Amazon had a significant head start in the voice-based personal assistant wars when it introduced its Echo speaker in 2014. That original Echo, it turned out, was just the first vehicle for Alexa’s artificial intelligence and natural language technologies. Three years later, Amazon is pushing Alexa in all sorts of directions, including new apps, platforms, and gadgets from both Amazon and third parties. And while much of that product vision comes from inside Amazon, the company is wisely looking for the service to be shaped by outsiders, too.
Enter the Alexa Fund. Born with a $100 million investment from Amazon in 2015, the year after the debut of the Echo, the fund is part of the company’s effort to identify more Alexa-showcasing companies and products around the world. It isn’t just a fund. It’s come to comprise three parts–a VC-style investment arm, an accelerator, and a university fellowship program. You can think of these as different ways of reaching innovators and entrepreneurs at various stages, from training to ideation to execution as a real company.
These startups aren’t just sitting around creating new Alexa skills. How the companies’ technologies fit into the Alexa vision is more nuanced than that. Some of the companies work with natural language in a way that could impact Alexa’s language skills and personality in the future. Others provide enabling technologies that help Alexa work in new places or use less power.
In fact, only about half of the companies in the fund’s portfolio make consumer-facing products accessed or operated via Alexa and voice. But all of the startups live in places where Amazon thinks digital assistants are headed.
“We try to lean forward in identifying companies that will map to the next domain, or the next set of experiences that are going to be relevant,” says fund director Paul Bernard, a former Nokia executive with a background in business development and investments.
The First Accelerator Class
When I visited the accelerator in its temporary home in October, inside the Startup Hall on the campus of the University of Washington in Seattle, I saw mostly young men scattered around the shared work tables in rows from the entryway to the back wall. A small sign with the Alexa Fund logo stood on a tripod stand at the front. Nothing fancy–just whiteboards, coffee, and code.
The accelerator’s managing director, Aviel Ginzburg, told me that a big part of his job is working through the Techstars network of accelerators all over the world to locate companies that might be right for the Alexa program. “We are looking for the entrepreneurial pull,” Ginzburg says, meaning that Amazon wants to talk to entrepreneurs who want to pull others into their own strong vision of how Alexa might be used in the future.
“We aren’t going out and looking for companies that are doing what we think should be done with Alexa,” Ginzburg says. “Startups are best at pulling the future forward.”
Ginzburg says he often asks startups if they’ve considered adding a voice capability to their product that allows users to talk to it instead of opening up and app and tapping buttons. Often, the startup founders haven’t even considered the idea, he says; for many of them, it flips a switch in their thinking. The best candidates for the accelerator, Ginzburg says, are those who can see that voice interaction could become one of the “main pillars” of their products.
Among the companies in the first class:
- Play Impossible. The company makes a $99 inflatable ball (“Gameball”) that contains motion sensors and a small brain (a physics engine, like in game controllers), and connects via Bluetooth to a smartphone app. The app guides users to play games with the ball, like Jostle, an enhanced version of keep-away. The ball can also connect with a Fire TV streamer to turn the TV room into a sort of workout or game space.
- Sensible Object. One of the Alexa Fund accelerator companies, Sensible Object, is developing board games that integrate voice in creative ways. In its When In Rome game you fly around to different cities on the board and Alexa asks you trivia questions about that city when you get there. And the voice you hear asking the questions bears the accent of the people who live there.
- Botnik. This two-person startup created a predictive text keyboard that uses AI to generate word suggestions based on a custom set of words (like the full set of lyrics from David Bowie songs or negative reviews of the Statue of Liberty). Cofounder Jamie Brew told me the tool can be used to help creative types (including comics and poets) get the creative juices flowing and face down the blank page. Brew was an editor at The Onion, while his cofounder, Bob Mankoff, spent 40 years as the New Yorker‘s cartoon editor–they’re pretty funny.
- Pulse Labs helps Alexa skill developers test their voice apps with users in the hopes of improving engagement and retention. Pulse recently received a $2.5 million investment from a group of investors including the Alexa Fund and CEO Jeff Bezos’s Bezos Expeditions fund.
- The accelerator’s first class also included Novel Effect, which has a skill that adds sound effects to a story that’s read aloud (by a human) from a book; Comet (formerly Semantica Labs), which helps developers track machine learning code, experiments, and results; and Tinitell, which makes a simple cell-phone wearable for kids.
Amazon and Techstars invest up to $120,000 in each accelerator startup in exchange for 6 percent of common stock. For the selected startups, their involvement also means lots of attention: Ginzburg says last year’s accelerator companies got the chance to pitch their products to 100 different investors over the course of the program.
For the accelerator’s second class—a 13-week session that will begin in the third quarter of 2018—Amazon and Techstars intend to recruit 10 companies that will work in Seattle, and another three “international” companies that will work in both in Seattle and in the London Techstars offices.
After the conclusion of the first accelerator class in October, Bernard, Ginzburg, and others took several months to parse the results, both tangible and intangible.
“From the applicants through to the selected companies, the companies showed us that there continues to be a lot of early-stage voice-related startups across several domains, and that activity is coming from all over the world,” Bernard told me in a January email.
Read more: As Amazon’s Alexa Turns Three, It’s Evolving Faster Than Ever
Like A VC, But Different
The Alexa Fund is run something like a venture capital business in some ways but with a slightly unusual emphasis. Its MO is to locate and support promising new technologies–its focus has mostly been on early-stage startups–but in a wider sense, it aims to extend and promote the Alexa platform. Or, as Bernard puts it, win hearts and minds.
