Chatbots and intelligent assistants can be essential customer engagement tools, and each has distinct advantages. So which do you need? Join the Director of Developer Relations at Slack and others for insight into which solution fits your digital customer engagement plan, and how to start leveraging it.

Register now for free.

The question isn’t whether conversational interfaces are the next big thing or not — it’s how to most effectively choose between the two major options, and then how to leverage it.

“I think we’re going to see two parallel processes happening,” says Amir Shevat, director of developer relations at Slack and author of Designing Bots: Creating Conversational Experiences. “I think we’re going to see more and more workflows happening in conversational interfaces, and in conjunction, we’re going to see advances in AI — but they’re not necessarily tied together.”

On the one side, you have the sexy AI-powered virtual assistant, what analysts are hailing as the future of human interaction with their digital devices. Armed with speech recognition and natural language understanding, personal assistants Siri, Cortana, Alexa, Google Assistant, Facebook M, and Samsung Viv are getting better every day at understanding user intent.

And while human-style conversation will likely remain the gold standard, that doesn’t mean that many of the entry-level questions and concerns can’t be handled by clever decision-tree software. In other words, you don’t have to whack a mosquito with an elephant gun.

“I think we’re going to see a lot of bots that realize simple workflows that make our life more productive that do not require AI,” Shevat says.

There’s two types of conversations, he says: Topical conversations and task-led conversations. Topical conversations require plenty of artificial intelligence capacity as well as natural language understanding.

But task-led conversations, even in our daily lives, don’t actually require a lot of intelligence. Often a task can be completed just by asking a few fixed questions. In fact, many conversations you have with a human service provider are simply scripted. And while artificial intelligence is becoming easier to leverage with new platforms and tools being released regularly, implementing it isn’t the daunting task it used to be, but it’s certainly a lot more complicated than setting up a decision tree­–style chat bot.

But that’s not the only consideration to make in the decision — or even the most important, says Shevat. Before anything, you need to choose a service or product that’s actually suited to a conversational user interface.

“Developers should start by pinpointing their products,” he explains. “A conversational interface is a user interface, and they’re only as good as the product or service they expose. Obviously exposing a great product or service means your chances of success are much greater.”

But it’s also about what makes sense. A digital agent chat bot could offer the best workflow if you’re booking flights for customers, for example, while service offerings might best be offered on your website footprint.

“In most cases you don’t need AI for the first instance of your bot,” Shevat adds. “Don’t start out using AI for the MVP — minimum viable product. As you see things progress, then you can add AI to deal with more complex use cases.”

And yet there’s still a lot of people flinging themselves on the AI wagon for the sake of saying they’re cutting edge — or perhaps, as Stewart Rogers, VentureBeat’s director of marketing technology has said, because they think they’ll get easier access to other peoples’ money if they use the letters “A” and “I” in that order. But none of it matters unless you get the other three letters sorted out, which are ROI.

“What we’ve seen is that bots are not uninstalled; bots are abandoned,” Shevat says. “Engagement here is very very important.”

Measuring that engagement means mapping out the conversational funnel, and measuring how many people have achieved a task successfully or gone through the entire process start to finish, and then optimize from there, removing steps wherever possible.

“You don’t necessarily want to have verbose bots,” he says. “You want to have a bot that finishes a task.”

But this conversation has only just begun: For more insight into the messaging platforms you should look at, crucial KPIs to achieve, and how to ensure natural language processing isn’t a human problem, don’t miss this VB Live event!

Register now for free!  

In this VB Live webinar, you’ll:

  • Understand the messaging platforms of the future
  • Learn which platforms people are using — and why
  • Measure the success of your chatbot through best-practice KPIs
  • Create personalized interaction between your organization and your customers


  • Amir Shevat, Director of Developer Relations, Slack and author of Designing Bots: Creating Conversational Experiences
  • Darrius Jones, Assistant VP, Enterprise Innovation, USAA Lab
  • Stewart Rogers, Director of Marketing Technology, VentureBeat
  • Rachael Brownell, Moderator, VentureBeat