HubSpot acquires Kemvi to go deeper into AI

Kemvi's AI tech helped sales and marketing reps create better relationships with clients

If you ask entrepreneurs and VCs about which technologies most excite them, you're going to hear a lot of the same answers: virtual reality, Internet of Things, and machine learning. One of the most anticipated is artificial intelligence, and Gartner predicts that, by 2020, AI will be in almost every new software product and service, and that it will be a top five investment priority for more than 30 percent of CIOs.

AI is going to change nearly every vertical, including sales and marketing. Inbound marketing and sales platform HubSpot has already made artificial intelligence a key part of its platform and is now going even deeper with that technology with the acquisition of Kemvi, an artificial intelligence and machine learning startup that helped sales representatives deepen their relationships with prospective buyers. No financial terms of the deal were disclosed.

Founded in 2014, Kemvi's algorithm was called DeepGraph, and it focused on extracting information from language. It would crawl the Web and extract information from blog posts, press releases, articles, SEC filings, and then structure that information.

For example, if a company releases a product, or if they change their CFO, Kemvi would recognize what was happening from a headline or the first paragraph of an article, and then put that information in a structured database, and then use it to identify which companies its customers should be reaching out to, as well as what they should be saying to them.

"Basically, what we were able to do was generate extremely customized messaging specific to individuals based on what's happening with either those people or the companies they work at so that you could send out large campaigns in a way that every single message was customized to the recipient. The effect of that, of course, was that your response rates go up, your open rates goes up, and that propagates through your entire sales funnel," Vedant Misra, Founder and CEO of Kemvi, explained to me. He is now a member of the engineering team at HubSpot.

Part of that the company was working on was prospecting, or giving its customers the tools to increase their open and response rates on their outbound marketing campaigns. By using AI to automate the process, it was able to dramatically increase those rates.

“If you're sending out cold campaigns, your response rate, on average, where you're sending a whole bunch of emails that are not customized is generally pretty abysmal. What we were seeing with some of our campaigns was response rates on the average of about 15 or 20 percent, sometimes as high as 25 or 30 percent. That was largely because the content we were sending out was, generally, very, very specific to the extent that it looked like a human had researched and read about you, but, of course, it was all automated," Misra explained.

"The way we think about this is that what's happening is that the bar for capturing people's attention is getting higher and higher, and that's because your e-mail inbox is an arm race and everyone's competing for your attention and attention is zero sum. That dramatically increases competition which raises the stakes to get people's attention."

Going forward, HubSpot will be shutting the company down and incorporating its technology into HubSpot CRM, while Misro and his co-founder at Kemvi will be working on the artificial intelligence machine learning group. Kemvi's customers will be transitioned over to HubSpot.

By becoming a part of HubSpot, the Kemvi team will be able to take its technology and apply it to inbound marketing rather than outbound, something that Misra calls, "a very exciting proposition."

"HubSpot has a truly colossal amount of data, tons and tons of happy customers and they have an end to end platform for SMBs to grow their companies. We're excited about bringing in the technology because what we were focused on was outbound, and that's really not what HubSpot does. Outbound, fundamentally, is less good than what HubSpot does, just because people don't like getting prospected. People like engaging with people who they have a relationship with. It's a little warmer, it's not just cold outreach. It's more compelling to bring the tech that was leading to great open rates and response rates into a context where people are actually excited to hear from you, they like getting your content," he said.

At the same time, HubSpot CRM doesn’t just focus on inbound marketing, but "the entire company growth process." That means Kemvi’s technology will be used to make the entire transaction more efficient.

"The tech that we built, customization was part of it, but other side of it is reducing inefficiencies in B2B transactions. If you think about it, the whole B2B transaction ecosystem is an extremely inefficient market, and it's gradually becoming more efficient as the transaction are getting costs lower," said Misra.

While sales teams 15 years ago were made up of people with prized Rolodex, now it has largely been automated.

"The tech that we built comes into that context, in terms of identifying where in the entire transaction ecosystem there are good opportunities for buyers to connect with sellers. To support that, not only do we have data from the customers using HubSpot's new CRM, but also tons and tons of data on what captures people's attention on the Internet. HubSpot is at the center of putting out so much blog content that what people click on tells you a lot about human preferences. Using that information to reduce inefficiency in transactions, we're going to be supplying that tech."

As I mentioned earlier, AI is becoming a part of every space in tech. So what does the future of AI look like? Misra sees a world where we are going to have to grapple with what kind of work we want machines to do.

"What I see in a five to 10 year time scale is a world in which the tools that we're using at work, become increasingly symbiotic with us. We retain executive control in the decisions we're making, but a lot of the content generation component, or collation work that we do, when we're collecting information, or generating material, that's all going to be increasingly facilitated by machines. That kind of work will become a lot more collaborative with machines," Misra said.

"We're going to start wondering what creativity actually is, given that, when we are being creative, largely all we're doing is combining things that we've seen already, and that's in fact something that we can get machines to do today. Generative models for images and text are creating things, so we're going to be asking ourselves some pretty hard questions about what it means to be creative, what it means to be a person, and where the limits are of what we're comfortable with machines doing for us. That's one way to think about how future is going to unfold."

Kemvi had raised $1 million in funding from investors that include Kepha Partners, NeoTribe Ventures, Seabed VC, Victory Ventures and numerous angels.

Do you have a startup in the spaces of Deep Learning, Machine Learning, Augmented Reality, Internet of Things, Virtual Reality, Artificial Intelligence, Analytics, SaaS? Samsung NEXT Start is seeking to invest between $100k and $1 million in pre-seed and seed-stage startups solving hard problems using bleeding edge technologies. Apply here today!

(Image source: deepgraph.io)