Is Conversational Commerce Really a Thing?

Lately, it seems like all of my meetings have been about bots, AI, ML, NLP and beyond. Conversational commerce is the buzzword of the moment, no question. And yet, Microsoft tried this UX years ago with the much maligned Clippy. So why conversational commerce now? And will entirely new companies really be built within these new messaging eco-systems?

My short answer is yes?—?but not universally across platforms, existing companies and start-ups. As with all new trends, each must figure out whether there’s truly a consumer value-add fit.

Why Bots? Why now?

Bots may be the latest and greatest iteration, but conversational UX and messaging as a platform have been around forever.

Unlike the touchscreen smart phone which changed user behavior overnight, conversational UX has been around for many years, assisting and training us for today’s commercial Chatbot moment. When I started as an analyst at Morgan Stanley in 1997, we had K1s and research reports printed centrally and delivered by courier. We also used Lotus Notes. But even back in those stone ages, we chatted on our PCs. The younger generations communicated regularly internally through both a propriety chat client (a rudimentary 20 year old Slack) and an alpha numeric pager.

That said, mass market consumer messaging is relatively new. I started dating my husband in 2004 and we managed to court and marry over 2 1/2 years without sending a single text. Today I communicate with my husband, my kids, my partners and entrepreneurs daily across imessage, messenger and snapchat with words, photos, snaps, videos and bitmojis. The UX is widely utilized and understood across generations and importantly it is an acceptable mode of communicating not just with your romantic partner or friend at work, but with professional relationships and commercial entities outside your immediate intimate circle.

So I contend that it’s this behavior shift, along with smarter and more organized AI, which makes bots a viable part of the commercial customer experience.

But conversational UX is a feature, not a product, unless you aspire to build a completely new platform, which will be done by very few. It is a feature that could catapult a limited number of focused companies with very specific assets - most likely those that have a marketing advantage in the real world and are able to profitably deliver against a use case.

How Bots Will Become a Business

Bots are no different than any consumer technology?—?to become a business, consumers must discover a bot, then have a positive experience such that they will use it again. Like any new technology, it needs to be better or easier than a user’s current options. The sweet spot for bot adoption is experiences that make it easier for end users to learn how to use a chat interface vs. a website or app.

At the core, this means ensuring that initial user expectations can be met by the AI. This is typically a narrow use case that is well understood by the end user up front and can be handled reasonably well in the first couple of attempts, by a smart computer or a human that will quick build content and hand off to a smart computer.

For example, when I use Operator within Facebook’s mobile messenger and am confronted with a stark white text box, Operator gives me broad shopping suggestions (this week’s electronic picks or mother’s day gifts) but doesn’t know what to do when I ask for recommendations for a 6 year old birthday gift. The results are a generic, frustrating way to browse. My somewhat narrow use case hasn’t been met and I move back to my Amazon prime comfort zone. Website?—?1. Bots?—?0.

And as we learn to communicate with bots, they will also need to learn how to communicate better with humans. Perhaps because we have been trained to chat with the ultimate intelligence, a human, we are trained to talk in “stranger syntax” (e.g., the way you might ask a question on Quora or text the parents of a child’s friend to coordinate a playdate). Our syntax varies widely depending on the “listener”?—?who they are and how we’ve communicated in the past. With new chatbots, in particular around new brands, neither we nor the bot have context, and so we are likely to be confused and disappointed if we are not trained. Getting the intent and the UX around onboarding right, and explaining how to use chatbots well, is actually much harder than sending a consumer to a mobile webpage with navigation terms like “about us,” “how it works,” “FAQs,” “our brand promise,” etc. We also may be hesitant to share payment information in a new way.

So until bots can be more like humans in our communication and relevance, the business will have a long way to go.

The Early Platform Winners

But of course, the adoption and communication syntax will no doubt evolve into an intuitive user experience for the use cases that matter most. S0 really, the platforms that will win are those who can tie great user experiences to making money for those providing it.

Benedict Evans breaks downs the players with natural assets here?—?those that own the client/terminal (e.g., Apple) and the back-end identity (e.g, Facebook). It is also critical to think about who has payment information and a tendency to manage an app eco-system with a heavy hand (Apple again, anyone?).

