Creating Conversational Agents: The Future of AI Voice

Welcome to this week’s Deep-fried Dive with Fry Guy! In these long-form articles, Fry Guy conducts an in-depth analysis of a cutting-edge artificial intelligence (AI) development or developer. Today, Fry Guy is exploring AnyCall, an AI-powered call system that can mimic human conversations. We hope you enjoy!

*Notice: We do not gain any monetary compensation from the people and projects we feature in the Sunday Deep-fried Dives with Fry Guy. We explore these projects and developers solely for the purpose of revealing to you interesting and cutting-edge AI projects, developers, and uses.*

“Please hold, as your call is very important to us.”

Have you heard this one too many times? If so, AI is here with a solution.

AnyCall is a platform that allows users to create AI voice agents for their call centers. These agents are trained to understand and respond to customer queries, providing a seamless and efficient experience, improving overall communication for both the HR team as well as frustrated customers.

THE MASTERMIND BEHIND THE PHONE

Chris Estreich, the creator of AnyCall, has personal experience and valuable insights to share when it comes to building startups and developing new products. In his college years at Stanford, Estreich began experimenting with product creation and witnessed a shift towards startups in the early 2000s. He noticed that people were gravitating towards smaller companies with younger teams, so he decided to follow that trend. He eventually co-founded his own company which aimed to create a social network based on a feed of purchase history. Estreich and his co-founder then went on to build a successful e-commerce product and a marketplace for Etsy sellers. After spending about 10 years on that business, they decided to explore the world of AI technologies and tools. As Estreich explains, “We were very excited about the types of products that you couldn’t build a year ago that you can now build [because of AI].” Their venture into the AI voice space led to the creation of AnyCall.

A UNIQUE AI AGENT

One of the key goals of AnyCall is to make AI technology accessible to a wider audience. Traditionally, early adopters have been the primary users of AI products, as they have the technical knowledge to navigate and implement complex models. However, AnyCall aims to bridge this gap and make AI technology user-friendly for individuals who may not have extensive technical expertise. Estreich emphasizes, “We wanted to find ways to bring this technology to people who are not necessarily early adopters.” With advancements in machine learning and AI technology in just the past year, it is now possible to build products that were never feasible before. These new AI voice products have the potential to revolutionize the way we communicate and interact with technology. Now, users who have no coding or programming experience can easily interact with this technology. In the case of AnyCall, it is by creating synthetic AI agents for their calls.

Implementing AI agents in call centers offers numerous benefits. AI agents have the ability to understand and adapt to customer preferences based on voice tone and previous interactions, actually learning about the callers and adjusting future interactions accordingly. Additionally, unlike human agents, AI agents are always available, ensuring consistent and reliable customer service where people never have to worry about office hours or being put on hold.

What sets AnyCall apart from other AI call integration platforms is the ability to customize the personality and behavior of the AI agents. Users can easily configure the AI agent’s voice, personality, and goals through a user-friendly platform. Estreich explains, “Imagine a Typeform wizard where you answer basic questions and at the end of it, you have this agent configured to your liking.” Unlike boring robots that have sometimes attempted to make their way into customer service, these AnyCall agents can be tailored towards the specific personality of the company in a simple description. For example, the AI agent could be “a young woman who has a witty sense of humor” or “a grumpy old man who has jolly undertones.” Whatever the user wants to create, the AI platform can provide this unique flair.

Another key feature AnyCall is its support for multiple languages. With high-quality speech recognition technology, the AI agents can understand and respond to customer queries in up to 30 languages, improving accessibility and inclusivity.

While there may be concerns about AI replacing humans, Estreich envisions a world where AI voice agents enhance customer experiences and empower individuals to create unique content and applications. The potential use cases for this technology are vast, ranging from customer support in call centers to personalized voice assistants for individuals. By allowing the AI agent to handle the incoming calls, it allows the human individuals to partake in more meaningful tasks.

WHAT ABOUT PRIVACY?

One of the major concerns for individuals looking for AI solutions, either for their business or for individual use, is the concern over privacy. Where is the data being stored, and for what purpose?

AnyCall provides options for call recordings, tailored to the preferences of the user. While call recordings can be useful for training and improving the AI agents, the platform allows users to control the storage and disclosure of these recordings. This ensures that privacy concerns are taken into account and that users operate with an adequate level of comfort.

