Be inspired by the best bot examples for conversational commerce, including Dyson, Gillette, Black Rifle Coffee and DAVIDsTEA
Say goodbye to dumb bots. There is no place to see this more clearly than with leading ecommerce businesses that are pulling out all the stops to create engaging and assistive conversational experiences for their customers.
TL;DR Competing Conversations Video
Check out our wrap-up of the 8 best ecommerce bots in the video below
If you are looking at designing your first ecommerce bot or giving your existing bot an upgrade, here you will find inspiration from a variety of companies trying different things in a space that still leads by experimentation and iteration.
At Meya, we want to democratize conversation design and UX by providing the resources and inspiration to embark on a bot project with confidence. Check out How to design an ecommerce bot.
Presented in no particular order...
When it comes to visually-engaging, assistive and brand-aware digital assistants, Gil leaves no stone unturned. Aesthetically, the bot is a seamless addition to the Gillette website, showcasing a consistent style and voice. Functionally the bot offers a variety of easy-to-find options including Account information, order status, product spotlights and FAQs.
Engaging Conversational UI - From matching brand colours, to the use of various content types including GIFs, cards, and buttons.
Mobile-view - full disclosure, it was difficult to find something that Gil did wrong. If there is anything that could be improved, we did notice some minor formatting problems on the mobile version of the bot that take away some of the finesse in the bot’s otherwise spotless execution.
Say hello to Mia, Mobilia’s digital assistant. As a FAQ-driven experience with a more form-based UX, Mia offers a lot when it comes to organizing questions into topics and guiding the user around . One other notable feature is Product Search which allows users to type in a product category like “Office Chairs” and be shown a number of in-stock options for the keyword.
Multi-channel support - Mia offers both FB Messenger and web chat to their visitors, offering customer service where customers spend most of their time. With more ecommerce businesses leveraging social media for distribution, having automation expand across channels increases accessibility and customer satisfaction.
Order assistance - Mia does not offer the user any option to view current orders or account information directly in the chat window, limiting the amount of functional context the bot can perform
Tile’s Support bot, Veronica, can only be found on the site’s product support page. It is clear that Veronica’s mandate is centered around the technical setup and troubleshooting of your Tile device, more so than product discovery or sales.
Topic-driven flows - With an abundance of FAQs and associated content, it can be difficult to structure a bot well in a button-driven UX. Tile does a great job of parsing the high-level topics of their content and creating branches that help customers find the answers they need.
No Knowledge Base integration - It feels like a missed opportunity that there is no obvious integration with the knowledge base articles available on Tile’s site. These articles tend to have more detail, including images and videos, which would give customers another option should the bot’s text-only instructions be unclear.
The team at Black Rifle Coffee have designed a functional and brand-aware bot around sales and subscriptions. It isn’t clear on first interaction but there is quite a substantial amount of FAQ content behind the curtain if you ask the bot questions directly, however, it seems that this is secondary.
Account modifications - Black Rifle Coffee’s bot supports a number of subscription modifications directly within the chat window including payments, delivery addresses and products. This serves at making the bot a one-stop shop for all account needs.
Coffee Quiz support - One of the main sales-driven tactics of the bot is to get users to complete a coffee quiz to determine which blend best suits their taste. However, this quiz is served on a separate page on the site instead of within the bot itself. Having the quiz embedded in the bot opens up opportunities for more personalized content.
DAVI, the DAVIDsTEA bot, is playful and chock-full of helpful questionnaires and FAQ content to help users with product discovery and support. With supporting both English and French languages, DAVI is a great example of how to build smart conversational UX for big libraries of helpful content.
Multi-Language support - Operating in Canada, DAVIDsTEA has made sure that their bot supports both the English and French languages. When serving markets where a large segment may not have English as their first language, ensure that your bot is responsive to the most common languages and more accessible for a broader audience.
Text-only - For a product with engaging visuals, animations and videos, DAVI neglects to take advantage of this supportive content. For many ecommerce businesses, it is difficult to inspire, convince or assist customers in an entirely text-driven UX so it is recommended that your bot mixes it with different content types to improve engagement and comprehension.
The Dyson bot is a great example of giving exceptional service for a brand with a large variety of product categories. Leveraging live chat, serial number identification across product categories/types and advanced troubleshooting, the bot can quickly route customers to what they are looking for.
Live Chat - Leveraging a live chat feature gives ecommerce business a new mode of communicating with customers, the main benefit being real-time messaging right when the customer needs it. Supporting live chat experiences can be time-consuming so it is recommended that only a subset of visitors get access to this service, such as existing customers, VIPs or those looking to make a purchase immediately.
Live Chat only - With the benefits of live chat listed above, relying on live chat only for certain use cases limits a businesses ability to scale and offer consistent customer service. In the case of purchasing a Dyson, it would be recommended that talking to a agent is an option but a bot-driven UX is the default.
Not all bots need to be big showcases of visually-striking content, advanced FAQ branching and account support. Db’s bot, Astrid, keeps things simple and functional, relying on ticket creation and feedback mechanics in case not all questions are answered.
Ticket Creation - Not all questions can be answered by the bot, in which case Astrid allows the creation of a ticket for agent follow-up. When allowing ticket creation directly from the bot, we recommend presenting a small set of triage questions that need to be answered before ticket creation so your agents can solve inquiries immediately.
Error prompts - When checking your order status, an incorrect order number will bring you back to the beginning of the UX with no message on what went wrong. This gives the impression that the bot is broken where it may just be the case that the order number could not be found.
Inside Weather has opted for a natural language driven UX, where the initial interaction is only triggered by a user’s question. This approach may encourage users to ask more questions of the bot, instead of relying on buttons, as this is the only mode available for interaction.
Account verification - Inside Weather is a great example of automatic account authentication for signed-in users so you can provide a personalized experience immediately. Most bots tend to be siloed citizens on a brand’s site so that no context is being passed between a user’s activities and the bot. However, with a deeply integrated experience you can create a seamless bi-directional transfer of data.
User onboarding - Unlike some of the other bots we’ve looked at, Inside Weather has chosen not to share much information about how the bot can help. Asking the bot questions in natural language shows an expansive library of content so it would be advantageous to showcase some of the FAQs and topics on initial load.