Of course, we can look at Kik, Line and Wechat for robust international experiences. While companies like Wechat have built a thriving chatbot economy, QR code scanning behavior and humans manning chatbots are also common. Cory Weinberg envisions a world in which Facebook both has payment infrastructure and in store reps using bots. This would make so much sense?—?and no better player to drive valuable usage than the company that got the world to make stickers saying “like us on Facebook” in the past.

U.S behavior does not always follow Asia in lockstep. Many waited for QR codes to become mainstream and it never happened. In the U.S., I think our best analog is customer service, and here I see the bot opportunity as Twitter’s to lose. Who hasn’t tweeted or DM’d brands with customer service complaints? I also think Snapchat has an interesting opportunity with the under-30 crowd. Younger generations are already chatting and interacting with brands in native Snapchat language.

In-store, human customer service has even more opportunity to leverage the technology as well. The next time I’m in Nordstrom, if the stylist tells me that I can reach her via the Facebook chatbot when I need shoes for an event, you can bet I am likely to use it since there’s a human face on top of an efficient channel.

The early, easy opportunity is no doubt for existing, larger companies to augment their customer service capabilities with high-performing bot communication. But as a VC, I am even more interested in the companies who are building to the feature intelligently as part of a broader mission.

The Companies that Will Emerge 

There are some interesting places where I see Conversational UX potentially catapulting existing companies beyond traditional customer service:

(1) “In Transition” Populations: Think about populations who are new to a life-stage and are looking for direction from someone they trust. Let’s take new parents?—?an audience that is overwhelmed and looking for information could easily be acquired in a bot setting by interacting with credible, influential sources. For example, if my pediatrician would let me text questions about nursing and sleeping to an integrated bot, you bet I will try it. Another interesting example is LA-based UR Welcome, who has built an onboarding chat-interface service for immigrants, introduced through consulates around the country.

These kinds of life stages and populations lend themselves to a universal chat interface and a common set of questions. These are also the moments in time when new brand loyalty is built. Operator has to win me away from Amazon, but “change of life” moments have to win me away from no one.

(2) Real World Venues: Consumers may also adopt conversational commerce for a specific and timely need in a real-world setting. Stadiums, for example, can easily teach their customers to order a beer without downloading an app as described by kik CEO Ted Livingston here. Gyms and hospitals are other obvious examples. I can easily see both the “I need a beer” user and the “I am becoming a Crossfitter” user onboarding to a chatbot, as both are driven by real world, immediate marketing.

(3) Enterprise: I would look at all the companies that are improving existing enterprise workforces (task management, HR forms/training, employee perks programs). To use one of the more entertaining examples, the Slack taco bell bot should just sit in front of Uber Eats, Amazon Prime or Seamless. Or think about companies that are pulling new data out of these conversations and making historically painful tasks easier and richer (e.g., 360 performance reviews).

(4) Industry specific anything: My guess is the most interesting applications emerge within verticals like health care, logistics, or manufacturing, and begin with new/confused users instructed to use the bot by an influencer (e.g., your boss or your doctor). For example, I’d love to have my reminder text from doctor’s offices also include a link to forms, or even an ID that accesses my common form info from a cloud database. It may need to be delivered first by a human, but over time these applications can be machine automated. The ROI should be strong for any office whose front end staff is slowed by paper work delays and poor information.

With all of the above, new brands must beware of platform risk and ensure that they own the customer. I look at companies like @service and remedy who are using Twitter and text chatbots to solve customer service and medical billing challenges respectively. Presumably both those companies are building up a smart back end with high quality content, workflows and API connections that can easily port into a messaging environment like Facebook in order to accelerate their brands and onboard new customers.

So yes?—?the bots are coming. But I see it as a step-function, not a revolution yet. This likely starts with companies that have real world assets to acquire customers and a human workforce who are already chatting in different forms that can educate the customers. And as always, it’s in the execution?—?if we overwhelm consumers with crappy chatting commercial experiences, the medium may be ruined for all and we will all move to a new clean communications channel. Perhaps brain to brain chat…