AI deployed to replace employees and cut costs - instead of as a tool to assist them - is a pattern we've seen before.
IVR systems (interactive voice response) were supposed to route customers to the right agents faster. Instead, some companies turned them into maze-like menus designed to exhaust callers into giving up. The result: customers learning to hate IVR.
We're watching the same playbook with AI in customer service. A company once known for strong post-sales and logistics is now generating widespread complaints because:
😡 Returning a product went from 3 clicks to a 1+ hour chatbot session
😬 AI has autonomous decision-making authority in disputes with no human override
😫 Sellers lose money when AI sides with buyers absent any evidence - photo proof gets ignored
😳 At least one seller's account was suspended by what appears to be an AI-triggered process, with no recourse
🤬 Human agents absorb NPS scores (Net Promoter Score) for interactions the AI handled - including neutral scores, which count as negative
The supervision relationship got inverted: humans now accountable for AI errors they cannot correct.
Customers are filing complaints directly with regulators (consumidor.gov.br) and review platforms (ReclameAqui) to bypass the AI entirely. Others are moving to competitors (Shopee).
The core issue is not AI itself. It is deploying AI with decision-making authority before it has earned it, and removing the human capacity to catch and correct its errors.
The goal is to reduce support costs, but destroying customer trust will end up as an expensive way to do it.
