Conversational computing platforms and its origin
Although, till date, most of the computing platforms have owed their existence to the US, conversational computing platforms might be something that would owe its existence to the east and China in particular. It wouldn’t be incorrect to say that China’s Tencent pioneered the concept of conversational computing platform through its WeChat application. The Chinese application is a world in itself and allows you to chat, send money, book rides and do a whole lot of other things from the app itself. All these things are facilitated by official accounts that users can follow and interact with to be notified about new offers or perform various interactions with a particular company. WeChat has been highly lucrative for Tencent, the average revenue per user on WeChat is pegged around $7 and WeChat is pegged to have around 800 million users according to Statista. Looking at the kind of success Tencent received with WeChat, American tech companies wanted to implement something similar as well. However, their approach differs and conversational computing platforms aren’t necessarily restricted to messaging alone.
Types of Conversational Computing Platforms
The core
Google has two forms of conversational computing platforms which they are trying to promote. One is Google Home and the second one is Allo. I personally feel that Google should try and focus all its efforts on Google Home. I have tried Google Allo and honestly it’s not a bad chat app. It has a few cool tricks up its sleeve and Google assistant has been pretty good so far. But chat apps require network effects to be successful and Google Allo lacks that. Very few people would be interested in using Google Allo if their friends are not using it. Sure, there’s going to be some initial excitement surrounding Allo but ultimately everyone would want to go back to their regular chat apps. Also considering that almost every country now has a dominant chat app, there’s no clear path for Allo to dominate.
Also Read: The only way Google Allo can Succeed is if it’s Everywhere Although Allo lacks the network effects, Google has a pretty impressive lead over its competitors when it comes to AI. This was very visible during my use with Allo when the app was able to suggest smart replies even when the conversation was happening in Hindi. Google’s lead in machine learning and AI is hardly matched by any other company as of now. Combine this with the vast amounts of data Google has about you thanks to the numerous searches you have made, email exchanges on Gmail, usage of maps, data from Android etc and Google’s bot can be smarter than most others out there, at least on paper.
After Google launched Google Home, they also released an SDK. Developers can tap into the Assistant SDK to make Actions for Google Home. Actions are basically what skills are on Echo.
The biggest benefit Facebook has in this war is obviously its user base. Facebook has two chat apps with a user base of 1 billion and above namely Messenger and Whatsapp. Although Whatsapp was destined to hit the 1 billion mark, Facebook’s decision to unbundle messaging from the core Facebook app did help a lot in Messenger’s growth. Starting in April this year, Facebook completely opened the floodgates for bots on Messenger. Facebook eventually plans to make Whatsapp a platform. It’s very clear here the Facebook is trying to take a page out of WeChat’s book. Just like how WeChat allows its users to perform a host of functions from right within the app, Facebook wants to enable the same on Messenger/WhatsApp.
There was a lot of enthusiasm in the beginning but as it turns out, the enthusiasm on paper didn’t really pan out in real life. Bots were riddled with bugs and had received largely negative reviews from the public. Most of the times, the bots simply weren’t able to comprehend and execute even the simplest of commands. This was made apparent when at this year’s TechCrunch Disrupt, David Marcus, head of Messenger at Facebook himself admitted that bots weren’t what they were hyped to be. Although Marcus said that the lack of quality in bots was down to the low preparation times developers had, the damage had been done; Facebook has added payments and web views but how far extra time and these extra features help in improving bots remains to be seen. Some people feel text-based conversational computing platforms might never take off in the US like they did in China because of the vast cultural differences between the two countries. There are many reasons why WeChat became successful in China. First of all, the Chinese directly leap frogged to mobile payments. It is literally possible to leave your wallet at your house and go out and make all sorts of payments for all sorts of services through your smartphone alone. This kind of ubiquity of mobile payments in China provided a solid base for apps like WeChat to be an all in one hub for all your needs. Secondly, apps in China are judged depending on the number of functionality they can provide whereas apps in Western countries are judged depending on how well the app can do that one function that the app claims to do. There are quite a few other reasons as well but the end point that matters is that bots on Messenger have had a bad first impression.
Apple
If iMessage is to be converted into a conversational computing platform, it needs to go cross-platform. Android commands more than 80% market share of the global smartphone market. I know some would point that global market share isn’t accurate as Apple is heavily disadvantaged in emerging countries and I agree with that. But even in its home country of USA, Apple has less than 50% market share and the same is true for several other countries where Apple is the dominant manufacturer. The market share of Apple wouldn’t have been a problem but other chat apps exist as well, and leaving iMessage, all of them are cross-platform. It makes sense for developers to build bots for the chat app that’s supported by smartphone OSes rather than develop for iMessage and miss out on as much as half the market. There are two things Apple can do, one is to make iMessage cross-platform just like how Apple Music is available for Android or weave iMessage even deeper into iOS such that competing chat apps such as Facebook Messenger aren’t able to provide the same experience. But even if Apple manages to do either of the above, I still feel Apple might not be able to do well. The biggest problem, in my opinion, is privacy. For bots in chat apps to be extremely useful, they need to mine a lot of data from the end device. If access to this data isn’t allowed then the job of the bot becomes even more difficult. Apple’s answer to this is differential privacy but I wonder if it’s going to help. After all, Google had to back from its initial privacy claims of Allo in order to make sure that Google Assistant would perform at its best. I am pretty sure, the folks at Google would have tried to keep privacy intact if it didn’t have a hit on performance.
Conclusion
I would again reiterate that the future of conversational computing platform is far from certain. For all you know, they can be a fad that might go away or they can also be the next big computing platform. But whatever be the case, almost every major tech company is invested in this in some way or the other. It would be interesting to see how all of this plays out in the years to come.
