If there’s one thing you can say about chatbots, it’s that there’s a lot of buzz surrounding the phenomenon.
There’s an excitement in the air that hasn’t been around for some time. The pending bot revolution also comes at the right time. Even with the success of Pokémon Go in mind, there seems to be a case of app fatigue. The cure just might be bots. But who knows? I don’t, because I cannot predict the future. However, I can look into the past. To a time before BlackBerry and Nokia were household names, in 1994. That was the year when professor James Utterback of MIT released his book Mastering the Dynamics of Innovation.
By looking at innovations in the fields of typewriters, ice harvesting, lighting, and photography, he formed an innovation model for assembled products. The model looks at transitions from one technique to the next. His model is as true today as it was back then. So let’s use it to predict the future of modern-day assembly.
Dynamics of Innovation model
Utterback’s model describes the processes within an industry during the total development cycle. The Fluid Phase consists of a great deal of change in which the outcomes are highly uncertain. Functional product performance is the basis for competition during this phase. Brand names are of little importance, and the number of competitors is small.
We enter the Transitional Phase if a market keeps growing to the point where the product innovation is accepted by the general market. A dominant design will emerge. As needs of specific users become clear, companies that focus on producing products to meet those needs will have the competitive advantage during this time. Companies will move away from product innovations and move towards a focus on processes.
The move towards process innovations eventually leads to producing a very specific product at a high level of efficiency. During this Specific Phase, the ratio of quality to cost becomes the basis of competition. The differences among available products are fewer than the similarities. Production is set in stone, and even small changes in either product or process are difficult and expensive.
The market during the phases
As we transition from the Fluid to the Specific Phase, the market and its companies will continually change as well.
The Product will at first see lots of variety. Some bots respond when you say, “tell me a joke”; others will answer you only with a blank screen. Some interact with buttons, others with text-based interaction. If this confuses you, you can always ask “help” or “menu” — to find out only some bots will actually open up a menu or help screen. Even the rejection of the rest will surprise you: from a blank screen to the shrugging of shoulders, with emojis to ask if you could repeat the question. This will keep happening until someone finds a mode of interaction that works. That mode will become the de facto way of talking to a bot, thus forming a dominant design. From that moment on, products will become standardized and innovation will only be incremental.
Manufacturing moves from a heavy reliance on skilled labor and general-purpose software to specialized equipment operated by lower-skilled labor. Just hooking up a bot to IBM’s Watson doesn’t produce a miracle. You need big data, A.I., and machine learning scientists to actually get a decent product up and running. When the market agrees upon a dominant design, we can stop tinkering and researching and focus on maintaining the service, writing the best copy for the bot, and marketing your company as different from all the others with a similar product.
The Organization will progress from an entrepreneurial organic firm to a hierarchical mechanistic firm with defined task and procedures and few rewards for radical innovation — the classic tale of three people starting in a garage working on a product, getting traction and funding, and eventually turning the operation in an actual company.
The Market will go from a fragmented landscape with lots of build-measure-learn moments to a stable field of largely undifferentiated products. The current bots are very good at understanding and remembering what a user says. This has led to bots telling you the weather for your location, giving you the news for the subjects you like, and delivering that one pizza you prefer. But there are only so many assistants a user will want to deal with. Just imagine having a secretary for every specific task. Remembering which one does what is about as much work as doing the task yourself. That’s why all the one-task bots will eventually be replaced by a couple of multi-purpose bots.
Lessons from the past
Towards the journey of becoming part of the oligopoly of firms making a multi-purpose bot, it might be wise to look at the lessons learned from other industries during their evolutionary path.
For an innovation to gain overall market acceptance, it helps to use existing forms from relating markets. Getting used to a PC was actually not that difficult for most typists, since the QWERTY keyboard they used on their typewriters was the same. That’s why we’ve already seen the emergence of the familiar “hamburger menu” in bots.
Be prepared to change. A lot. The adjustability of the design is as important as the design itself. Quickly moving away from what doesn’t work to what does is vital. Keep your product design dynamic. Don’t create static user stories or decision trees. Bots have already seen the addition of buttons. You just know finalizing payments within the conversation will happen. So don’t design a process which can’t be easily and cheaply adjusted to handle payments.
Think systems. The invention of the lightbulb would have never been a success if Edison hadn’t thought of a system to power up all those lamps. If I start a conversation in Slack during work, I might want to finish it on Messenger when I get home. Which assistants can do that right now? Humans. Bots won’t be able to help you outside of the walls of a single messaging app. In order to be useful for the masses, you might want to start thinking about a solution right now.
Handling all the changes, adjustments, and updates is critical to keep serving your customers and thus staying in business. To do so, you need to have the right competencies on board. Innovations generally enhance or destroy current competencies. Bots are stuck within apps for now, meaning they all work with the User Interface (UI) provided by the app. If you hire a UI designer to work on a bot, is that competency enhancing or destroying? Online copywriters can work magic for websites and apps. But a bot’s copy will be presented in a radically different form. Does that enhance or destroy the competency of a copywriter? In answering these questions, keep in mind that competence-enhancing innovations come from established firms as well as outsiders. Historically, competence-destroying innovations nearly always come from outsiders.
One design to rule them all
The most critical part of the model is the emergence of a dominant design. Apart from the content of the experience, visiting CNN.com and FoxNews.com isn’t all that different. A logo at the top of the page. A menu with different topics. A search bar. The most recent and important stories featured with a catchy title and picture. Older and less relevant news as a collection of text links. I can even navigate towards the most important stories of the day on the Spanish site of El Pais without speaking Spanish. That is because the format of (news) websites have a very distinct dominant design. If you know how to use one, you know how to use them all.
Now try getting the breaking news from three different bots. Do I ask “Could you please give me breaking news?” or do I say “breaking news”? Or should I press the menu button? And will it update me with the news tomorrow, or do I ask for it again? As soon as you’ve figured it out on one bot, you can start figuring it out on the next, because that one has different rules. As long as the general public doesn’t know how to talk to a bot or what the usefulness of one is, we will never reach a point of market acceptance, and bots will go down in history as a failure.
So keep making those mistakes, keep iterating, keep designing, and either come up with the dominant design or be prepared to change your entire structure to accommodate the dominant design. That’s how bots will go from hype to the cure for app fatigue.