Planet of the Bots

To understand the new paradigm that this announcement creates, imagine you are driving along when your car breaks down, and you must look for the closest mechanic approved by your insurance company. Today, your options are:

• Use a search engine to find the insurance company website, find the number for the call centre, press 2 followed by #, then 3, wait for a free representative and (do your best to…) explain where you are, wait to get an almost-relevant response by e-mail or phone, without being able to change your request once it’s made.
• Download the insurance company’s application, use the “Forgot Log-In?” feature, wait to receive your information by e-mail, click through to confirm your request, go back to the app, log in, and hope that you can search for approved mechanics in the area by GPS.

Tomorrow, it will be simpler. Just scan your insurance sticker, which will automatically open a Messenger conversation with the insurer’s chatbot. Send your location in one click. The chatbot will send back a list of approved mechanics closest to your location.

This will be simpler, faster, and more efficient for the user, and cheaper than an app or a call center for the insurer. Bots seem like an obvious solution. What’s more, conversations with bots will be recorded and consultable for analysis, with the user’s first and last name, which allows for new possibilities in client knowledge.

These bots will one day provide technical assistance and store location, distribute coupons, help customers in choosing the right product, and eventually manage payment (already the case for over 400 million Chinese WeChat users). The possibilities are endless for every brand across every industry. Incidentally, that’s where the challenge lies: though the first turnkey solutions are appearing (such as Salesforce, Liveperson, iAdvize, or Shopify), their ability to respond to diverse needs is likely to be discriminate, so as to not limit a chatbot to a simple order tracking system or a remote live chat.

From relationship marketing to conversational marketing

Another considerable challenge is finding the right proportion of automation to human interaction. Natural language processing algorithms are progressing with such speed that we might one day have 100% autonomous chatbots (Facebook’s M assistant is a type of precursor to this), but for now clients still need to be able to talk to another human to ask open-ended questions.

Though it is definitely too soon to measure the impact of this launch, we are willing to bet that 2016 will be remembered as the year when relationship marketing became truly conversational, and where contextual data meant brands could better serve their clients with automatised agents, if only in part. Watch this space – it’s sure to be exciting.

This column was originally published on April 19th 2016 on Journaldunet.com, and translated from the original French by Niamh Cloughley.