WeChat Bots

Considering the role bots play in the WeChat UX

Introduction

First, a helpful disclaimer. The notion of “WeChat bots” is a bit of a red herring. Yes, bots do exist on WeChat, just not that many of them, and they don’t have a very central role in the overall user experience.

WeChat started off with bots, but has since moved away from them in order to deliver a better user experience. In this article, we’ll discuss how bots are integrated into WeChat:

 

The misunderstanding

WeChat’s user experience in 2013 featured bots prominently.  Since then, WeChat has moved its UX toward web views and an “app-within-an-app” display with tabs and menus, rather than focusing on the development of “conversational commerce” with bots at the center.

This was surprising to me. In the United States, when bot mania really took hold of the technology world in early 2016, there was a notion that WeChat’s success, aided by bots, proved the viability of the bot technology model, with frequent discussion of which chat platform was going to become “the WeChat of the West”. This discussion seemed to imply that WeChat was overflowing with bots, transacting business, helping users, and generally pointing the way toward the future for messaging apps.

This misunderstanding has been pretty difficult to displace. Even in early 2017, major news outlets like USA Today were giving the impression that bots are central to the WeChat user experience:

In launching chat bots, U.S. companies took a page from huge social networks in Asia, where businesses have been able to insert themselves into the growing volume of conversation on texting platforms. On China’s WeChat, for instance, doctor’s appointments and shoe shopping can take place alongside conversations.

It turns out this is a misunderstanding, at least in relation to bots.

From bots to “app-within-an-app”

Bots do exist on WeChat, just not that many of them, and their use is limited:

There are very few bots in China. While there are over 10,000,000 official accounts for organizations in China (similar to a Facebook page, but only available on a mobile device), the vast majority are menu-driven, not bot-driven. Where bots do exist they tend to be very shallow and only react to very specific keywords.

In early 2016, Andreesen Horowitz had a podcast about bots with their China lead, Connie Chan, mobile lead Ben Evans, and Chris Messina. A central thread of the discussion was a tendency to conflate WeChat’s success with the coming dominance of chat-first interactions, including bots.  Connie had this to say (emphasis mine):

Facebook is starting to look like WeChat official accounts did back in 2013. When WeChat first launched Official Accounts, a lot of them were text-based, with very simple graphics. They didn’t have the menus on the bottom yet.

 

A lot of Official Accounts weren’t focused on the web view concept, and the “app-within-an-app”. But it quickly evolved into that, because the text-based chatbots just didn’t work out, and very few of them got much traction at all.

 

In China the app-within-an-app is much more focused on web views than bots… Messenger should be thought of more as a shortcut, a link to a web view that can give a much broader, richer experience.

It’s interesting to see that, in this context, WeChat started with bots, but evolved toward a model that (for their users) represents a better user experience.  The model they’ve arrived at happens to be more consistent with an app-like experience, rather than chatting with bots:

wechat user interface

Booking a doctor appointment via WeChat Wallet: image credit

What’s also interesting is that another major platform player, Facebook Messenger, seems to be following in their footsteps.  In March 2017, Messenger announced a major release with “Enhanced Menu Structure” functionality.  Essentially, this is a way for users to choose a menu-based user interface, rather than chatting with a bot:


In January 2016, Dan Grover, an American who moved to China to be a product manager at WeChat, wrote an excellent post about mobile UI trends in China. In 3,800 words, he only devoted 250 words to bots, summarizing by saying:

Though bots have been fashionable, I’m not sure there’s evidence that real users actually prefer interacting with apps this way.

I think Dan’s point is an excellent one. And I think it’s really interesting that two of the biggest messaging platforms on the planet are moving toward more app-like interactions.

It also suggests, to me, that all the “AI” hype around bots is massively overblown. Add to this the mis-branding of bots as “chatbots”, which sets bad user expectations, and you have the perfect recipe for a promising technology that can’t (currently) deliver the “chatty-AI-knows-exactly-what-I-need” user experience that users are expecting.

What are bots on WeChat doing today?

Today, bots in WeChat seem to follow one of a few models:

The traditional “chatbot” model

Thousands of men in China refer to Xiaobing as their “virtual girlfriend”, but it’s really just a chatbot developed by Microsoft. Xiaobing is a precursor to Microsoft’s Tay, minus the racism. Microsoft describes her as “an emotional companion”, and she’s good enough at carrying on conversations that the average user talks to her 60 times per month. This is another example of the canonical “chatbot” model, where there is no “point” to the service other than to carry on a conversation, and pass time.

xiaobing bot in wechat

Image credit

Some companies are looking to extend this chatbot functionality for more practical purposes, like learning English. However, even in the context of discussing how chatbots would be a good fit for this use case, the author notes that “when it comes to more commercial or service-oriented WeChat accounts, chatbots are rarely implemented.”

