The Conversational Commerce Bill of Rights


Share on LinkedIn

There probably isn’t a more buzzy buzzword out there today than “Conversational Commerce.” Okay, maybe Pokemon Go, but after that, definitely conversational commerce.

The concept is nothing new. At its core, conversational commerce is the consumer asking for something, like a mocha frappucino at Starbucks. But now these requests are happening in messaging apps – Whatsapp, Messenger, Kik, Snapchat, etc. – between humans and bots designed to process orders.

That’s great for convenience, but as businesses owners we need to think carefully about the level of service provided within these transactions. How do we ensure that we’re respecting the rights of our customers?

Conversational commerce is still in it’s infancy, so basic tasks are functional, like ordering a taco from tacobot, or ordering a beer at a baseball game like Kik co-founder Ted Livingston suggested, or finding a bot geared toward information and entertainment in a bot shop.

But what about something a little more complex, like buying a pair of pants? In a recent experiment we sent a sent a reporter out to test the waters of conversational commerce and buy a pair of pants using only chatbots. His experience in Messenger broke down when he got to a question about inseams that reached the limits of natural language processing (NLP).

This would have been the perfect time for a human to swoop in and save the sale. Alas, it didn’t happen and Ted went away pantless. (I mean, he was still wearing pants, but he wasn’t quite able to buy new pants using chatbots.)

So what is the future of customer service in conversational commerce?

Blending automated and human
“Those who cannot remember the past are condemned to repeat it.” – George Santanya

At some point in the history of customer service, businesses realized that no interactive voice response (IVR) system could ever fully handle a customer’s needs. They were good at handling some requests, but required human intervention for others.

Why then do we think that a fully automated system on a mobile device will succeed where IVR failed?

The challenge of providing great service within conversational commerce isn’t technical – the challenge is finding the right blend of customer service technology and customer service organization to help customers where and when they need help.

Chatbots are a powerful new channel to increase sales, but they are also a minefield of mismatched expectations. To head off these mismatched expectations we at Olark created our customer service Bill of Rights for conversational commerce.

We hold these truths to be self-evident, all customers have:

  • The right to be heard – There should always be a way for the customer to ask a question when making a transaction. Don’t hide the ‘help’ button.
  • The right to be understood – There are limits to natural language processing. If a customer’s question reaches the NLP’s limited set of commands, they should be transferred to a human. Like saying, ‘operator’ in an interactive voice response (IVR) system.
  • The right to fast (near instant) responses – As Kik co-founder Ted Livingston says, ‘the future of chat isn’t AI, instant interaction is key.’ Customers expect communicating with a business to be like communicating with their friends: quick.
  • The right to clear expectations – Customers should clearly understand how long a response will take, not a range of time in which a business might respond. It should also be clear to a user whether they are speaking with a real human, or a chatbot. Don’t try to deceive the customer.
  • The right to be treated as a unique individual – Messaging apps might seem less personal than a phone call or an email, but your customers have intimate relationships over messaging. Customers don’t expect the same treatment from a brand, but they will want you to know their order history, understand their preferences, and flag the products they are interested. If an automated interaction fails, make sure your team is ready to pick up mid-conversation.

The jury is still out on the role that chatbots will play in conversational commerce. What we do know, with some level of certainty, is that consumers trying to place an order through a messaging app expect more than what can be done by chatbots alone.

As your business prepares to integrate with a messaging app, make sure you’re considering how you’ll support that channel, and whether you’re respecting the rights of your customers.

Ben Congleton
Ben Congleton is the CEO and co-founder of Olark, a leading live chat and messaging company.


  1. Hi Ben,

    while I agree that conversational commerce, or chatbots, or particularly the AI’s behind them, are not there where they could be really useful in fully autonomous sales- and especially also service scenarios – what makes you think that it will always be like this? Just the fact that IVR’s (another, even more complex conversational UI than a chatbot), doesn’t really prove the case here. it only states that technology is not yet advanced enough, imo.


  2. Hey Thomas,

    Good question. To eliminate the need for humans to be involved in customer service, we would first need to first develop Strong AI / Artificial General Intelligence. (

    That is, AI will need to be as good or better than a human handling not only normal customer service cases, but exceptional cases that require judgement, context, and decision making under uncertainty. I don’t presume that this is impossible, but I do presume that there are going to be a lot of chatbots deployed with Narrow/Weak AI ( in the meantime. The state of the art AI developed by IBM, Google, Microsoft, Facebook and others is only able to handle a narrow scope of situations.

    The rationale behind the conversational commerce bill of rights, is to guide chatbot developers for the foreseable future.

    There is a lot of uncertainty about when we will develop Strong AI, the only certainty is that no one knows for sure, and most experts predict it is going to be a while before we get there. The best source of general knowledge on the topic is a 2014 survey done by Müller and Bostrom who surveyed 550 experts in the field of AI. Roughly 50% of those surveyed predict we will have AI with 3rd grade human intelligence by 2030, and that passes the Turing test by 2040.( )

    Would love to hear your thoughts too,


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here