Speech to Intent vs Speech to Text
We are often asked “what is the difference between speech to text vs speech to intent?” Our patented speech to intent system utilizes AI to analyze the entire audio file – a complete voice utterance – to get to the right intent. This methodology brings intelligence to the way we speak, which is far different than the way we type. We created our speech to intent solution due to the experience we had in the AI/chatbot space and the exciting opportunity we uncovered.
What’s the Opportunity?
Over the last five years chat bots have gone from a let’s try
tool to a must-have part of a business CX roadmap. It has been proven that digital AI solutions can answer over 30% of the questions that would have ended up at the call center, but many customers never try the chat bot, they just call. In fact, over 70% of customer conversations are still using voice. There is a vast opportunity to use AI to answer these questions within voice channels.
What’s the problem?
We have all seen speech to text systems vastly improve over the last three years. Microsoft’s research has seen human parity results in transcription
. It would seem simple to just use speech to text to connect IVR to AI solutions like chat bots and get all the benefits realized in the digital world. In reality, this approach does not deliver as expected. Transcription can create a type of interaction, but transcription is not actually intelligent.
Speakeasy AI’s solution is Speech to Intent™.
||Speakeasy AI Speech to Intent
|Method of operation
||The recognition process is divided into a pipeline of different micro-services that is mapped directly onto the corpus used to train the system.
||Our speech to intent system bypasses the issues of traditional speech to text where accents or poor audio signal affects the outcome greatly. The Speech to Intent system only matches against known content in the AI system, giving a much better match percentage.
||The existing corpus of alternates in the AI system is used to set up the Speech to Intent system.
||The system can be setup quickly because the content already exists. All of the special words and products are immediately recognized by the system.
||We have seen over 80% accuracy in testing with a well-developed corpus.
||Within the Speakeasy AI admin console it is easy to see questions that have been misunderstood and instantly assign alternates to improve the accuracy. Additional content can be added and in production within minutes.
||The matching can be set up specifically to understand what is being said in the context of the business deployment. Maintenance is instant and completely controlled by the company.
||The SpeakeasyAI Speech to Intent system requires fewer resources than speech to text systems.
Our mission is to make it easier for businesses to understand and respond to their customers’ needs in voice with AI. We accomplish this mission by using the world’s first and only Speech-to-Intent™ solution. Combined with our end-to-end reporting, our solution provides real-time insights into understanding customers’ intents, needs and outcomes. And since an AI platform is only as good as it’s improvement cycle, we enable rapid updates to ensure wins are delivered on the day you launch. With our voice AI solutions and our team’s proven expertise, we work tirelessly to provide better voice experiences and deliver understanding as a service.