Facebook and others: Lots of Bots

In April this year, at Facebook’s f8 developer conference the company launched its AI bot framework for Facebook Messenger and the Spring API for interacting with its bots. In doing so, it is looking to turn Messenger into a platform where users can interact with companies in order to find and purchase products, track deliveries and make bookings in restaurants and cinemas, among other applications.

Facebook is one of a number of companies that have similar aspirations. For example, Tencent’s WeChat was the original messaging-as-a-platform service and has a broad range of functions that enable users to organise or buy through an app. Amazon has also ventured into this field, using voice commands to its AI assistant, Alexa, on the Amazon Echo smart speaker. Kik messenger too launched its bot store earlier in the year, as well as Slack which has used Slackbot within its enterprise messaging service while also integrating with a number of other services. In addition, Microsoft announced at its Build developer conference in April that it will use its AI technology to power 'conversation-as-a-service' and in May, Google announced Google Assistant, powering both its new Google Home smart speaker and its new messaging and communications apps.

So what will these technologies allow us to do?

For all of these players in varying degrees, the vision is that users are starting to move on from the world that the iPhone introduced, in which we have a single app per company or per brand, and into a new era. Here, we will interact with companies, their services and websites in natural language using voice and messages across a variety of devices without having to download separate apps, setup an account and then login. Instead we will operate through voice commands or a messaging app that we are already using a lot of the time.

In theory this method should help make the interactions with companies more efficient, as routine queries are dealt with through AI-driven bots rather than using human interaction. It should also bring other benefits from being less dependent on any single device or operating system. This of course is the very reason that companies such as Facebook and Microsoft are seeking to boost the standing of their respective messaging platforms in an iOS and Android dominated world. The 'OS war' is done. We are now entering the next phase.

However, there are also plenty of challenges to overcome to enable these new technologies to flourish. Firstly, just as there are plenty of badly set up interactive voice response (IVR) systems and call centres, we should also expect to see badly set up bots and they will be just as frustrating for users to deal with. Secondly, we expect more services to be set up to use bots, for example, the likes of Apple services could be powered by Siri-bots, Whatsapp and Instagram have not yet been given the bot treatment. Other search companies such as Yandex and Baidu will not want to be left out. And thirdly, we may see the arrival of a number of bot stores to enable users to find and use bots more easily.

More competition in the market

Brands and website owners will need to work out which of the bot options they will and won’t support – just as they had to with apps and app stores – and many of the same issues will apply. These organisations are unlikely to have the time, energy, staff or budget to support them all so we can expect winners and losers in the field. Similarly with app development today, we expect to see cross-platform software products that enable website owners to do one development project and then work with a variety of bot frameworks.

Going forward, it is crucial that users understand that these bots will automatically use their identity on the relevant messaging system or platform that they are using, for example their Facebook or Amazon login. To protect user privacy, controls will need to be provided to manage the information that flows from the use and application of bots. Security with bots is also going to become a major issue. Therefore regulators will have to pay close attention to their growth in the market and adapt their regulatory frameworks to cope with them.

The long and short term

Having said this, we do not expect the arrival of bots in the market to lead to a rapid transformation of how we interact with the companies we deal with and we are not expecting the AI element of bots to cause large numbers of redundancies among call centre and other support staff, at least not in the short term. Instead, we expect companies to have bots as a supplementary option, which may result in the need for more staff at the outset. Companies will take a number of years to fully reap the benefits of applying bots as part of their interactions with customers and along the way there will be lots of different approaches tried, lots of blind alleys, plenty of mistakes, a high number of bad bots and some very angry customers.

In the long term we do expect AI to have a profound effect in many areas and routine interactions with our suppliers really should become much easier than they are now. In the meantime however, there will still be a need for customers to be able to interact with a human who is providing sales or support for the large number of more difficult questions, highlighting that - although we are moving in the right direction - the underlying human interaction is still vital as these technologies become more advanced.

It is therefore vital that organisations embrace the arrival of AI. Rather than shy away from its potential, companies will have to ensure their business models are able to adapt and evolve to ensure they remain at the forefront of the ever-increasing digital economy.

Martin Garner, Senior Vice President, Internet at CCS Insight