excel file and a question: what should this be as a product?
No brief beyond the file. No access to the employees who wrote the responses. So the workbook became the research — 1,181 rows, 22 columns, one campaign’s worth of complaints about work.
The work stopped in three places.
Scoring and classification were complete — every row had a pain score and a cluster. The time-loss column was filled to submission 788, then stopped. And five columns — process stage, effort, impact, gate, owner — were empty on every single row.
0 of 1,181 entries were ever assigned an owner.
The spreadsheet could describe a problem in detail. It could not hand it to anyone.
One in five people didn’t know where their own problem belonged.
212 submissions answered “Keine Angabe / Ich weiß nicht” for segment. Asking the submitter to categorise wasn’t working.
Almost everyone rated their problem as the biggest one.
64% said it affected the entire organisation. 38% chose the top of the time scale. 38% said it happened daily.
When most answers sit at the top of the scale, the scale can’t rank anything. What could rank things was repetition: 106 separate people described the same broken app experience. Volume was the honest signal, not self-assessment.
Almost everyone rated their problem as the biggest one.
64% said it affected the entire organisation. 38% chose the top of the time scale. 38% said it happened daily.
When most answers sit at the top of the scale, the scale can’t rank anything. What could rank things was repetition: 106 separate people described the same broken app experience. Volume was the honest signal, not self-assessment.
