As pandemic-related budget cuts come to call centers, organizations turn to chatbots like IBM's Watson to filter calls, reducing the need for human operators
Karen Hao / MIT Technology Review : Tweets: @_karenhao , @digi_ad , and @cortnie_cdo Tweets: Karen Hao / @_karenhao : Call centers have long been a frontier of workplace automation, but the pandemic has accelerated their vanishing act. https://www.technologyreview.com/ ... Andrew McStay / @digi_ad : Call-centres, an under-appreciated space for tracking workers with emotional AI and usage autonomous of empathic technologies. Looks like Covid-19 boosting the latter as “intent” of callers is responded to with appropriate scripted answers. https://www.technologyreview.com/ ... Cortnie Abercrombie / @cortnie_cdo : AI + Automation replacing call ctr jobs reminds of folklore of John Henry v steam engine. Humans aren't made to be machines. Machines aren't made to be humans. 1000 calls a days on COVID symptom FAQs, not human. Panicked Mom fearing the worst, human. https://www.technologyreview.com/ ...
Context & Ripple Effects
Call centers were already a frontier of workplace automation before COVID-19: an [[a:952955|FT report in late April found the coronavirus accelerating the shift to chatbots and AI across the industry]]. This piece extends that arc with a concrete mechanism — budget-strapped organizations deploying IBM's Watson to filter inbound calls so fewer reach human operators at all.
The longer tail of that decision shows up in the corpus three years on: a [[a:842344|US AT&T agent working alongside AI-generated transcripts and suggestions openly wonders whether she is training her replacement]], while [[a:836326|Wall Street Journal reporting finds conversational AI splitting agents into those who find it helpful and those left with more complex, intense workloads]]. IBM, meanwhile, was publicly framing its pandemic-era AI push in ethical terms in an interview with Executive VP John Kelly III.
First-order effects
- Organizations facing pandemic budget cuts can route routine calls through Watson-class chatbots, directly reducing headcount demand for human operators at the front of the queue.
- IBM gains a crisis-driven deployment window for Watson in a segment where cost pressure, not capability demos, is doing the selling.
Second-order effects
- Remaining agents inherit the residue: as WSJ-reported deployments show, filtered call centers leave humans handling the harder, more emotionally intense contacts — changing the job rather than only shrinking it.
- Vendors of agent-assist tooling (transcription, suggested solutions) find a second market inside surviving call centers, selling augmentation to the same buyers automating intake.
Third-order effects
- If filtering becomes standard, the call center consolidates around software vendors and away from labor, and monitoring tools — including the emotional-AI worker tracking Andrew McStay flags in his commentary — extend from customers to the diminished human workforce.
- The pattern generalizes beyond telephony: any high-volume, scripted customer-contact function becomes a candidate for the same filter-first architecture, with the human role migrating to exception handling.
The trend: Communication-cost-driven automation is using each crisis as an accelerant, converting call centers from labor-heavy operations into chatbot-filtered systems where humans handle only what the scripts cannot.