Chorus.ai wants to improve sales talk

Micha Breakstone Photo: Javier Lorrenz

The Israeli startup, which has already raised $22 million, is transcribing and analyzing the conversations of sales personnel.

Do you have a good product that provides value for consumers? It might not be enough. One of the important factors affecting the state of companies is their ability to sell their products. Good sales personnel can do wonders even with mediocre products, while less talented colleagues are liable to fail even when the merchandise is good.

Can technology be used to improve the work of sales personnel? Israeli startup Chorus.ai says yes, and estimates that 80% of sales talks fail to yield results. Chorus.ai cofounder and head of R&D Dr. Micha Breakstone says, "Chorus.ai is a company dealing in artificial intelligence. Our goal is to understand how people make decisions and instruct sales people accordingly." The other company founders are CEO Roy Raanani and CTO Russell Levy.

"To sell better"

Founded two years ago, Chorus.ai has since raised $22 million from Emergency Capital and Redpoint Ventures. The company is currently active in the B2B (business-to-business) market, rather than the B2C (business-to-consumer) market. Breakstone says, "For example, we're talking about global technology leaders selling their services in deals that can be as much as $100,000. These are fairly complex deals that can take several weeks and include several meetings.

"We record the talks in real time, transcribe them, and extract information from them in order to help sales people and their managers do a better job of selling. This information is also given to the company's management, which can derive insights from it."

"Globes": How does it actually work?

Breakstone: "In the companies we work with, there are usually hundreds of sales people sitting in Silicon Valley. The management of a company like this has to make decisions about each product - what price is right, what features are important to emphasize in order to persuade the customer, and so forth. Before we came, there were no technical means of coping with this problem.

"In the current situation, we teach the sales people how to sell, and there are training processes, but that's not enough. If we think about a manager who has 10 employees, each of whom holds dozens of hours of conversations a week, we get 200 hours of conversations. Obviously, no one can listen to that. The managers listen to conversations partially and arbitrarily.

"Our product analyzes the content of the conversations. We have a dashboard that displays data about all of the conversations, such as the average length, whether the seller or the buyer talked more, etc. We check which questions arose during the conversation, and what aroused the buyers' opposition - something that can help managers realize that there is a problem."

Are there no similar tools in the market?

"There are attempts at optimization, but they rely on technical things, such as what time of day you should call, how many e-mails to send, and what number to dial so that someone in large companies will answer. For companies with sales in the billions of dollars, these conversations are in effect a black box, and at Chorus.ai, we're developing an X-ray device that completely reveals the data.

"We regard sales personnel as our customers, and it's important for us to offer them an effective product. After the transcription, we can see which questions create a positive interaction with the customer, and when there were hesitations in the conversation. We have developed a search engine that makes it possible to search for any word or concept. This makes it possible to go over the conversation in 60-90 seconds, instead of listening to the entire recording."

What patterns do you detect in sales conversations?

"One of the things we saw was that if the competitors are mentioned during a conversation, it's highly likely that the conversation will lead to closing a deal. The biggest enemy of a sales person is not the competition; it's the possibility of his customer doing nothing. If someone mentions the competition, it's likely that he is already ready to make a deal."

"We worked on the vocabulary"

Breakstone adds that good salespeople do not usually mention the competition. "They wait for the customer to bring up the subject, and then they have to know how to answer him and present the relative advantages of the product they're selling," he says. According to Breakstone, a good sales person, someone who sells 30% plus more than the average, should be taking up 40-60% of the conversation with his talking.

Some of the sales processes sometimes take place in face-to-face meetings, but Chorus.ai does not record the dialogue in these cases. Breakstone notes that many negotiations today take place in video calls (using software like Skype). In a case this this, Chorus.ai's product is able to identify who is taking part in the conversation. In the future, the company is likely to also analyze the participants' body language and derive insights from it.

Breakstone goes on to say, "We're responsible for developing all of the technology. We did it all in-house, and according to tests we conducted, we get better results by a 15-25% margin in transcription than generic products existing in the market. It sounds unrealistic for a company that has existed for only two years and has raised relatively little money, but we improved the engine and worked on the vocabulary and the formats."

Chorus.ai has over 50 customers worldwide and 45 employees, 20 of whom work in R&D in Tel Aviv. The other employees focus on marketing the product from offices in San Francisco.

Why don't you develop a similar product for companies aiming at the end consumer?

"In such cases, the size of the deal is simply of no interest. For example, if you take the sales personnel in cellular or television companies in Israel, you'll see than most of them leave after six months and go on a long overseas trip. They have a limited range of options in this case. You go from option A to option B. You don't have to be a great sales person for that. When deals of NIS 100,000-200,000 are involved, the technology most relevant to you is for you to have a printed page with all the information."

Aren't sales people afraid you will steal their work from them?

"In the distant future, you will talk to a robot sales representative, and you won't know that it's not a person. If someone manages to develop a product like that, and I believe that it will happen in only 20 or 30 years, it will be a Google or a big company, not us.

"Sales people are crazy about us. First of all, they don't have to write things down during the talk. Think about getting a transcript of the interview when we finish talking. In the bottom line, sales people are paid according to the volume of the deals they generate. If you offer them a solution that will save them time, they'll embrace it."

Published by Globes [online], Israel Business News - www.globes-online.com - on August 14, 2017

© Copyright of Globes Publisher Itonut (1983) Ltd. 2017

5 Comments
View comments in rows
Update by email about comments talkback
POST
Comments
Micha Breakstone Photo: Javier Lorrenz
Micha Breakstone Photo: Javier Lorrenz
Twitter Facebook Linkedin RSS Newsletters גלובס MAD Conference 2017