Sherpa, a personal assistant startup that has carved out a niche for itself by focusing on the Spanish-language market (alongside English) and predictive suggestions, is expanding. The company is launching a set of APIs called the Sherpa Platform, which will let other businesses tap into its predictive recommendations and use them in their own consumer-facing services. Sherpa — based out of Bilbao, Spain and Palo Alto — is also announcing its a customer for the service: Porsche, which plans to use the service in its connected car services in its luxury vehicles.
In addition to the automotive sector, Sherpa plans to target the home and mobile segments with its Platform APIs, and it has some deals specifically with other automotive companies and telecoms carriers in the works.
The expansion comes at the same time that Sherpa has passed three million downloads of its app, which currently has 800,000 active users, with 80-90 percent of those Spanish. Founder and CEO Xabi Uribe-Etxebarria said that growth has been largely word-of-mouth, and that the primary aim up to now has been not scaling out — the app has been free and not trying to monetise — but gathering enough users to help train its systems as it continued to build out its product.
“There were two reasons for launching Sherpa Platform: one to start monetising since we hadn’t before,” said Uribe-Etxebarria, “and two because we saw that we had a lot of interest from telcos and car manufacturers for a B2B2C product.”
It’s also in the process of raising money. Sherpa has so far grown up on a very modest $6.5 million of funding, from Alma Mundi Innvierte Fund, FCRE, and unnamed private investors (“celebrities” says Uribe-Extebarria). The company is close to completing a bridge round for later this year of around $8 million, ahead of a larger Series B. Uribe-Extebarria says he has spoken to “all the usual names” in the US — he splits his time between Spain and California — and also a number of investors in Europe.
There have been a number of companies doubling down on using machine learning and natural language processing to develop personal AI systems that respond to voice commands to either provide information or carry out simple digital tasks, either on their own devices or on those of third party hardware makers: Amazon has its Echo speakers and Alexa; Apple has Siri and a range of hardware that runs it; Google’s Assistant goes everywhere that Android does; Microsoft has Cortana; and even Samsung (once a close partner of Sherpa’s) has rolled out its own Bixby assistant.
Sherpa has, in that context, worked to differentiate itself in a few ways.
The first of these was being an early mover — Sherpa was founded in 2012 — in building out a system for Spanish speakers, covering a number of different regional dialects in what is considered to be the world’s second-most popular language after Chinese. Several of the large tech companies that have built personal assistants, such as Google, Apple and Microsoft, all now also support Spanish, although Alexa does not; but none have as extensive support as Sherpa, which currently has seven types of Spanish dialects, and four for English. “We are focusing more on quality than quantity of languages,” he said. That said, the company does have plans to add more languages next year — an effort that it will be raising money to target
The second of these has been a focus on predictive technology, not just answering questions dictated into the app, or carrying out small tasks, but also providing a voice-based interface that talks to the user in the same way that a human personal assistant might do: it learns what kinds of things you might want to know about, and then it proceeds to tell you about these, before you ask. This could be notifications about new incoming emails from specific people that are then read out to you — a handsfree experience that comes in handy in situations like driving — or news about subjects that you follow.
Uribe-Etxebarria says that the predictive engine is currently the crux of the company’s technical development: it actually uses third-party technology for automatic speech recognition and text-to-speech capabilities. “What we focus on is using natural language processing for predictive recommendations and conversation,” he said. “For us, the other tools are now commodities, since there are other companies that do these well, and that are cheap. That’s why we want to focus on things that others do not do, and providing recommendations, and cross-domain recommendations, is what we are good at.”
The third way that Sherpa has differentiated itself to date is that it doesn’t have any skin in the game in the way that the other businesses that have launched personal assistants do.
“We only focus on personal assistants that provide predictive capabilities,” he said. “And I’m sure that the others will do this better and better, so even this isn’t our complete advantage. Ours is that we are an independent platform. It means that companies like Porsche or others do not see us as a potential competitor, but a partner.” This is especially notable, considering that the larger tech companies have long been seen as competitors among carriers, and now more recently automotive makers, who may need to rely on them for some connected car capabilities but do not want to fully relinquish ownership of their customers in the process.