Alex Levin is the Co-founder & CEO of Regal.ai.
In five years, talking to a human in a contact center will feel as quaint as dialing a rotary phone.
I believe voice AI agents built on large language models (LLMs) that answer immediately and know everything about you will handle the majority of customer interactions. Not clunky chatbots or email autoresponders using natural language processing but smart, empathetic agents built on GenAI that sound (and reason) like humans.
This isn’t a sci-fi fantasy. The shift is already underway and accelerating. In my opinion, companies that cling to human-only support models are setting themselves up for extinction. Here’s why:
1. Consumers will demand instant, voice-first interactions.
No one wants to type a support email and wait four days for a scripted reply. People want to say what they need and get an answer immediately.
Talking is faster. It’s more natural, and it’s already embedded into how we live; phones, cars, homes and even appliances are voice-enabled. In the next decade, I believe voice AI agents will become the default interface for customer interactions, embedded in every device we own. Think ChatGPT voice mode version 100.0 but tailored to your brand, product and customer.
These agents should solve problems in seconds. No hold music. No transfers. No need to repeat yourself about what you told them last time. Once people experience this kind of interaction, there’s no going back.
2. AI agents will talk like humans but work like software.
Voice AI agents have infinite learning potential. They can instantly remember every customer interaction. They can speak any language and understand any accent. They can detect sentiment and respond with the perfect tone. They don’t get flustered, annoyed or distracted.
Human agents need weeks of training yet can still make mistakes. Also, the job can be tough, as answering the same five questions all day can lead to burnout. AI can handle the repetitive interactions so that humans can focus on high-value tasks instead.
3. The cost of human labor can’t compete.
A single contact center agent costs around $40,000 a year in salary, training and overhead, according to ZipRecruiter. Globally, that’s nearly $1 trillion annually for a model that can be inefficient and inconsistent, and it’s becoming increasingly outdated.
Our voice AI agents cost between 10 cents and 20 cents, which is 90% less than the human cost—and the price should continue to go down. More importantly, you’re only charged for actual productive talk time, not training or after-call work.
In a decade, the 60-inch TV that was $10,000 for mediocre picture quality is now $1,000 for unreal picture quality. Just like flat-screen TVs over the last 20 years, the price-performance curve for AI agents is accelerating.
4. Personalization at scale will become the norm.
AI agents will access real-time customer data (purchase history, preferences, prior issues) and offer seamless, hyper-personalized support instantly. While human reps need time to review accounts, AI agents can know exactly who customers are and what they need before they even finish a sentence.
5. Resistance to ethical, regulatory and emotional concerns will fade.
Today, many companies hesitate to go all in on AI due to compliance concerns or fear of replacing humans too quickly. However, the barriers are falling.
Regulations are already allowing AI in most industries, and the technology can stay in compliance. Meanwhile, new AI-adjacent jobs are emerging for human agents who may feel burnt out—allowing them to be redeployed to higher-value, more meaningful work.
Real Blockers Exist
I often get asked some version of the following: “These voice AI agents are amazing. Why isn’t every company using them?” It’s hard to do well, and companies that want to implement agents still have material work to do.
Some companies get bogged down trying to build their own infrastructure. More get bogged down due to their knowledge base and policies not being well-documented, as AI agents can’t use “tribal knowledge.” Most get stuck focusing on a use case that is too small to have a material impact on their business, which limits the willingness of internal teams to support the changeover from human agents.
Conclusion
Even with these challenges, I expect the best customer experience to be built on GenAI agents, and I don’t believe companies can afford to let these stop them for too long.
The big change in voice AI in the last 12 to 18 months is that voice AI agents are performing as well as human agents for the first time. You can now have high-quality, lower-cost voice interactions available to all.
I expect the companies that lean into this transformation today to gain massive operational leverage and unmatched customer satisfaction. The ones that don’t might want to start putting their growth plans on hold—just like they do their customers.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?