In 2006, Dharmesh Shah founded HubSpot with his friend and MIT classmate Brian Halligan. HubSpot provides inbound marketing and sales software to companies all over the world. By 2011, HubSpot was serving more than 5,000 customers, placing the company at No. 33 on the Inc. 500. But they didn't stop there: In 2014, HubSpot went public and has a market cap of $1.75 billion.

Yes, billion with a "b."

By any standard, Halligan and Shah are living the entrepreneurial dream. So why has Dharmesh been in, as he describes it, "mad scientist mode" for the last several months? He's been working on what he says is one of the most ambitious projects he's worked on in recent years: GrowthBot, a chatbot (digital assistant) for marketing and sales. GrowthBot was just released into public beta this week.

In short: What is the founder and CTO of a publicly traded, billion-dollar company doing still writing code? Is that really the best thing for HubSpot? And why a chatbot, and why now?

Those are all great questions. So I asked him.

Let's start with the big "why?" Why build a chatbot?

I've experienced three pretty big shifts in technology during my professional career.

The first was in 1997, when I encountered the internet. That was really something. I realized then that this whole "access information with a Web browser" was going to be a Very Big Deal.

I was running a financial services software startup at the time. So I started building a software application that let people access their retirement savings accounts online, using this newfangled thing called the internet. It was the first of its kind, and the idea turned out to be a hit. It powered the growth of that company, causing it make the Inc. 500 list three years in a row.

The second tech shift was a decade later, in 2007, when the iPhone was launched. It was the breakthrough year for the smartphone. Mobile apps let people walk around with the internet in their pocket.

Now almost another decade has passed and the third tech shift is happening. It's what we're seeing today in messaging and chatbots. Chatbots allow people to access software using a messaging interface, like SMS, Facebook Messenger, and Slack.

This, too, feels like a Very Big Deal.

What makes chatbots so special? Will they kill mobile apps?

I don't think chatbots are going to replace all mobile apps. What fascinates me about them is how they remove friction and let people go straight to what they want to do.  

Mobile apps have to be downloaded and installed. Then you figure out the interface and try to find where in the app the feature is that you need.

Here's an example: You're standing in line waiting for an Uber and an idea for a blog post strikes you. Assuming you already have a mobile app for your blogging system, you then have to navigate to that app, tap to go the feature that allows you to start a new blog post, type in the title, and then hit "save."

Contrast that to a chatbot.  You are likely already in a messaging app. You simply message "start blog post Why Bots Are Awesome" to the chatbot. Just one step. No download. No login. No hunting for the feature. You just type what you want and the software figures out what you want to do and does it. It can be magical.

You're the CTO of a publicly traded company. Why are you still coding? Haven't you risen above that, so to speak, by now?

I just love building software. It gives me such joy. I'm a classically trained software developer. My undergraduate major was Computer Science, and I've been writing software for well over 25 years. I see coding as a craft, and as with any craft, you get better with practice.

I think I'm a better CTO when I stay close to technology and the craft of coding. Distancing myself from it doesn't make me happy and doesn't help the company.

It's always puzzled me that we expect engineers and technical entrepreneurs to just move into management and stop coding. Why? In other disciplines, we encourage people to develop their craft and keep practicing. If you're a writer, you write. If you're a surgeon, you practice surgery. If you're a musician, you play. Yes, life requires other things of you sometimes, particularly if you're trying to build a business. But the goal shouldn't be to grow so you no longer do what you love.

The goal should be to build a team around you so that you can do what you do best--which is likely what you love.

Is GrowthBot intended to be an extension of the tools HubSpot already provides, or does it take you in a different direction?

Though the audience for chatbots reaches beyond current HubSpot customers, it is very closely related to what we do. HubSpot sells marketing and sales software. GrowthBot is a tool built for marketing and sales professionals. My hope is that not only will the tool be useful to existing HubSpot customers but that it also extends HubSpot's brand and reach to prospective customers.

HubSpot believes strongly in the idea that you should create value for your market before you to try to extract value. We have reached millions [of users] through our free content, free education and training, and free tools. GrowthBot extends that idea by offering another useful (and free) tool built for our target market.

Having an idea is different from acting on that idea. Talk me through the nuts and bolts of developing the bot.

I have been obsessed with the idea of chatbots for a long time. Back in 2004, before I started HubSpot, I was tinkering with a startup idea that involved building a text- or email-based interface for enterprise software systems.

Basically, a chatbot.

The problem was that back then people weren't used to using messaging systems yet as a way to interact with software. And there weren't any libraries or other frameworks for helping with the Natural Language Processing (NLP) that chatbots need to really be approachable. At the time, most text-based systems required users to learn some arcane syntax. This made them inaccessible to mere mortals.

Things have changed.  First, people are using messaging apps a lot. Facebook Messenger alone has 900 million active users. More and more people are using messaging apps and they're spending more and more time in them.

And they're not just using messaging apps to talk to other people, they are increasingly using them to shop, buy things, do research, and to access a variety of different services. Plus, the technology has evolved to the point that it's finally possible to deliver relatively good natural language processing. People don't need to learn some secret language to make a chatbot work. They can just ask for what they want the way they normally would. That's huge.

So what wasn't really viable 12 years ago is all of a sudden not just viable but actually inevitable. We're all going to be using chatbots to interact with our favorite companies and services in the coming years.

Of course, not all traditional Web apps or mobile apps will be replaced by chatbots. Just the ones where a text-based interface is quicker and more natural to use. It'll happen fairly organically, now that the groundwork has been laid.

So the real question is: Why wouldn't I work on a chatbot that helps people do their jobs? I mean, really.

How can people check out your chatbot? And what does it actually do?

Trying GrowthBot is easy-breezy and free. There's nothing to download, nothing to install, no form to fill out, and no signup. Just visit GrowthBot.org and click a button to get going. It takes less than a minute.

In terms of what it does, here are some real-life examples of messages you can send:

  • Give me company info about hubspot.com (or any company)
  • Tell me about dharmesh@hubspot.com (or any person, based on their email)
  • Show me law firms in Boston that use Google Apps
  • Add contact dinesh@piedpiper.com to CRM
  • What are the top blog posts about growth hacking?
  • Show me a marketing cartoon

Those are just a few examples. The bot is learning how to do new things every week.

Published on: Jul 14, 2016
Like this column? Sign up to subscribe to email alerts and you'll never miss a post.