Chat-Bots will replace shops

Short idea on the future of chat-bots

The average shop website is not very smart. Based on user input it shows the user some output. True, if you have a genius algorithm, then you are able to create a very valuable output like google. But we are still interacting with a “site” on a screen, which is still a paper-like experience (since the beginning of WWW).

It is not really surprising, that chat bots pretty much copy that mechanism. New technic, old behavior. User input, website output. Wow. Difference is: The output channel is not a website it is a messaging site like facebook messenger or wechat.

But what the fuzz about Chatbots?

Let´s keep in mind what a shop user cares for:

• get the right thing
• get it in the right way
• get it right away

Chatbot Version 1 — dull bots

Thesis: chat bots fit in our day-to-day behavior because we more and more like to communicate via short text messaging platforms. But if you look at a simple Hotel booking experience, the input a chatbot needs to deliver a hotel booking is the same input a website needs. No time is saved doing hotel bookings via a chat bot version 1. Even worse, because as user you are limited in using filters and other options, so it will be even more time consuming (see also chatbots vs. google).

Look at the hotel booking flow of the bot platform Flow XO: no difference to a hotel booking website:

Chatbot Version 1a — FAQ Bot

A valuable iteration is to use the Bot as FAQ gateway on a website. Pre-defined flows in combination with the users website behavior can help to guide the user. It is a passive bot. Mostly activated through the user who is desperately searching for help.

Often this bots are connected to customer service. A lot of these requests need to be handled manually, because the focus of the bots is still to get the user through the shop NOT convert within the bot.

Advatange: user somehow feels personally engaged, because chatting with “someone”. But it is still a rulebase FAQ -filter to either get the user back in the shopping funnel or provide human help.

Chatbot Version 1b — Conversion Bot

Something I haven´t seen so far is a kind of Conversion Bot which guides you through the website. You do not have to follow the website flow. You can rely on the bots help in order to get what you need. The bot is a translator of the website syntax. Staying with the hotel example, the bot would realize based on the users website behavior, that he may look for a business hotel and suggest a predefined search instantly.

It is an active bot using the real-time input in order to present multiple outputs and present other, maybe better options of how to use the website.

Still it is a very limited input-output bot. But it could help to bridge the next evolutionary step.

We are still in the era of dull bots, which handle only very simple tasks.

• Chat bots help the user to get the right thing, but not to get it right away.
• Users do not get things in the right way, because they do not trust a stupid chat bot.

Chatbot version 2 — AI Bots

More interesting will be the next iteration of chat bots. They will know the user better than he knows himself. They will get the right thing fast. And users will trust the bot, because they simply understand the value of it.

One fundamental difference between a form field on the screen (“paper”) and a chatbot is, that a bot is always reacting and is not dependent on a certain output device. If the user does not do anything, the chat bot will react and ask questions. The bot is profiling the user.

i. Profiling

In order to get to know the user, bots will collect information about him. Be it social media open source information or by asking the user directly. Three questions and bots knows the user better than anyone else (e.g. see cambridge facebook like study). Without knowing the users website behavior bots predict exactly what a user needs. Besides collecting data the bot can be trained using psychographic models like the OCEAN model, which is pretty simple and already used to influence elections.

ii. Artificial intelligence

I am convinced that the input-output-paper-style of buying things via the internet will be disrupted. Artificial intelligence helps us to overcome this old communication style and build multiple real-time flows administrated by algorithms (example below).

• Bots learn fast through machine learning
• Bots will be able to connect with other bots (bot net) and help each other optimizing the output for the user
• Bots will be able to communicate with the user in his personal language, because they learned it. Someone speaks your language? Zhe user will trust those bots.

To frame it as an example for hotel business:

It is 2pm, busy working day. I receive a push message from my Business Bot.

Bot: “Hey Ulf, should I book a hotel for the upcoming meeting in Barcelona?”

Ulf: “one night. Not sure where to stay.”

Bot: “Don´t worry. Checked event & arrival location. I just booked three different hotels for you. Check your inbox. Talk to you tomorrow.”

After checking mails in the evening, I see the three Hotels booked for me. The right style, category and price. I reply the next day.

This really saves time and I trust this bot, because it knows me very well. I received the information asynchronous. I do not want to follow a synchron input-output flow. I will be able to decide when, how and if I want to reply to that.

Please comment, share and discuss.