The viral LinkedIn version, Salesforce cut people while IKEA reskilled them, gets shared as proof that AI automation fails and human judgment is the moat. Read the two halves again. Both are stories of AI working.
Why the comparison is junk
Different technology, a decade apart. IKEA's Billie launched in fiscal 2021 as an NLP intent router, before ChatGPT existed, and the reskilling and the €1.3 billion channel figure are from 2022. Salesforce deployed Agentforce at the start of 2025, a different class of system. The story's own punchline, "same technology, different results," is the one claim in it that is flatly untrue.
A misread statistic. The famous 47% is Billie's containment rate (inquiries resolved without a human), not a quality score showing AI beat people at the easy questions. Different number, different argument.
Unsourced load-bearing claims. The "rehiring support at 1.5x" line and the Benioff "we moved too fast" regret quote are the beams holding the whole narrative up, and neither is on the record. What is on the record: support headcount down, costs down 17%, no backfill, people redeployed, and thousands of new sales hires.
Both companies ran the same play
IKEA moved call-center staff into a remote selling channel. Salesforce moved support staff into professional services, sales, and customer success, and hired more sellers. Both took people out of deflectable work and pointed them at revenue. One was announced as reskilling, one as needing fewer heads. The mechanics were nearly identical; the press coverage was not.
The real lesson is not "augment instead of replace." It is that freed capacity becomes a new business line only when there is unserved demand sitting adjacent to the work you automated. IKEA's leftover queue was full of "will this sofa look right in my room," a sales motion nobody had staffed. A B2B support ticket about an integration failure usually is not. For the full breakdown, read the opinion piece.