Demo · step 4 of 5
People type messy names. AI turns them into one truth.
Ask 20 people about the same project and you'll get five spellings. WorkLens uses Claude to merge the variants — so your counts are real, not fragmented. Raw answers are never modified, and every run is audit-logged.
“Payments Replatform”
“payments replatforming”
“Payments replatform QA”
→
Payments Replatform
5 people · counted once
“ERP Migration”
“ERP migration (SAP S/4)”
“ERP migration testing”
→
ERP Migration
4 people · counted once
“Mobile App Rewrite”
“Mobile app rewrite (v2)”
“Mobile rewrite design support”
→
Mobile App Rewrite
4 people · counted once
Why this matters
Without normalization, your biggest program looks like three small ones. Acme's Payments Replatform showed 3 people as raw data — the real number is 5. A 40% undercount on your largest investment is how budget conversations go wrong.
- ✓ One click — runs in seconds
- ✓ Original answers preserved, mapping auditable
- ✓ Understands context: “SAP S/4” = your ERP program