What’s really happening right now with chatbots
Image Credit: Shutterstock.com/BlurryMe
Source: Originally published in Venture Beat
By Sanjit Dang
June 3, 2017
One night in Corvallis, with the Oregon State basketball team heading to its 14th straight loss of its season, the talk around the hotel bar moved from the television screen to a far different form of communication — chatbots.
No one used that word, of course. But the technology took front and center in a conversation about Alexa, the voice-activated digital assistant from Amazon that can do everything from play music to order groceries. Give Amazon credit: It has pushed hard to win the hearts and minds of U.S. consumers, teeing up more than 100 mini ads for just that purpose.
Software that can respond to voice or text commands is a high-growth investment sector about to explode in scope and penetration worldwide. Look no further than the response since Mark Zuckerberg announced last year that Facebook Messenger would be opened to bot development. With Messenger drawing 1.2 billion users per month, its corporate parent is now actively encouraging developers to plug in, and more than 10,000 have already done so.
The rise of chatbots also signals a landscape-altering reality: the decline of standalone mobile apps. Alphabet, Facebook, Snap and others see a world in which a spoken or typed word can trigger software that is intelligent enough to handle airline reservations, make clothing purchases, or a reorganize someone’s calendar. It will be the difference between opening an airline app to get your flight information and simply typing or saying, “Grab my boarding pass.”
And it’s already happening. Onvoya, for example, can find a flight based on your travel preferences. Talklocal lets you request a plumber like you’re ordering pizza. And Dotin can use publicly available posts, photos, and other data to help companies pinpoint everything from buying behaviors to good employment matches.
In fact, you can argue the messaging layer of the software stack will be the de facto operating systems of the future, with bots handling most the chores of today’s phone-based apps.
For investors, the growth of chatbots has two important implications. First, it will fuel a rising demand for artificial-intelligence software (AI) to make businesses bot-friendly. Second, it will create a standardized and low-data approach to reach markets still challenged by connectivity speeds.
In other words, bots will make it as easy to sell in India as Indiana.
Battle for the intelligence layer
For the next decade, the key business technology driver will be bots that can connect to both local and cloud-based data centers. That will put a premium on AI routines that can connect voice and text commands to everyday processes.
Startups are already battling it out to become the go-to provider of affordable, AI-driven bot technology. One such example is Botworx. The California company provides businesses with bots that can be tuned to the kinds of conversations they have with their customers everyday — bots that can plug into any messaging service.
There are also niche opportunities in the bot-filled future. One of our portfolio companies, Reflektion, makes the search routines of retailer websites more intelligent — able to match, for instance, a photo of a piece of clothing to available inventory.
Another company we’ve backed, True Fit, uses algorithms to learn why customers of certain body types buy and return clothing, so retailers can provide personalized and highly accurate size recommendations. Imagine taking a photo of a shirt, sending it in a simple text message to a bot, and getting recommendations of three different brands — all guaranteed to fit.
Accelerating the availability of these smarter technologies is a growing set of open source AI tools. Remember when Linux was just a technology for hobbyists? Now it powers nearly all the world’s supercomputers, most of the servers behind the internet, and the majority of financial trades worldwide — not to mention tens of millions of Android mobile phones and consumer devices.
AI is going down the same road. Baidu, Facebook, and Google have released sets of open source machine-learning algorithms. Another group of high-tech heavyweights, including Sam Altman, Elon Musk, and Peter Thiel, have launched the OpenAI initiative. And universities and tech communities are adding new tools to the mix. Simply put, AI routines are poised to be the next great commodity.
Are there still challenges? Absolutely. While we can translate spoken words into text, for example, it’s very difficult to parse their meaning consistently — to pinpoint the right verbs and nouns and take action. But make no mistake: As more progress is made, even more sophisticated AI tools will become freely available.
It will also help drive more commerce worldwide — especially in regions with connectivity challenges.
More commerce as connectivity challenges fade
In many parts of the world, mobile commerce still revolves around low-end smartphones. Available bandwidth is low, data plans are limited, and iOS and Android devices are too expensive for most people.
In fact, the 22 percent of Indians who have smartphones — about 292 million at the end of 2016 — generally use a single device that’s shared by an entire family and holds about 10 apps. The nation’s remaining population lives in villages, well out of reach high-bandwidth connections.
But even though the infrastructure is lagging, people won’t wait. They want the internet to provide them with everything from weather forecasts to a life partner. Dil Mil, an online matchmaking app very popular in the South Asian community and growing rapidly, has a bot that’s going viral because it provides a seamless user experience and works easily across lower-bandwidth connections. A similar service, Neargroup, is gaining popularity in the Philippines.
At the end of the day, people in these regions want information very much; they just can’t afford the data load.
That’s why Facebook launched a bandwidth-friendly version of its namesake product two years ago in India. In February, Facebook Lite hit 200 million monthly users. Similarly, Alphabet has rolled out a beta version of YouTube Go, which lets people with low-bandwidth connections save and store videos offline for future viewing. Because chatbots require the relatively small bandwidth of text and voice data, they can also sidestep connectivity challenges until fatter-bandwidth networks like 5G are widely available.
Every market, every industry
If the potential of chatbots feels familiar, it should. At its core is a promise to connect businesses and customers more intimately and to ease the process of buying and selling. It’s the same promise that fueled the rapid expansion of email, text messages, and today’s mobile apps.
What sets chatbots apart is a layer of deepening intelligence, one that can get people what they need with a few spoken or typed words. This intelligence is the reason businesses of all stripes will start deploying bots rapidly. With deep, vertical data stacks of data already in hand, it’s a logical next step.
And it’s why — in the not-too-distant future — we’ll wonder why we ever used apps at all.
Sanjit Dang is an investment director at Intel Capital, driving investments in user computing across the consumer and enterprise sectors.