As we get closer to the Holidays, customers get increasingly concerned that their orders are not going to be delivered on time. So the clever chaps at Best Buy are experimenting with an order status chatbot that will reduce the load on the call center by providing updates on customer orders. At 1080bots we take a good look at ecommerce chatbots to figure out the most helpful ecommerce conversational experiences.
The Best Buy Best Bot greeted us by name and let us know exactly what it did. The initiation sequence was short and informative. It also let us know it was a chatbot and not human. The opening question asked if we had an order, to which we replied negatively. The chatbot responded with an interesting tech tip. It then offered us a chance to track our orders, and to authenticate us it collected two pieces of information – a name and phone number.
The chatbot then asked us for an order number, and let us know where we’d find that number. We entered the correct number and it told us that our purchases were ready for pick up. That was a fast and effective query fulfilling the chatbot’s primary purpose.
Typing ‘HELP’ got us a Facebook link to a human. Given the Best Buy chatbot’s declared scope, this was the right thing to do. We repeated the order status inquiry, this time giving the wrong phone number or the wrong name, so unauthenticated we could not access any order details. Having a chatbot respect privacy is an important consideration in building confidence in the service.
In his post on Conversational Commerce, Jay Myers at Best Buy stresses the importance of effective connections to the back-end, sound authentication, and efficient messaging. We know that the forward thinking team at Best Buy Open has been solving these problems for some time.
Prompted by Jay mentioning “Shopping Helper” and “Context Machine” we decided to follow the backcountry trails. The chatbot understood our request for Store Directory.
We then tried to search for a product but as noted, this prototype was not hooked up to the Product Catalog. It quickly let us know that it was not what it specialized in. We threw some negative sentiments at the chatbot and it responded with an apology although it didn’t fully engage. It’s better not to engage with unhappy customers than via an automated medium. Finally, it also handled the random phrases and positive sentiments very well.
Best Buy is gathering great feedback from this prototype Order Status chatbot. We’re sure we’ll see a more complete and effective chatbot in the near future. Providing a little unsolicited commentary on this we’d like to see the following:
If you’re thinking of creating an ecommerce chatbot that deeply integrates into the back-end retail systems and provides a great service experience, then drop us a line at firstname.lastname@example.org