Good bot, bad bot
A paper exploring what makes a good bot opportunity today
Bot technology has been hyped up recently, we have all been talking about it and most importantly we all want to start delivering on it. Now, if we refer to the Gartner “Emerging technology hype curve”, under “Conversational User interface” we can see how this technology is today in the Innovation Trigger Phase, showing mainstream adoption within 5 to 10 years. Successful companies looking to innovate in this space will have to find a way to cut the curve, avoiding not only the peak of inflated expectation but most importantly the “Trough of Disillusionment”. In this phase, we see companies building solutions for technology’s sake and not aligning their expectations with the current capabilities that the technology offers.
In this expert blog post, we discuss what a good opportunity looks like for this technology in its current state of development. A practical scorecard to assess opportunities within your business moving forward is available upon request.
Before we go into the detail of what makes a good opportunity, let’s discuss what bots are and what they are not, and why we are talking about them. Why is bot technology currently being hyped?
Chatbots are not a new concept. Back in the 1960s, Joseph Weizenbaum, a Massachusetts Institute of Technology professor, developed the first ever chatbot, Eliza. This bot was designed to interact with users as a psychotherapist, using scripts to recognise patterns and generate responses, marking one of the first uses of artificial intelligence (AI) in the machine-human interaction scenario. So what has changed since then?
What we are seeing today is a proliferation of bot platforms allowing brands and products to engage with users in a completely new medium. Facebook, Google, Slack and Amazon have released their natural language processing platforms, allowing companies to develop their own experiences. Just like Apple allowed developers to build native apps over iOS back in 2008, opening a new delivery platform, the major players in bot technology have opened the floodgates, allowing us to reach audiences on a large scale. The table is now set to create a new ecosystem of developers, service providers and 3rd party services.
Back in 2008, the companies that implemented mobile technology successfully were the ones that were building value for users and not just tech for tech’s sake. What we need to focus on is the user and how we will make user’s life better through this medium.
It is important to also mention that mobile platforms and sandboxes (the SDK and rules for building these experiences) have changed enormously since 2008. We are expecting that the new natural language processing platforms will also change and adapt to the market moving forward. A successful implementation today is one more aligned to change than to last.
Finally, and before we come to our our point of view on conversational interfaces, we want to separate this from AI as there is a fair amount of confusion today on this. The terms ‘conversational interfaces’ and ‘bot technology’ do not refer necessarily to artificial intelligence. Conversational interfaces refer to the way the user is interacting, not the cognitive technology that sits on the back of that interface and reacts to the user’s requirements. In this way, conversational interfaces can be built over simple decision trees and simple skills without the use of AI.
What we are investigating in this blog post is what makes a good opportunity for bot implementation today. To do this, we have defined a series of variables that make a use case attractive to this type of interface. Analysing a particular opportunity over these variables will help us expose the opportunity’s value and alignment to the technology.
Sporadic engagement: diametrically opposed to mobile apps, where a good app can be seen as a toothbrush, something we use twice a day for a very particular defined objective, conversational interfaces are more attractive when the engagement is sporadic. They are a little bit like a Swiss Army knife sitting in the drawer of our side table. We use it twice a year and the use cases can vary. What we are looking to provide here is an umbrella of tools that can be fit for different purposes. This tool set will vary in time as our platform expands and contracts.
Task based: aligned with the “sporadic engagement”, users will engage with us to execute task-oriented use cases such as reservations, simple changes to orders, issue triage etc. Introducing use cases that are transactional in nature will make a good opportunity.
Mimic human interaction: bots are good when a human intention would be the immediate user reaction. Typical engagement is expressed above transactional customer service.
Process automation: moreover, and in support of the previous variable, having a clear decision tree or workflow that can be automated makes bots easier to implement and more valuable for the user. This means bots could be useful to deliver and cement new processes or engagement methods.
Broken experiences: given that bots can wait for us eternally; users can drop off and pick up the communication at any point of the workflow without having to explain what has happened before or what the outcome needs to be. If we compare this technology with human agents in chat situations, where the chats times out and the history is lost, bots make a great solution for broken experiences where users drop in and out.
The right audience: you are targeting a young inquisitive demographics that can be reached through a bot platform (Facebook Messenger, Amazon Alexa, Slack, etc.). Bot savvy demographics will change fast, so stay tuned to your users to know when appetites change.
Acquiring users is tough: when your user case does not allow you to acquire a user before they have a requirement, it makes it difficult to release particular tools. Here is where bots come into their own and they can wait in the shadows for the user to engage with them. A good example of this is an insurance claim, a customer complaint or issue resolution. People won’t download your issue resolution app when they initially get your product as they are not anticipating that event.
Aside from these variables that can help you identify a good case for a bot, there are some special key success factors that need to be taken into consideration.
The right platform: once the audience has been defined, it is vital to identify on what platforms they congregate and build over these. One of the key values of using an existing platform is reach, building a bot over your own website can be valuable, but providing this functionality over social media and other platforms like Amazon Alexa can provide a much smoother user experience.
The right personality: bots come with personalities; building good engagement is having a character, good scripting is key here. The combination of the decision tree and the personality of a bot will define the user experience. To show how important this is, Google on-boarded the story artists of ‘Brave’ and ‘Monster’s University’ to help develop the Google Home and Ok Google personalities.
Getting the word out there: when it comes to promoting your bot, conventional marketing can only get you so far. Bots are great at sitting in the shadows waiting for users to have a requirement. It will be key to communicate this new medium when users are looking for you, not before. Identifying and capturing key business moments is the best way of doing it.
Finally, we have prepared a scorecard to do a high level analysis on a use case for bots. If you are interested in obtaining this scorecard and learning more, please get in touch.
Mubaloo is a specialist mobile partner, consulting on business transformation through mobile and the delivery of bespoke, enterprise-grade applications.
Mubaloo has a wealth of cross-industry expertise, which comes from experience working across sectors developing innovative mobile strategies and intelligent mobile apps. Mubaloo helps to transform business processes, productivity and customer engagement, all through the use of mobile.