WHAT’S IN THE WORKSHOP?

As AI voice technology continues to advance, AnyCall is working on staying ahead of the curve with some innovative improvements to the project. Some of these improvements include:

Addressing interruptions and natural language in conversations:

One of the key challenges in AI voice technology is handling interruptions and natural language in conversations. Teaching these agents to talk like humans is like teaching someone a new language, so it takes time, effort, and a whole lot of brain power. For example, in normal human conversation, we use slang, wavering tones, and interrupt each other often (sometimes too much). It is important to develop systems that understand the nuances of natural language and can handle interruptions seamlessly to provide a more human-like conversation experience.

Reducing latency:

Reducing latency (or delays in response time) is crucial in AI voice technology to ensure real-time interactions and improve the overall user experience. By optimizing the processing and streaming capabilities of the AI voice agents, latency can be reduced, resulting in a smoother and more responsive conversation. This step is crucial for handling aspects of human conversation like the aforementioned interruptions. The AnyCall team wants to create a seamless experience where interruptions feel natural and conversations with AI agents are as smooth as possible, with optimal response pauses that mimic human behavior.

Local AI voice models:

Don’t you hate when your wifi causes videos or audio to buffer? It is the same with these AI voice agents. The potential for local AI voice models on devices will be crucial as the industry develops. By implementing AI voice models directly on devices, it eliminates the need for a constant internet connection and reduces reliance on external servers, resulting in faster and more efficient voice interactions.

As the AI voice agent industry continues to emerge and grow, AnyCall’s Estreich believes the companies that can successfully capture these next steps are going to be the ones that capture the market and truly revolutionize the way humans converse with AI.

CHALLENGES SURROUNDING AI VOICE GENERATION

AI voice generation and manipulation has sparked heated conversations all across the space, as there are important ethical and regulative considerations to address. One of these considerations is the balance between copyright concerns and innovation. Who owns an AI-created voice? Is the question even answerable? With the ability to clone human voices and create realistic AI voices, there is a desire by many to ensure that intellectual property rights are protected while also encouraging innovation and creativity. How to best do this is a challenge that nobody seems quite sure how to solve.

Another important aspect of this emerging use for AI, along with any other AI developments, is the debate between open-source models and closed-source models. Open-source models allow for collaboration and the sharing of knowledge, fostering a more inclusive and diverse AI landscape. On the other hand, closed-source models may offer more control and security but can limit accessibility and hinder innovation. They also draw out concerns about insider data usage and privacy. Striking a balance between these approaches is crucial for the future of AI voice technology.

Lastly (for now), and some think most importantly, there have been numerous emergences of AI voice manipulation in the public sphere. We have seen this in fake ads from Mr. Beast, Taylor Swift, Scarlett Johansson, and Tom Hanks, where their image and vocal likeness was manipulated to deceive people into thinking they were promoting certain products or services that they had no business with. We have also seen this in the political sphere, as deepfake image and voice generation has been used to trick people into thinking Joe Biden, Donald Trump, and others have made statements they actually never made. How this will play out as the future unfolds and AI voice continues to improve will be quite interesting, to say the least.

The future of AI voice generation holds immense potential, as we can see with emerging projects like AnyCall. It is set to transform industries, elevate human interactions, and drive innovation. And while there are challenges to address, it is undoubtedly true that the power of AI voice is creating a future where communication is rapidly different.

THE FUTURE OF AI VOICE AGENTS

AnyCall is paving the way for enhanced communication and customer support in various industries. By making AI technology accessible to a wider audience, AnyCall is driving the growth and adoption of AI voice for those who don’t have a background in tech. The platform allows users to easily configure AI voice agents in mere minutes, enhancing customer support and improving communication.

As AI voice continues to evolve, it has the potential to transform industries, elevate human interactions, and drive innovation. AnyCall is at the forefront of this innovation, providing a user-friendly platform for creating AI voice agents and shaping the future of AI voice.

WANT TO SEE THE FULL INTERVIEW?

Our FryAI team sat down with Chris Estreich, founder of AnyCall, to discuss the project as well as the future of AI agents and voices. You can view the full conversation here: