A new wave of AI-powered voice agents is transforming how debts get collected, according to a recent investigation by WIRED, handling tens of millions of calls monthly and forcing a reckoning over compliance, consumer perception, and the future of the collections workforce.
The piece, which went hands-on with several AI collection platforms and spoke with industry professionals, academics, and consumer advocates, offers one of the more detailed public looks at how deeply this technology has already penetrated the space.
The scale WIRED documented is striking. Domu, a collections-focused AI startup founded in 2023, reported 70 million connected calls in March alone. Altur, whose platform powers what it calls a “human-less call center,” processes more than 2.5 million debt-related calls per month for major bank clients in Mexico.
The most sophisticated platforms WIRED profiled go well beyond basic scripting. Domu told the publication its agents adjust tone based on context, shifting from a “friendly, neighborly” reminder mode into a distinct hardship conversation posture depending on what a consumer says. At least one platform adjusts Spanish dialect based on whether a consumer is located in Mexico versus Colombia. Moveo, another vendor featured in the piece, goes further still, building what it describes as “psychographic profiles” on consumers by analyzing transcripts of prior conversations before a call is even placed. Most platforms have escalation protocols that route calls to human agents when a consumer mentions bankruptcy, illness, or a death in the family, though WIRED noted that thresholds vary considerably across vendors.
On compliance, every vendor the publication spoke with claimed their systems are designed to meet FDCPA requirements. But not everyone is convinced that’s sufficient. One credit counselor told WIRED he views the industry as “a minefield of legal trouble,” pointing specifically to the risk of a buggy agent inadvertently disclosing a debt to the wrong person. Another advocacy group quoted in the article is actively pushing for legislation that would make companies directly liable for the behavior of their AI collectors. The group’s legal director told WIRED the core concern is scale: unlike a human collector constrained by an eight-hour shift, a single AI agent can run thousands of simultaneous conversations.
The consumer response picture that emerges from the WIRED
