AI Agents Reshape Customer Control and Financial Access

Zendesk CEO Tom Eggemeier said his personal AI agent contacted service providers 37 times over four days and saved him $3,150 by renegotiating cable, streaming, internet and family mobile plans. He said he hadn’t contacted those companies for three or four years before trying the agent. Now, he plans to have it check with them monthly.

The example points to a shift in how customers reach businesses: personal agents may make repeated negotiations and service requests for them. “What’s changing is the amount of interactions is just going to explode,” Eggemeier told Newsweek. Zendesk sees interaction volume rising among customers with the heaviest AI adoption, he said.

Companies must decide which requests to leave to self-service, which AI can handle and where a person must apply judgment or company policy. Automated service can run around the clock and in languages that businesses may previously have supported only during limited hours.

Customer Contact Moves to AI

Eggemeier said customers may increasingly seek help through personal agents or interfaces such as ChatGPT and Claude, rather than visiting a company’s website or using its support tools. That could mean fewer website visits and less human-to-human contact. Businesses would still need to preserve their information and customer experience when another AI system mediates contact with customers. “I don’t think closed ecosystems are gonna work that well,” he said.

Companies should measure whether automation improves service, not just how much work it handles, Eggemeier said. “If you’re getting higher resolution rates, but lower satisfaction, that is gonna impact your business in the future,” he said. Measures of success vary by interaction: whether an issue gets resolved, money is collected or a sale is made. Chief executives should weigh those outcomes alongside customer satisfaction.

Some situations call for empathy, not the quickest decision. If an AI agent leaves a customer “in a kind of circle loop” without resolving the issue, Eggemeier said, that customer should be able to reach a human being.

Voice Systems and Business Adoption

A separate Newsweek analysis said businesses should prepare for customers who speak rather than type. It cited a June 2026 YouTube video in which Boris Cherny, creator of Claude Code at Anthropic, said roughly half his engineering work happened on his phone. He said he could manage AI agents remotely and direct them by voice while walking in the woods, getting coffee or sitting on a couch.

OpenAI reported that more than 150 million people use ChatGPT through Voice and Dictation each week. Its July 2026 launch of GPT-Live added voice functions that let the system listen and respond at once, handle interruptions and delegate complex work during a conversation. xAI’s Grok Voice 2.0 emphasized speech-to-speech conversations, transcription and tool use. Tesla recently added it to cars for hands-free interaction.

The analysis cited a 2026 study of more than 10,000 U.S. adults published in Scientific Reports. Adjusted for errors, median typing speeds were 37.8 words per minute on computers and 29.4 on smartphones. NCVS put the standard benchmark for general American English conversational speech at about 150 words per minute. Voice has limits, too: it can fail in noisy settings, misunderstand instructions or raise privacy concerns when people speak sensitive information aloud. Typing remains useful for detailed writing, precise editing and tasks involving visual control.

Financial Safeguards and Returns

Charles Schwab Corp. Chief Executive Officer Rick Wurster said the company is actively blocking unauthorized AI agents trying to move clients’ money, including Meta Platforms Inc.’s Muse. The extracted report said bank executives welcome consumers’ use of AI in their financial lives when agents follow the rules.

A Wharton and GBK Collective study reported that 67 percent of enterprise leaders whose departments had fully embedded AI, adopted agents and significantly redesigned workflows reported strongly positive return on investment. The figure was 19 percent among leaders whose departments hadn’t done so.

The Newsweek analysis described Model Context Protocol, or MCP, as a way to connect directly to major AI language models and create voice experiences for websites, inbound calls and outbound calls. It said customers could ask questions, schedule appointments, receive text links, make payments or connect to representatives through spoken instructions. The analysis urged businesses to test voice AI on real customer tasks and measure completion and satisfaction.

One business example involved Aisata Diallo, 24, who developed Tressfecta after finding limited access to braiders and textured-hair specialists while attending Northeastern University in Boston. Its first project, Plait Pilot, is a prototype designed to separate hair from packs before braiding. Diallo said a hairstyle can take eight hours, and the physical toll contributed to her stopping braiding. Tressfecta was still pre-launch; Diallo had bootstrapped it, drawn on grant and accelerator funding, and was preparing to fundraise.