It’s most of the way through 2018, the landscape and tools for businesses to leverage automation and NLP (Natural Language Parsing) in their business messaging channels is maturing, making developing a ‘chatbot’ easier.
Typically we see brands start this journey in one of four key areas of their business, usually their highest pain point, current challenge/objective in the business or looking for bang for their buck.
It may be led by marketing, sales, operations or customer service; each one of these business functions benefitting from automation to some degree.
Building a conversational experience (‘chatbot’ is a little primitive in this context) that can deliver the objectives for only one of the key business functions isn’t too challenging. Your audience is well known, what they are aiming to achieve is defined and more often than not the process can be mapped out in a straightforward manner. Most importantly when engaging your audience, the user’s context (why they are there) remains the same.
Running this over multiple channels like Facebook, Twitter and live chat can be solved using the right tools as well.
A period of negotiation usually kicks in when the channels of communication with prospective and existing customers are the same, when the different business units have slightly different objectives but must play in the same space.
Where the train starts to derail is when the different business functions all want to run their own conversational dialogues, on the same channels and with a different context about why that customer is there to chat about, let alone do it in multiple languages.
Here is why we believe creating a single purpose ‘chatbot’ isn’t too hard, orchestration is the challenging part. Orchestrating multiple business units with different objectives, differing conversational flows for users with vastly different contexts as to why they are chatting to you, do this across three or four channels and then add multiple languages for good measure.
Here’s an example. Marketing wants to run a short term event activation, sales is trying generate long term leads, customer service is dealing with a volume of existing customers who can automate some of their queries and others need handing off people for support. All this in two locales; English and Traditional Chinese.
You have an orchestration problem that a simple ‘bot’ won’t cope with.
We haven’t even begun to analyse how many people engaged with the event experience, what happens to the leads acquired and how they get out of the ‘chatbot’ system and when does the ‘bot’ know to stop talking in the customer service handover flow.
Orchestration is, down right hard. Making all this work together is where business come unstuck when they embark on building their ‘chatbot’.
Here are some core orchestration use cases:
- FAQs > Lead acquisition > External CRM;
- Intent discovery > Customer service flow > Human handover > Loop the user back to the beginning;
- Platform ad > Known context > Deep in the conversation flow.
Take all of the above and run them simultaneously on web chat, Twitter and Messenger in 3 languages.
If business’ are to truly leverage what can be done in this space, ‘bots’ are not the problem we need to be solving — orchestrating all the moving parts required to make an amazing customer experience is.
Developing Iris we’re looking beyond simple the problem of a singularly focused ‘chatbot’, more to the future and the (very) hard problem of powering the many interconnected business challenges. How do you run a multi-faceted and dynamic business solution in this constantly changing environment of business messaging?
An orchestra can only as good as its conductor.
For more about about how Iris powers meaningful conversations at scale across your messaging and voice channels, visit Iris product details