Why we invested: Voca

Today, we’re excited to announce our investment in Voca, who’s on a mission “to enable businesses to reach their customers at scale, with a personal human voice.”

To set the scene behind this investment, we’ll start with one unsurprising fact — four years ago, no one in the world owned a smart speaker. What may be more surprising is that by the end of this year, one in four US households will own at least one, and by 2020, Gartner predicts ownership will reach 75%. People are taking to smart speakers significantly faster than they did with smartphones. Similar to the way Facebook trained us to scroll unconsciously, the Amazon-led proliferation of home speakers will mark another profound shift in consumer behavior — we increasingly feel comfortable speaking to machines, saying “thank you”, “please”, and “sorry” to conscious-less algorithms. While these new interactions between consumers & computers may have captured the hearts and wallets of many U.S. households, they have largely failed to produce meaningful conversation. Other than serving as a music player and an audible news feed, “smart” speakers have thus far been pretty stupid. We think this is about to change.

In May 2018, Google released a demo of one of its stealth initiatives, “Duplex”, which successfully booked an appointment on behalf of a human. Although the tech giants have been racing to develop effective virtual assistants for the last decade, this was the first time we saw a virtual assistant lead the conversation, rather than responding to one-off requests. This demo resulted in public outcries, as many people were uncomfortable with their inability to distinguish whether Duplex (sounding eerily similar to Dupe) was a human or computer.

Duplex was the result of significant advances in machine learning and data acquisition — more accurately, as machine learning became more sophisticated as a methodology to understand data and make increasingly accurate predictions and decisions to the millisecond, so too did the corpus of voice data available to Google. The combination between machine learning & voice data will inevitably lead to a set of algorithms capable of simulating two-way conversation, penetrating every industry, increasing productivity, improving customer experience, and enabling new business models.

Enter Voca.

Voca Founders, Einav Itamar (CEO) and Alan Bekker (CTO),along with VP R&D Danny Ruchman (left to right)

Led by CEO Einav Itamar, former co-founder and CTO of Corrigon (acquired by eBay) and CTO Dr. Alan Bekker, one of Israel’s leading deep learning researchers previously at Intel, Voca has built a unique technology that believably replicates human experience in two-way conversation. Given the endless ways in which humans interface with voice, Voca’s award-winning technology has thousands of potential applications, but the company has chosen to focus its initial efforts on improving a truly broken industry — call centers.

Think about the last time you spoke to a call center agent on the phone. Who was more frustrated? You, or the human on the other end repeating the same robotic script for 8 hours a day? There are more than 3m agents employed in US call centers commanding an annual salary of $150b, fraught with accent & language issues, high turnover (33%/year), emotional fatigue, and regulatory violations due to human error. Voca automates the parts of the conversation where humans are forced to act like robots and give enterprises the ability to personalize voice interactions with their customers at scale, without losing the human touch.

Voca’s virtual agents, change the first point of interaction on the phone from a robot lacking empathy and natural language understanding to a virtual assistant that can adjust the voice, language, accent, pitch, and tone, as well as its responses to provide consumers with a pleasant, helpful, and non-judgmental experience (see Voca in action here).

Voca has seen substantial interest from a variety of prospects looking for support in automating subscriptions renewals, ID/account verification, customer upsells, payment reminders, appointment scheduling, customer satisfaction surveys and reducing the time SDRs spend speaking to unqualified prospects, to name a few.

Within this broader shift towards conversational AI, I am incredibly excited to welcome Einav, Alan and the growing team at Voca to the lool family and supporting them in their mission — to enable companies to reach their customers at scale, with a personal human voice.