Rajiv Parikh
VP of Marketing

Rajiv explains how important accurate STT is to Nytro.AI’s pitch intelligence platform for sales enablement. They tested other STT solutions and only saw 75-80% accuracy but with Deepgram they have seen 90%+ consistent accuracy.

Read the case study


Elerian AI

Craig Akal, Co-founder and Director

I’m Rajeev Parikh. I am Vice President of Marketing at a company called NytroAI N Y T R O dot AI. We are in the sales onboarding, sales readiness, sales coaching space, and we have a pitch intelligence dashboard that enables sales reps to practice and perfect their pitch. And we use Deepgram for their speech-to-text technology in our in our backend. So Nytro.AI  is in the sales training, sales onboarding, sales readiness space. And we’re solving a critical pain point, which is really enabling sales reps, SDRs, account execs, BDRs, etcetera, to really practice and master their sales pitch. And we use artificial intelligence and we created something called Pitch Intelligence, which enables sales reps to really understand how well they’re doing when it comes to a particular pitch. And we use AI to grade score and rank that across a number of important dimensions.

So our Nytro.AI’s pitch intelligence platform involves includes three main things. Number one, it’s it’s really identifying that visual body language, the facial expression of that of that sales rep. How well are they, you know, sort of, performing, if you will, on camera or on a call. Number two is sort of the vocal quality. So the tone, the clarity of speech, the words per minute, how well are they annunciating that pitch? And the third and probably most critical part is where Deepgram plays a part, which is that text to to speech. Translating that text so that the the algorithm can actually then use that translated text to identify the key components, the clarity of speech, the tone, the words per minute, the keywords that were covered. So accuracy of that that speech to text conversion is really really important for a platform like Nytro.AI.


When we discovered Deepgram, we found that the accuracy was ninety percent plus.

There are a lot of different open source offerings out there that exist today. And those open source offerings are pretty good, but their accuracy rate is what we found to fall with between the seventy-five percent to eighty percent accuracy. When we discovered Deepgram, we found that the accuracy was ninety percent plus. So so Nytro.AI is, you know, we’ve already seen this great success with Deepgram’s general speech model as part of our core offering. And now we’re beginning to explore possibilities of building tailored domain specific speech models with Deepgram. So, Nytro.AI is really already exploring, integrating Portuguese, for example, and Spanish models, to to address, you know, more of the global SDR. And with, you know, the partnership with Deepgram, that that accuracy has been on point. And so companies, enterprise companies as well as midsize companies out there are benefiting from our technology. Thanks to the partnership of Deepgram.

You know, the future is really around voice analysis, around AI, that predictability, that accuracy. So I would recommend, you know, if you’re if you’re considering some of a technology like speech to text definitely give Deepgram a look.

Read the case study