Speakeasy AI: Bringing Intelligence to Interactive Voice Response Systems

Follow Speakeasy AI on :

Moshe Villaizan, CTO

Today, the meteoric rise in the development of voice input systems is turning a concept that was considered science fiction a few decades ago into a paradigm-shifting reality. Further, with innovative technologies such as artificial intelligence (AI) and natural language processing (NLP) becoming increasingly relevant in organizations, technology innovators are now looking to utilize these technologies with voice systems to develop intelligent solutions for a diverse range of industries. Among the applications, the one most sought after is the use of AI-powered voice systems in online channels for businesses to enhance an organization’s customer experience. The development of an intelligent voice self-service system that delivers a smooth customer experience, however, has become a challenge for AI solution providers due to the lengthy processes of training an AI to perform effectively in an interactive voice response (IVR) system.

To overcome these challenges, Connecticut-based Speakeasy AI offers a solution that helps companies understand and respond to their customer needs with an AI-powered voice system. Combining end-to-end reporting with real-time insights into customer intents, needs, and outcomes, Speakeasy AI provides voice AI solutions to deliver better voice experiences to their customers.

In an interview with CIO Applications, Frank Schneider, CEO of Speakeasy AI, and Moshe Villaizan, its CTO, discuss the company’s inception, growth, milestones, and research and development in the field of voice recognition technology.

What spurred the inception of your company?

Frank: Back in 2017, Moshe and I worked at a company called Creative Virtual USA—a digital AI chatbot developer. During our time with the company, a majority of our customers began asking us if we could incorporate digital AI into voice channels. This question, coupled with a brief discussion with the former CEO of Creative Virtual USA —now the primary Angel investor of Speakeasy AI—inspired us to start working on a product that would bring AI into IVR. After we exited the company, we turned this revolutionary product into a company, since we realized that our platform had market validation to fill a gap that most other IVR solutions could not.

What are the key features of your voice recognition solution?

Frank: Our system has been developed to integrate any AI system to your corporate IVR. Our speech-to-intent solution allows for more customized and personalized experiences without the transcription based limitations of speech-to-text.

Speakeasy AI’s solution acts as middleware for enterprises to assist them in making natural language IVRs better by integrating any IVR with any AI system

Our design leverages the power of MRCP integration to bring intent based routing and self-service solutions to your legacy IVR. Self-service opportunities increase dramatically, improvement cycles are quick, and enterprises can drive improved voice customer experiences without ripping and replacing existing investments in IVR and AI.

Moshe: Our solution bypasses the transcription step, understands the customer’s voice, and routes to an appropriate intent within the relevant conversational experiences that are already built within the chatbot. Speakeasy AI’s solution acts as middleware for enterprises to assist them in making natural language IVRs better by integrating any IVR with any AI system.

What are some of the market pain points that Speakeasy AI addresses?

Moshe: Most companies in the voice recognition solutions industry offer their clients with a full suite of products and bring in their own NLP systems and conversational experience. This leads customers to build a complete experience in their ecosystem from scratch. Our solution allows companies that have invested in building digital chatbot experiences, to integrate their disparate systems and built-in experiences with IVR. Speakeasy AI achieves this by offering a proxy server that can be utilized by both, the telecoms and IVR team, and the API and chatbot developers to understand the complexities in the architecture of the technologies that the other team is working with. The proxy server acts a middle ground to make it simple for the developer team to collaborate with the IVR team using the appropriate protocols and build intelligent chatbot APIs.

What are the applications of your voice recognition offering?

Frank: Many natural language voice solutions today do not meet customers’ expectations. In many cases, solutions require customers to change their behavior to interact with the voice system. Therefore, we have built our system in a way that flips this equation on its head; our customers interact in a way that is comfortable for them and the platform adapts and learns from these interactions. Speakeasy AI enhances the potential behind the collaboration of different departments—telecom and software developers—to build and improve solutions with a customer’s own voice. This is why we are seeing a 10 to 15 percent increase in call deflection rates in our clients and a 40 percent expansion in opportunities for their voice self-service offerings.

Moshe: Our platform has a learning curve where it starts by using the built-in responses and then gradually learns from client customer inputs to become more specialized and accurate over time. It can integrate services for security and voice authentication with the ability to become progressively better at understanding intent.

How does Speakeasy AI envision the next 12-18 months?

Frank: We are expecting an explosion of opportunities as large companies see that they can bring the benefits they have seen from their chatbots to voice. From a partnership standpoint, we are trying to drive the creation of an ecosystem that allows enterprises to reposition their assets and choose best-of-breed components from the AI and voice realms. In addition, by utilizing the native agility of our platform, Speakeasy AI is planning to collaborate with voice biometrics, analytics, and workforce management solution providers in the coming months.

Moshe: On the product side, we are working to make our APIs publicly available to allow innovators the opportunity to build and connect with our tools and create solutions that can change the IVR and communications market forever.