Cognitive AI Chatbots are enabling communication (and e-commerce) to revert back to its roots: Conversations

October 22, 2020 • 5 minute read

In its 5 million years of existence on Earth, humankind has radically transformed, at an ever-increasing pace. It is generally acknowledged that we have transformed our lives more in the last 100 years than we have over all those millions of years ago. 

One of the hallmarks of this transformation is communication, from disorganised grunts and cave paintings all the way to today's latest social media applications and augmented or virtual reality events and media.

Communication being the foundation of the social species that we are, it naturally became the foundation of trade, as trade can’t happen without an agreement that an exchange is mutually beneficial. Trade itself transformed radically from exchanging objects and services, where one person would only exchange with another if they had exactly what they were ‘on the market’ to have an intermediary system, called money, that propelled commerce to the next level by erasing the reciprocity requirement. At that time, the conversation at the stand then became a question of negotiating how much of that currency would one part with for the desired product, in other words: haggling. 

Fast forward a few hundred years further, and things started taking the shape they have now, where the increased economic movement and population coupled with ever-growing companies led to standardisation. One does not haggle with a big corporate and thus; the conversation was relegated to pleasantry and potentially awkward exchanges, as conversations lost the power to influence how much money they had to part with.

The way commerce works today is fundamentally against human nature

Then came what we could label the fourth shift in commerce, where the conversation essentially disappeared with the introduction of new intermediaries such as catalogs, shopping channels, and eCommerce portals. In this fourth stage of commerce, the conversation was relegated to exceptions, only coming into the picture if something was not right (delays, returns, or complaints). The conversation became labeled with negative outcomes.

We argue that while conversations have been relegated in many industries to exceptions for the sake of scale, consistency, and efficiency, this is essentially trying to shoe-horn fundamentally social creatures away from the social aspect.

 “It was so enjoyable filling this form to say my shoes were the wrong size. So many drops-down boxes and radio buttons, I had such a blast” said no one ever. 

Consumers still can have a conversation if they so wish, but it usually involves a rather lengthy set of on-hold music while agents are serving other customers (a scenario made far worse by the pandemic the world is experiencing), and not every organization will embrace their phone lines as a sales channel as it typically represents the most expensive channel in their support strategy.

In other words, conversations have lost the prestige they once had, the influence and power they had, and business became far more transactional in nature, especially if compared side by side with the first two iterations of commerce, where negotiating and finding a mutual ground was at the pinnacle of it all.

Thankfully, there is a better way & consumers want to experience it! 
We firmly believe that it is time to go full circle and go as close to the original roots as feasible without having businesses compromise on their top or bottom line. We know that going back to the roots and adapting to how we function as human beings are not only more engaging and deliver better experiences; it has genuine ROI as it reduces costs and improves business performance.


The way this can be done is through a conversational interface, either in a chat window, on your favourite messaging application, or through the phone. 

We call this conversational commerce and we hail this as the future of e-commerce, where a business is powered by conversations rather than forms and checkout processes across several pages.

Xiaomi Conversational Ad

In a nutshell,

  1. A user lands on any of your websites, your mobile app, your WhatsApp, your Facebook Messenger, or your web app. 
  2. They write, in their own words, what they are looking for
  3. The Cognitive AI assistant absorbs the information, extract the requirements from the user and revert back with either a selection that matches the criteria or follow-up questions to qualify what the person is looking for
  4. The user and the AI assistant continue the conversation until the user lands on what they want with the help of the assistant who answers queries they have and helps them refine the search
  5. With the choice made, the assistant offers products that normally go well together or people tend to buy in tandem, offer some type of promotion, or maybe even some upsell to convert a higher average shopping basket.
  6. The shopping assistant takes on the required details (shipping address, shipping method, preferred payment method)
  7. The assistant takes on the payment & confirms the transaction
  8. The user, delighted, purchased exactly what they were looking for and potentially a little more.

From a technical perspective, what happened is that we layered that agent on top of the existing infrastructure, very much like a live chat agent would do. They have access to all the same information to aid the customer. 

Evidently, this is where the comparison stops as the virtual assistant has a virtually unlimited bandwidth for concurrent conversations, is much faster at reviewing the inventory and items specifications against client requirements, and before long, would sit atop millions of similar conversations to inform its judgment and decision making in the interaction. 

Yes, it is a machine and no it cannot capture and correctly react to every single thing that the people can throw at them, so unless your conversational designer trained it to have a defined preferred music genre or a favourite movie, it won’t go in that small talk avenue, unlike a human. 

It does not mean the conversation cannot be humanised, on the contrary. The virtual agent can still assume a persona and be faithful to a brand’s tone of voice and add as much personality as the brand ought to have.

One easy way to sort the wheat from the chaff in terms of humanising a conversation with an AI chatbot?

The Swear Test. We find that users like to test the patience of bots and try to trick them and one of their favourites, and we must acknowledge, probably the most enjoyable way of doing so is by swearing at it. Very few humans would be on the receiving end of abuse and not at least acknowledge this as part of the context of the conversation.

Which means that, if it doesn’t understand the user, thinks it is the user’s name, or continues as if that was a valid answer/input, you know you are dealing with a technology that lacks the sophistication and maturity to genuinely deliver humanised conversational commerce. 

The chatbot or virtual assistant should be able to answer threats and abuse however you’d like your brand seen reacting to such behaviour. For example, it could answer with just a single image and leave it at that (worth noting that none of our clients has gone for this yet), like the below.

chabot meme

The 5th shift in communication brings the biggest benefits to date :
As we move towards a conversational economy, businesses will see dramatic results. Indeed, there are several reasons to rejoice at this new approach for both the customer and organisations.
Starting with the customer who can finally go back to a more unstructured approach, as they can say, in their own words, in their favorite channel, what they are looking for, and let the seller do the work of interpreting, refining, and short-listing suitable products. It will also benefit from the rich insight gathered through technology and data, as they are offered the most suitable complement to that purchase or enticing promotions and upgrades. The entire customer experience will just become far more natural, seamless, and all-around enjoyable. 

And from the perspective of the business, the cost to serve that individual customer will decrease further again, the conversion rate will shoot up. Exception handling, when required, will be focused solely on the more challenging and intellectually stimulating scenarios rather than routine ‘wrong size’ type of queries, amping up the engagement of employees and further boosting customer experience through heroic service recovery.

To illustrate this point, one of our clients has not only saved millions in operating costs (which was the reason to engage in the program all along), they successfully generated on top of this several millions of dollars in added revenue that can be traced back directly to the chatbot - needless to say, the ROI figures for this specific client are nothing short of awe-inducing. 

Are you ready to experience the power of omnichannel conversational commerce and reap the benefits it brings to innovative organisations? Wait no more, get in touch with our team… before your competition does!

Drop a mail @ [email protected]

Gustavo Imhof

Gustavo is Vice-President at Yellow Messenger with responsibility for transforming enterprises into success stories across Europe and Latin America. Previously to Yellow Messenger, he was named a 30 Under 30 in Customer Experience by the prestigious CX Network and was a sought after speaker and a published CX expert having worked in Financial Services, Logistics, Consumer Electronics and market research

Read more

Subscribe to Newsletter

Stay up-to-date on what’s happening at

Latest Insights

Build your first NLP powered intelligent chatbot in under 10 clicks