AI Agents - Today the Phone, Tomorrow the Web

Telephone agents: changing communication with customers

The rise of AI phone agents is rapidly transforming the way businesses handle calls. These are intelligent voice assistants that answer incoming calls or make outgoing calls with natural, human-like language. They work around the clock, speak many languages and never get tired. Companies are adopting AI agents for their virtual call centers not only for the obvious cost and scalability benefits: as indicated in our business plan as early as 2021, the de-humanization of the relationship between companies and customers requires the introduction of AI agents capable of managing customer relationships in a natural way. We are not talking about replacing flesh-and-blood agents: those, especially if they are internal to the company and trained, will always be indispensable. But the volume of inquiries cannot, as of long ago, be met solely by them. If 50 years ago it was normal to talk directly to one's butcher, today it is clearly not possible to communicate with an e-commerce representative; tomorrow (or today, for companies using MrCall or similar services) we will be able to interact with intelligent AI agents.

Leading startups are rapidly emerging in this ecosystem. MrCall offers telephone agents who sound almost human, able to converse in any language and with rates as low as a few cents per minute of conversation. MrCall agents can be "hired" to handle sales calls, schedule appointments, and support customers, operating like a perfect employee who never sleeps. Other players, such as Bland AI and Retell AI, also offer low-code platforms to enable companies to create customized voice agents who answer calls and perform tasks such as booking appointments. All of these companies use advanced language models (LLMs) as the "brains" of the agent, coupled with advanced speech synthesis technologies, enabling fluent understanding of natural language.

These AI agents are not simply IVR (interactive voice response) systems of the past; they maintain context and can perform entire business processes during a call. The MrCall system, for example, integrates with business tools to book appointments, update a CRM or send a follow-up SMS, all in a single conversation. This allows a caller to reschedule a meeting or place an order by talking naturally, while the AI simultaneously records information in the corporate database. Response latency is now under a second, and the technology allows users to interrupt the AI as they would in a human conversation. With MrCall, we have observed that users have moved from using short sentences to expressing themselves in increasingly natural phrases, favored by the high voice quality and low latency.

AI Agents as Interface

The success of phone agents anticipates a broader change in the way information and services are accessed. Back in 2018, in various talks at AI conferences, we argued that AI agents will become the primary interface with machines, likely taking the place now occupied by websites and apps. A recent analysis predicts that if AI agents deliver on their promises, the internet will go dark: consumers will no longer visit websites directly for many activities. This does not mean the complete disappearance of sites, but the importance of Search Engine Optimization does. The year 2025 has been called the year of AI agents for a reason.

How could MrCall's AI agents replace websites? The fundamental reason is the convenience and natural language of communication. Instead of navigating to a site or app, users simply express what they need, and the agent performs all the necessary actions on the back-end. For example, instead of filling out forms on an airline site, we could simply say, "Book a round-trip flight to Rome for next Monday afternoon," and the AI agent would handle everything.

Our vision at MrCall is to build AI agents that become the primary point of contact between the company and its audience, across any channel. Today's phone answering agents will be extended to web chat, messaging, and more, creating unified, multimodal conversational interfaces.

Agent-to-Agent Communication

Looking further ahead, an even more radical vision emerges: AI agents autonomously communicating with each other on behalf of humans. In the future, the AI personal assistant could talk to the AI agent in the restaurant to reserve a table, completely eliminating the need for human intervention.

This is not science fiction, but a natural extension of current trends. Standards may emerge to facilitate this communication between AI agents, similar to APIs or shared protocols. Companies with the best AI agent or agent infrastructure will become crucial, as they will handle a large part of future commerce and communication.

In MrCall we see our mission in exactly this light: today we build AIs that answer the phone, tomorrow these same AIs might negotiate with other AIs or handle transactions initiated by personal assistants.

Re-imagining the Web

In conclusion, call-answering AI agents represent only the first phase of a deeper transformation of digital interfaces. MrCall's vision goes further: to become the central platform of business communication, progressively replacing websites and apps with conversational, multimodal AI agents. This is our bet on the future of digital communication, which we want to share with our users.