Where a traditional venture capital firm might focus on helping companies find their markets and eventually exit well, the Alexa fund also promotes a cause.
“The main focus is how do we get this storytelling around what’s possible with Alexa by getting the companies excited, as well as the people who hear about it,” Bernard says. With a few big exceptions, the fund’s estimated average size of investment so far has been in the range of $250,000 to $500,000, according to Pitchbook.
The fund, says Bernard, doesn’t try to be the lead investor in any round. “Our priorities are more about supporting them [portfolio companies] on their Alexa integrations.”
These investments often create situations where portfolio companies can help Amazon develop the Alexa platform. For instance, the Fund made an investment in Ecobee, which is Nest’s close competitor in the smart thermostat space. Bernard says the fact that Amazon was an investor in the company opened the door to the two companies working closely together on defining the Alexa API for smart thermometers. In simple terms, the API contains all the technical hooks necessary to let Alexa initiate the control and monitoring of a thermostat based on voice commands given by the user.
“It helps [Ecobee] because they’re getting what they want into the API, and it helps us because we’re getting a lot of access to customers who know what they want,” Bernard explains. “That type of thing happens a lot easier if we’re an investor, because we can tell the organization that our interests here are one.”
For the portfolio companies the capital is important, but the access they get to groups within Amazon may be equally important, Bernard says. The fund provides them with a single point of contact who helps them communicate with all of Amazon’s groups, something the startups are encouraged to do.
That closeness with Amazon hasn’t always been seen as a good thing for participating startups, however.
Amazon faced questions about its potential conflicts of interest last May when it released its new Echo Show device, which looked suspiciously similar to an existing product produced by an Alexa Fund portfolio company called Nucleus. Amazon said it had been working on the idea for the Show well before it made its investment in Nucleus, and that it informed Frankel and company of its product roadmap.
Some have also pointed to similarities between Amazon Key, the company’s smart lock product, and a device built by August, another startup that Amazon invested in through the fund. Amazon released Key in October 2017, over a year it reportedly made a $100 million offer to buy August, an offer the startup turned down.
Amazon insists it doesn’t use information gained through its investments to develop competing products. For instance, its “experience in retail, hardware, cloud technology, and logistics led to the creation of Amazon Key,” a company spokesperson told The Information this week. Still, following the release of the Echo Show last year, fund managers sought to calm the potential anxieties of other portfolio companies by telling founders that a “firewall” exists between the Alexa Fund and Amazon’s product development teams, CNBC reported.
Sometimes the fund’s investments don’t appear to have anything to do with the Alexa itself. It’s very hard to see any connection between Botnik’s text-based AI and Alexa’s voice-based platform, for example.
The same might be said about what’s likely to be the fund’s largest and most high-profile investment to date. This came last August when it took part in a funding round for Essential Products, the consumer hardware company founded by Android creator Andy Rubin. Essential makes a high-end smartphone and a smart home hub. The total size of the round wasn’t disclosed, but Bernard told me it was “in the hundreds of millions of dollars.”
Last week, the investment yielded a modest partnership: Essential announced a new version of its first edition smartphone, a “Halo Gray” phone that comes with the Alexa app preinstalled and will only be available for order at Amazon. There’s also a chance that the as-yet-unavailable Essential home hub will support Alexa in some way.
Alexa Goes To College
The investment fund and the accelerator are the main parts of the Alexa Fund, but it has one other main spoke. Bernard and his team realized that engaging with promising technologists–and potential Alexa ambassadors–after they left school may simply be too late.
To that end, the Alexa Fund’s fellowship program sponsors a PhD candidate or a post-doctorate in a university’s artificial intelligence department or natural language lab.
“There’s an Alexa fellow, who is like the Alexa czar for the AI department, and they create a curriculum for building on Alexa,” Bernard says. So far, the Fund has sponsored fellows at Carnegie-Mellon, Johns Hopkins, the University of Southern California, and the University of Waterloo. Bernard says there’s been lots of interest among other universities to host the Alexa fellowship.
While in some cases it’s not easy to immediately see how the fund’s investments help Alexa, it’s clear that Amazon is trying to carefully seed the parts of an already-sprawling ecosystem.
For startups, it’s an attractive bargain. As we saw yet again at CES in January, Alexa is the place to be if you’re a developer with a product or service that’s naturally summoned up by the human voice. Amazon has put a lot of energy into making integration with Alexa easy for skill developers and device makers. Many of these companies see Alexa as an important new aggregation point for consumers, a natural voice gateway where people access all sorts of internet content and services, and the internet of things. It also doesn’t hurt that Alexa comes with built-in connections with both Amazon’s massive e-commerce platform and Amazon Web Services.
Amazon, after failing with its own smartphone, happened upon a voice platform that people like to use and has so far protected its lead in the smart speaker and digital assistant market. Maintaining that lead in the coming years may depend on how creatively and wisely it pushes its natural language interface into new domains and new use cases. That takes vision, and lots of hard work. One of Amazon’s best sources for that stuff may be the little companies camped out in Startup Hall.
CORRECTION: A previous version of this story said that Amazon’s investment in Essential Products was in the “hundreds of millions.” That was an approximation of the size of the whole round, with Amazon being one of the smaller contributors.
Read more: Can Mycroft’s Privacy-Centric Voice Assistant Take On Alexa And Google?
Recognize your brand’s excellence by applying to this year’s Brands That Matter Awards before the early-rate deadline, May 3.
