Chatbots, whether revolutionary or AI-powered toys, are more than just talk

Jonathan Lehr is not just an investor in x.ai, he’s a power user of the artificial-intelligence-powered scheduling chatbot. Last year, Lehr scheduled more than 1,000 business meetings using x.ai.

“We are a small fund and I don’t have a personal assistant,” says Lehr, a managing director at Work-Bench, which recently participated in x.ai’s $23 million Series B round with Two Sigma Ventures and DCM Ventures. “I used to stay up from 11 p.m. to 1 a.m. every night just trying to schedule meetings.”

Not anymore. Now, Andrew, his x.ai personal assistant, handles all the back-and-forth required to schedule meetings, saving Lehr countless hours every week.

Bots, like the ones developed by New York-based x.ai, are one of the sexiest sectors right now, with investors pouring money into the most promising startups. In the first seven months of this year, 15 bot startups raised their first institutional capital, according to research firm CB Insights.

Bots are computer programs that act a lot like humans, simulating conversation with human users over the Internet or via text messaging. They perform a series of automated tasks, such as helping users schedule meetings, manage finances or find the best Chinese-takeout joint.

Some bots even serve as personal shopping assistants. Mezi, for instance, is an AI-powered free shopping app that recently raised $9 million from Nexus Venture Partners, Saama Capital and American Express Ventures.

Mezi blends AI and human expertise to learn a user’s shopping preferences and create a personalized shopping experience over text messaging. The company maintains that its app simulates the experience of shopping with your best friend through intelligent chat conversations.

“We loved the idea of an intelligent personal shopper that is helping consumers simplify their life and also has the ability to learn from past interactions,” says Rohit Bodas, founding partner at American Express Ventures. “When it comes to shopping, there is too much information overload, leading to decision fatigue.”

Searching on Amazon.com, for instance, can result in hundreds, or even thousands, of product options, and sorting through them can be difficult and inefficient. “That’s why there’s a need for more intelligence on the consumer filtering side,” Bodas says.

Conversely, brands and retailers need to target consumers more efficiently, delivering more personalized and relevant products and services. A chatbot like Mezi makes the buy-sell process a lot more efficient, and on the business-model side it takes a cut of every transaction it facilitates.

“We also really like that Mezi is focused purely on commerce,” Bodas adds. “AI is a very difficult technical problem, and trying to solve it for every category is extremely difficult. So having a narrow focus on a particular category makes a lot of sense from an investor perspective.”

Another interesting niche for chatbots is personal finance. Digit, which makes a bot designed to help people save money, recently raised a $22.5 million Series B round from Ribbit Capital and General Catalyst Partners.

Every few days, Digit analyzes a user’s spending habits and puts money in a savings account. The company says it has helped users save more than $125 million since its launch.

Trim, another personal-finance bot, pocketed $2.2 million in July from Eniac Ventures and the actor and investor Ashton Kutcher. The bot helps consumers cancel unwanted or forgotten subscriptions to everything from video-streaming services to gym memberships.

The company estimates that if all Americans had a personal financial assistant, they could collectively save more than $100 billion per year.

“Most people would love to save money. Yet they have anxiety around things they should be doing to save money, but are not, like shutting down that renter’s insurance from two years ago,” says Tim Young, one of the founding general partners at Eniac.

“Having a back-end AI system that proactively searches, finds and eliminates all those unwanted charges is a phenomenal opportunity.”

Because the company knows how much users are paying for things like car insurance, it can partner with insurance providers to offer better deals to consumers. “The system can see all your payments and how much you spend, so it can benchmark that data against the community and identify people who are paying more than they should be for a particular service,” Young says.

Bots aren’t just for consumers; they are also invading the enterprise. Slack Technologies, an enterprise-software company, has launched an $80 million fund to invest in office-productivity bots. To date, its investments include Sudo, a bot that enables employees to communicate with their companies’ customer-relations-management systems, and Butter.ai, a knowledge-sharing bot.

It’s these enterprise-focused bots in particular that are energizing investors like Phil Libin at General Catalyst. He made two bot investments recently, Butter.ai and Growbot, which enables employees to give feedback and kudos to team members in an organized way.

In a LinkedIn post in July, Libin outlined two kinds of bot startups: fart bots and consequential bots.

Fart bots “are either experimental toys or were rushed to market with use cases that don’t make any sense,” he says. “You don’t need a chatbot to tell you the weather.”

Consequential bots offer simple solutions to real problems, Libin says. “The consequential-use cases take more than a weekend to build,” he says. “Luckily, we’ve now had more than a weekend, and quality bots are starting to emerge.”

Producing a good bot requires a lot of thought and engineering, Lehr said. “With a scheduling assistant like x.ai, there is a lot of nuance around how much the machine can understand,” he said. “When I say, ‘Let’s meet at Starbucks,’ does it know which Starbucks? Can it learn my preferences over time? When I need to reschedule a meeting, how does it handle that?”

He says these are all things that are programmed into x.ai. “Whether I’m on a mobile device or desktop, I copy my x.ai assistant, and he does the scheduling,” Lehr said. “He doesn’t get sick, he doesn’t go on vacation for a week. He doesn’t forget. And if someone does not respond to a calendar invite, my x.ai assistant, Andrew, will send reminders and will notify me after three non-responses.”

Lehr says x.ai is addressing a huge market. According to his research, about 1 percent of information workers have personal assistants. He says x.ai makes it easy for the other 99 percent to access this service. For instance, human assistants who work remotely can run about $1,000 per month. X.ai, which plans to announce pricing later this year, says its service will cost a fraction of that.

“The market here is massive and democratizing,” Lehr says. “Everyone deserves more time back in their day, not just executives. The beauty of a fully automated solution means that x.ai can grow into a massive company without breaking the bank for users as other competitive services do.”

One challenge for all bot companies is competition from the tech giants, who are eagerly eyeing the space. Google, for instance, is developing a chatbot called Google Assistant that can respond to voice queries and engage in a fairly natural dialog with users.

Microsoft, despite the epic fail with its hate-speech-spewing Tay chatbot, recently introduced “Conversations as a Platform,” which is aimed at creating AI bots that work with Cortana, the personal assistant for Windows 10.

Facebook, for its part, is encouraging developers to build chatbots for its Facebook Messenger platform. This initiative has served as a significant catalyst for startup activity in the space. Trim, for instance, is one of the first chatbots for personal finance on the Facebook platform.

Of course, these tech behemoths will surely be natural acquirers of any chatbot company that gains traction in the market. “The large companies are building technologies in-house, but they will always be looking to complement it with some sort of inorganic acquisition,” American Express Ventures’ Bodas says.

Tom Stein is a Palo Alto, California-based contributor. He can be reached at tom.stein@yahoo.com.

Downloadable Data: startup-targets-developing-ai-powered-bots

Photo courtesy ©iStock/Viktoriya Sukhanova