The “Siri” model

Chumen Wenwen is an app that lives inside WhatsApp in the same way that Siri lives in an iPhone, and has very similar capabilities. You can ask Chumen Wenwen to find restaurants, movie times, and whatnot.

Similarly, some developers are building their own “virtual assistants” that operate inside WeChat, but are intended to help with organizational duties:

Jiarui began developing WeChat bots about half a year ago to automate group management for her dance community on the messaging app. Her bot can automatically pull friends into groups and welcome new members, as well as hold basic conversations.

Automated “bot spam” model

Any time there’s a thriving online ecosystem, there will be actors who attempt to extract value from it through malicious automation, or a “bot spam” model. In the case of WeChat, there were bots posing as women, attempting to extract donations from men directly. There’s also the standard ad-spamming activity common across the internet:

But unsanctioned bots run amok on the social network, often tactlessly spamming groups or artificially inflating a brand’s likes and follows. The Chinese tech firm clamped down last year on thousands of accounts, denouncing the use of external software to alter WeChat. Tencent did not reply to queries about its policies on bots.

These artiifically inflated accounts are then used to extract advertising revenue:

“If I grow a WeChat account to 3,000 to 5,000 friends, one ad on my Moments could be worth between US$14 to US$140,” explains Li Jiarui, a Javascript and WeChat bot developer based in Beijing. “It depends on the quality of your account.”

So it seems like WeChat/Tencent see bots more as potentially malicious external software than they do as a central part of the technology or interaction model within the service.

In fact, if you look at WeChat’s Developer documentation, the word “bot” is only mentioned 1 time. It’s on the Acceptable Use Policies page, discussing how it’s not allowed for bots to log into Official Accounts in order to verify certificates.

Summary

So bots do exist in WeChat, just not that many of them. The ones that do exist tend to follow familiar patterns.

WeChat is a fascinating ecosystem. There’s a popular meme of “the WeChat of the West” which surfaces on a regular basis in technology circles in the US, always in the context of how massive WeChat is, how complete their market penetration, and how, therefore, any Western messaging app should aspire to assume their mantle:

WeChat is the dominant messaging force for monopolizing mobile users’ attention, via a platform that plugs in all sorts of additional third party services, from ride-hailing to banking to food ordering. (Gartner’s poll puts WeChat’s popularity there at a staggering 95 per cent usage). This so-called ‘super app’ model was pioneered by WeChat, but players such as Facebook are clearly gunning to clone the approach for mobile users in the West…

Can WeChat’s ‘super app’ model be replicated in the West? After all, China’s firewalled Internet and differences in cultural attitudes could also be playing a role in WeChat’s success. But Ekholm reckons it’s only a matter of time before the US has its own WeChat-esque go-to platform.

Is this a realistic expectation? I don’t know. My guess is that it’s unlikely, at best. Forces specific to a market are hard to generalize out to other geographies.

For instance, over 20% of WeChat users are also users of the WeChat Wallet feature, enabling all payments seamlessly inside the ecosystem. Hundreds of millions of Chinese consumers use the service for payments each month.

Also consider that many Chinese users only get 500MB of data per month across 2G infrastructure. That means stripped-down “app-within-an-app” data delivery works much faster trying to access bloated web pages.

Are those conditions ever going to be replicated in the U.S.? The market here is intentionally fractured.

People self-select into whatever platform they feel attracted to. It’s hard to get all your close friends on the same messaging platform, much less get a substantial number to use mobile wallet functionality that makes the messaging app useful as a hub for commerce. In the U.S., only 12 million people use Apple Pay, the most popular mobile payment solution.  Personally, I’d rather keep my money in a mattress than have a Facebook wallet, with them tracking every click I make and dollar I spend. And most data plans in the US *start* at 4X – 8X what an average Chinese user has access to.

Beyond that, there’s even an argument that WeChat is a less a perfectly conceived ecosystem, and more of a reaction to the failings of the major mobile companies to serve the needs of international users:

What WeChat has achieved can be much more instructively framed as an adept exploitation of Silicon Valley phone OS makers’ growing failure to fully serve users’ needs, particularly in other parts of the world. Chat apps have responded by evolving into “meta-platforms.” Many of the platform-like aspects they’ve taken on to plaster over gaps in the OS actually have little to do with the core chat functionality. Not only is “conversational UI” a red herring, but as we look more closely, we’ll even see places where conversational UI has breached its limits and broken down.

WeChat bots exist, but in a very different context from what we’re accustomed to in the United States. WeChat is certainly showing how a “super-app” messaging client can expand to take over much of a nation’s attention on mobile. But they’re not doing it with bots, and it probably won’t be something that can be replicated.

mautic is open source marketing automation