Catching Up With NJ Startup SpeechTrans: Transitioning to Services for Company Growth

Lyndhurst-based SpeechTrans is a New Jersey tech startup success story. The company began with a speech language translation app that took the Apple app store by storm. However, as most app developers know, apps alone don’t make a business.

NJTechWeekly.com caught up with John Frei and Yan Auerbach, cofounders of SpeechTrans, in early January. The two explained how they had pivoted from makers of an advanced app to a full-service company focusing on “Results as a Service,” as Frei called it.

“We had to pivot to this new kind of company to enable sustainable growth,” said Frei. “It’s not just software, it’s not just hardware, it’s both. We’ve done this successfully. We’ve quadrupled our revenue and increased our valuation from $1 million to $10 million in our current round of financing, and we see nothing but hockey stick growth ahead.”

“We are integrators and automation experts,” he added. “It started out as speech recognition, but now we have a productivity play. We come in and provide added value to our clients. The speech recognition and language translation is always there.”

Auerbach explained that, in addition to its traditional products, the startup now offers new services that fall into two categories: those for companies that already have applications and those for companies that don’t. For the first category, “we offer companies the ability to do it themselves,” he said. “ We have a SpeechTrans API and SDK  [software development kit], and companies can incorporate our products into their existing services. That’s what our customers Hewlett Packard Enterprise (in HP MyRoom) and Kingsoft (makers of WPS Office) did. They had existing software that does Web conferencing, and they just wanted to add our speech translation on top of that.”

If a company doesn’t yet have an application of its own, the startup can build one for them and incorporate the SpeechTrans backbone into it, said Auerbach. “We are working with a corporation from Scottsdale, Arizona, that is building a brand-new electronic health record (EHR) system. They are opting in for our services where we use our software as the baseline, and we custom-build the application for the customer. They are hosting it on our servers.”  

Frei said that the startup is actually becoming a virtual CTO or CIO for other companies in some cases, customizing solutions for them based on tools and automation originally designed for SpeechTrans. Smaller businesses can’t afford to hire a full-time CTO and CIO, he noted, so SpeechTrans is able to do the work on a much cheaper basis.

He added that the Film Society at Lincoln Center has used SpeechTrans’ hardware and software integration services. “There is a new law that said that the cinema had to provide closed captioning for all of the movies that showed at the theater. This was very expensive to do, especially for a nonprofit.”

Sony was the only company that could provide appropriate hardware and software for this application, Frei explained. The Film Society would have had to switch out all of their equipment and acquire more than a half million dollars’ worth of Sony equipment.

SpeechTrans had developed a cloud solution for the hearing-impaired that delivers closed captioning at a lower Software-as-a-Service price point. “We came in as a subcontractor for Hewlett Packard Enterprise and got it running,” Frei said. “This is a very complicated operation with extreme latency considerations. If we were even a half second late in bringing the content to the viewers, it would have been considered a fail. The text has to show up at the exact same time that the words would be spoken.” 

The closed captioning is displayed on a headset, Auerbach noted. “We are now using Epson glasses, but it could be Google glasses or any other augmented reality headset. The viewer sees the caption in AR right in front of the screen.”

Frei added that, due to its language translation roots, the software can provide closed captioning in 144 languages. “Now we are looking to expand that success to other theaters throughout the nation,” he said.

One of SpeechTrans’ traditional speech recognition and translation customers is the Visiting Nurses Association, in Brooklyn. Some 900 nurses are currently using the system for live human translation for Medicaid and Medicare purposes. In this case, the company provides a language telephone line for a fraction of what other language lines cost because it has streamlined the backend services, offering a guaranteed answer time of less than a minute in all 144 languages, according to Frei.

This is as low tech as it gets, he said. “We use live translators because we need to be able to service enterprise clients and medical facilities that require both. We do the low tech stuff, get our foot in the door and transition them to automated language translation when they are comfortable with it. It costs them much less.”

The company has been working with Zapier and the Arizona-based EHR company, which has integrated the SpeechTrans APIs into its service. The upshot is that doctors can dictate patient medical notes to any device. “We tie it into Zapier and plug the notes into the medical records system,” Frei said.

The company has also tied wearables into its system. For instance, SpeechTrans has its own branded smart watch. Doctors can engage in live video conferencing with the watch or dictate their medical notes into the device. And if they have a patient they need to converse with, they can speak into the watch, and the translation will be spoken back out loud to the patient.

So what is next for SpeechTrans?

“We are focused on hands-free commands like Cortana and Siri, integrating these commands into our voice recognition and translation tools for customers who are using heads-up displays and want to be more productive, said Frei. “We are looking to grow the services business out as well. …We have some interesting opportunities working with telcos, although that is a longer sales cycle.

“As you know, we partner with HP, and are already on 10 million HP devices. We partner with Intel as part of the Intel Business App Portfolio, which will be on 40 million Ultrabooks and tablets. With partnerships with Vodafone and Kingsoft, we have the opportunity to scale to over a billion devices in the next two years! So we have a lot of exciting growth opportunities.”

SpeechTrans is also eying the possibility of opening a new office at the Bell Works facility, in Holmdel, and is in discussions about this. “We are probably going to transition over there after the first quarter,” Auerbach said. “We are currently at Bergen Community College, in a good space. The problem is that they don’t have 24-hour access, and since we are building out our own servers we need to get into the space in case they go down. That’s the impetus for the new location.”