BOTs are hot these days, you cannot visit a single tech website or tech event without being overloaded with stuff about BOTs: it's like every company in the whole world already owns several ones, and you're only company without a single one . Like, if you don't have one as a company today, you will be out of business even before the next morning will come.
Well, don't worry (yet): the soup is never eaten as hot as it is served. Nevertheless the great potential and fast development, only a few companies are able to turn their chatbot idea into reality and make it a differentiator and success towards competitors and customers.
So, why are BOTs such a challenge ?
First of all: marketing always runs in front of technology
Customers will only talk to your BOT when it's able to return valuable information and communication for their questions. The success of your BOT depends on the output and value of its communication: understanding words and sentences is one thing, but interpreting what is being said or asked is what differentiates humans from technology (today). I must admit, A.I. is very promising, but still not mature: the A.I. systems and Natural Language frameworks needs to be developed and trained (a lot). And still there will be a lack of understanding dialects and mixed languages which we use during our daily conversations.
So, the marketing machines of Google, Amazon, Microsoft, Facebook, ... are showing us great potential and opportunities ... but, until today these demonstrated scenario's only exist in an isolated 'ideal' world.
Secondary: the challenge is about how to connect the DOTs to the BOTs
If we take a closer look at many companies in place IT and application landscape, we can state that at least a few challenges and hurdles needs to be taken. Now, take into account that a BOT can only be as smart as the information and services it can access and provide as a response ... we instantly know the possible bottleneck(s) towards your first BOT-success. This actually means that 'the BOT' needs to be connected to 'the DOTs': other sources and resources, like internal ERP systems, warehouse systems, product information systems, transport systems, ... For many companies this is a major showstopper, even if it could technically be possible to connect your BOT, nobody can guarantee or even prevent the risk of being hacked. Since a BOT lives on the 'Wild Wild Web', the risk of being hacked is a fact and only a matter of time in today's online and connected world.
This often means that the oldest application is your biggest risk: it will always keep you in the present world. API's are like communication keys between applications. So, if your back-end applications are suffering a poor API, this often means your game is over before it even started.
Last, but not least: your BOT is not an out-of-the-box customer magnet
A fool with a tool, remains a fool: creativity and simplicity are key for your first BOT's success. Your BOT will always remain a tool, if you don't know the 'why I need this so badly' of your BOT ... it will die a quiet death. Your BOT needs to deliver the WoW-effect towards your customers, it needs to be smart, easy and likeable. Therefore your BOT needs a vision, a goal, a purpose, a why and a how: questions that needs to be clear and answered.
Your BOT needs to be part of your online and customer strategy, it will become a powerfull customer touchpoint, so: think before you act. Keep in mind that the customer has no patience and has zero tolerance for failure.
After all, your BOT will be your company Tamagotchi: it will always need attention and care.
My advice: start playing with it, try it, do it. Try the already existing BOTs and think about some small, but creative scenario's where your BOT could make a difference. Create your BOT and select a few testers to play with it and to train it. Experience yourself what a BOT can, could or should do.
Prepare for tomorrow, and be ready for the day after tomorrow!