The Last Shift

Twelve hundred stations pulsing with conversation in a Bangalore call center. Twelve hundred problems being solved in unison. By 2003, that number was 5,000. Not because IndiCore was good at customer service. Because they were good at being 75% as good for 20% of the price.

The Indian call center industry weaponized VoIP, turning voice into data that doesn't care about distance. Companies saved a few dollars per hour per seat, with acceptable quality. Meanwhile, thousands of workers in Iowa, Indiana, and Idaho received pink slips. Dorothy had 23 years of perfect attendance. It ended with a PowerPoint on "strategic realignment."

The pattern completed its first cycle: expertise extracted from one geography, modeled, compressed, and deployed in another.

Cedar Rapids, Iowa, 2004. Empty cubicles. Outdated motivational posters. Dorothy walking the floor one last time. Somewhere in India, the next shift was starting.

This was the first large-scale demonstration of what would become the Extraction Pattern: the systematic documentation, modeling, and transfer of expertise from expensive humans to cheaper humans. The call center diaspora of 2000-2005 displaced an estimated 250,000 American workers. The playbook was simple: document the scripts, train the replacements, cut the originals.

The same playbook is now being run with AI as the destination rather than a different geography.

Priya, now 34, managed a team of 1,200 in Bangalore. She spent a decade climbing from agent to team lead to operations manager. She has certifications in project management and AI literacy.

AI doesn't need managers. It doesn't need motivation. It doesn't need anything that decade of climbing trained her to provide. The same offshoring that displaced American workers in 2000-2004 created opportunities that were themselves displaced by AI in 2020-2025. The Extraction Pattern doesn't stop. It just changes destinations.

The 1,200 seats still exist. 47 are occupied by humans who handle escalations the AI can't resolve. The rest run at 3 AM with no shift differential, no sick days, and no severance packages.

Terminal. The call center displacement is complete in its current cycle. The remaining human roles — escalation handlers, quality auditors, and edge-case specialists — represent approximately 4% of the original workforce. These roles are themselves being compressed as AI systems improve their ability to handle ambiguous situations.

The prognosis for adjacent sectors — technical support, insurance claims processing, appointment scheduling — is 18-24 months behind the call center timeline. The pattern is identical. The destination is the same.

Read the full pattern analysis for a deeper understanding of the forces driving this assessment.