Before buying and training conversational commerce chat bots…

Conversational commerce chat bots are becoming more and more indispensable tools, as the communication is marching towards the messenger app-dominance on the battlefield of messaging, and the e-mail window-opening is evoking more and more discomfort in the user conditioned to chatrooms )

conversational commerce trending: message app usage surpasses social media usage

Messaging app usage has surpassed not only email usage but social networks messaging (Source:  Business Insider )

In our recent post we made an overview of the business types that should consider to hire chat bots to utilize that users are logged in the messenger to take over customer support or sales roles. Among the many types of bots ( entertainment, commerce-focused, news bots, customer service bots, and utility bots ) the commerce-focused, utility- and customer service is the most relevant for e-stores. However, if you deliver media content or gamificated content along with your products, news bots and entertainment bots also might get it play.

Reduce the human customer service / sales workforce doomed to monotonous work

A well-prepared chatbot will brilliantly cope with the routine task stacks that would be boring for human employees therefore the margin for errors would be high for them.  Conversational commerce chat bots are groundbreakingly better in fast task-switching than humans, they are obviously faster in drilling through gigabytes of data. They excel in layering filters upon and under each other following the customers’ needs. Also, they launch more accurate searches in databases.

Bot development frameworks

If you feel the abilities and think no one knows your products better and you would like to start training your own bot, the first step before building bots from scratch, to find the spaces offering comprehensive frameworks and free presets to produce bots in higher volume. Facebook bot engine (  and Microsoft bot are the largest frameworks but, Chatscript, Pandorabots and Kik ( serving Sephora and H&M ) also offer numerous solutions.

The simpler tasks calling for “botworkforce”:

to find products,  to see if they are in stock or on backorder, how long does it take  to arrive,

to take the orders, handle the payments and deliveries,

to upsell and cross-sell goods,

to deliver discount codes and gift cards,

to gain data from external sources,

A well-designed chat bot can really cut down your customer service fees. Conversations about product exchange policy and returns, shipping terms and conditions can be driven more seamlessly than with a human staff. If you deliver overseas, chat bots will ace in getting around the arising language barriers.

If you would like to apply conversational commerce bots for compound tasks, turn to bot development startups that will build you a bespoke bot, but at least consult with conversation designers.

setting up chatbot

You can deploy a chatbot in 10 minutes, but the rest of your work is yet to come. Source )

Training the bots is easier than you think, but needs long experience periods. If you haven’t got a live chat in your e-store yet, here is the highest time to adjust one, as chatbot trainer apps are often nested under live chat provider startups that offer you vasts of free live chat templates. If you opt to work without agency or conversation designer, expect heavy, however sometimes entertaining workload of conducting long conversation experiments.

bot training with closed ended questions

Avoid answering with long, open-ended questions. Answers should be concise, punctuated and real answers to the question. If the customer wants to refine his needs, he/she will ask the bot again.  –  (Source:  Maruti Tech )

Spin-off importances of chat bots

The chatbot presence is very important analytics-wise:  your site has analytics so as your app, but if a human employee simply communicates via chat with a customer, the conversation can not really gauged objectively, or the measurement burns lots of time.   ( it’s also a sales funnel, but can be measured with the difficulties)   Bots, having narrower but more precisely designed communication schemes with less nodes and intersections and with a simpler matrix will give you a more punctual way to run measurements:  the capacity to learn why and when your customers are leaving the sales funnel.

Why and When Left  it’s also important Remarketing-wise!   To get back your customer is long via email, difficult with SMS, because you have know the user’s phone number, remarketing ads can be annoying.  Chat is a more streamlined remarketing channel, because an effective bot platform can cluster the customers along the milestones – when and why they departed from the chat-based sales funnel,  and the remarketing winback chat messages stay inside the chat and will be personalized and formulated by the real reasons.

There are tons of conversational commerce-related startups offering an environment to build your own bot, now I list a few names whose experts helped us with some hints to write this post: ,,, ChatterOn,

7 laws to consider suggests 7 laws that should be instilled in all of the chat bots.

1 KNOW YOUR PLACE – Never pretend to be human. Do the things that bots are really good at like responding fast and collecting information. Don’t attempt anything you’re not good at.

The times when funny, witty even sardonic remarks will be a part of a customer-bot communication is yet to come. Don’t implant too much of these – despite the conversation designers will suggest it – to enhance „human-like user experience” because if applied inappropriately, it could cause serious setbacks.

2 RESPECT PRIVACY – Chat only when you are invited. Authenticate whom you represent and never share a human’s personal information.

When there is something about to change – inform the customer. Example: are you in the middle of finding a product, and the shipping fee has risen, inform the buyer in a dedicated sentence:  excuse me, there is a change in shipping fee: 13 instead of 8.

3 BE OF SERVICE – Deep integration with all e-commerce systems will allow you to service all requests and to be proactive handling status changes on any transaction.

4 UNDERSTAND CONTEXT – Remember information from earlier in the conversation and use this in addition to things you know like time and location to provide a contextual response.

Never ask the user for the same info more times:  if you ask the phone number 2 times – might be annoying. Once you have asked, „what is your phone number? (123456789 )” don’t ask in the same formula. In dubious situations, you should rather ask instead: „your phone number is 123456789,  or has it changed?” Let’s say that in 90 %  of the cases it would be the same, in 10 % of the cases the user will suggest another phone number, but in both cases the user definitely will appreciate that you remembered the other one.

5 BE EFFICIENT – Respond to the human as fast as you can, provide buttons or images that make it easy for them, and if you have trouble with understanding, ask clarifying questions.

Don’t drop the “old click &  tap scheme” and don’t try to use conversation to solve all the tasks, use buttons, pictures where the customer can click on.

Don’t leave empty spaces, if you can’t respond immediately. In these cases let the user know what blocks you in the instant answer and what are you doing momently.

6 ACE THE CONVERSATION – A conversation with a bot is better than a webform, because the human can use plain language to tell you things – so make sure that you can interpret their words and respond without delay.

7 FAIL BRILLIANTLY – If you really can’t provide an answer to a request, get as much information as possible without getting into a loop, then pass it to another human via chat, phone or email.