Are You Building Your Business BrAIn?

By Stephen Garside on 4/3/2018

In this article we are going to explore the idea that organisations should be considering a concept I have called a 'business brain'.

We will look in-depth at just what this means and also take a look at a brAIn in action!

The article is available as a powerpoint presentation to download or you can just sit back and watch the video!

I have deliberately kept the content non-technical and where possible 'code free' - its more about the concept rather than actual implementation.

What do We Mean by Business BrAIn?

We can easily describe this concept in the following sentence...

“a knowledge architecture that securely interprets and distributes business intelligence to people, software and IOT”

Who Will Use Your Business Brain?

Anybody who has interactions with your business could utilise your brAIn. This includes your customers, employees, the general public and even your competitors! There will be other less obvious users such as your customer's systems for example.

How do we Connect to the Brain?

Connections to your business brAIn can be made from various sources including chatbots, voice interfaces such as Google Home or Amazon Alexa, etc, websites and internal business systems. 

Each of these connections are basically methods of gathering input from the user and readying it for passing on to the next step which is 'interpretation'.

The Intent Interpreter

The next component in your knowledge architecture is something called an intent interpreter.  This component is responsible for taking natural language input from various sources and attempting to understand what is being asked.

This natural language processing (nlp) component is arguably the most important in the entire process. Its fair to say that understanding intent is everything!

Example of an Intent Interpreter

My Intent Interpreter of choice is the excellent Dialog Flow from Google.  This software takes natural language as its input and attempts to match it to a pre-defined intent. When it finds a match it then tries to derive key values required in order to fulfill (answer) the intent.

The example shown here is an intent (a question) for a customer who wants to buy a garden gate.  Various examples of how the customer might ask this question are provided, together with the key values required to provide an answer.

Intent Values

Following our 'Gates' example, a customer has asked to buy a garden gate. To provide an answer we need to know the quantity of gates required, their length and colour.

The Intent Interpreter understands that we are asking to buy and gate and also knows that it needs these three values to provide an answer, so will prompt for these values before passing the request onto the next step.

Meet the Intent Processor

Now we have an Intent, the next part of the puzzle is the Intent Processor.

This component's responsibility is to take the intent (and it's key values) and attempt to provide an answer using its understanding of the business / organisation.

A Working Example of the Intent Processor

The image below shows an example of the Intent Processor working its magic. 

A customer enters their question about buying a garden gate into a chat window on a business website. The Intent Interpreter matches the question and values to a pre-defined intent.

The information is passed on to the Intent Processor that understands to answer this intent it needs to call the Product and Pricing service.  The information provided by these services is then packaged into an answer that is passed back to the customer via the intent interpreter.

What About Knowledge Security

They say "Knowledge is Power", so you want to share your business's  wisely. 

As a basic example of knowledge security, you might want to restrict pricing and exact inventory information to internal business systems only.  Certain FAQ's might only be for internal employees. Using a combination of device and role should give you the control required.

One of the key functions of the Intent Processor is to check who / what device is requesting the inbound intent, and whether or not to divulge the information.

 

Feeding Your Business Brain

You you have your knowledge architecture in place - what do you feed it?

There are the obvious candidates - product and inventory data, frequently asked questions and answers, business logic, digital assets (images, pdfs, videos etc) etc. 

Less obvious information includes trained Machine Learned Model Endpoints and Html / CSS based rich content.

A Business BrAIn in Action

So, its time for a few examples of a business brain in action!  To demonstrate we are going to use a business called 360Jungle who provide online virtual reality and vr tours

360Jungle have exposed their business brain to customers via voice apps for Amazon Alexa and Google Home. Customers using these apps can ask questions about their products, latest news, support questions etc.  By using the integrations from DialogFlow, the brain is also exposed to customers via chat-bots on their Facebook and Twitter pages.

Business Brain Via Amazon Alexa

Business Brain Via Google Home

Using Your Business Brain To Drive Website Content

Earlier in this article we discussed distributing web content (HTML and CSS). The 360Jungle Customer Interact feature is a great example of using a business brain to expose this type of rich content. 

Rather than navigating a website using a traditional menu, the Interact feature uses customer intent to pull rich content into a chat stream, offering a unique customer experience.

In the example below we can see how media depicting latest news articles are being injected into the chat stream on demand. Generating web content in this way means taking a different approach to building and serving your html and css so it can be used on both your website and also in brain responses.

Other Examples of Customer Interactions with your Business Brain

So what other questions can your business brain fulfil - here are some more examples, and remember, these can be used by customers, employees, or systems:-

  • Tell me about product x?
  • Can I get a price for service y?
  • How much inventory do you have of product y
  • What delivery option do you offer?
  • Show me pictures of product x in colour y
  • How do I install product x
  • Tell me your latest news
  • How does service x compare to service y
  • How do I process this type of transaction?
  • What were the sales of x yesterday?

In Summary

Hopefully this article has provoked thoughts on how you can implement a business brain in your organisation. I will leave you with a quote from earlier in the article, and please feel free to get in touch if you want to discuss ideas or a bit of advice.

“a knowledge architecture that securely interprets and distributes business intelligence to people, software and IOT”