Determining the Business Relevance of Artificial Intelligence

Deep Red vs Augmented Attic Possum. Lucy Adelaide

When determining the relevance of artificial intelligence in your business, exploring what AI is good at is one place to start.

AI, through improvements in information processing, is currently good and getting better at the following:

1. Automating. Good candidates for this include support interactions, activities in your CRM and the use of chatbots that integrate into data records and hand off to support agents.
2. Predicting. Identifying patterns in data and learning based on more data processing. Examples include churn predictions, customer action triggers, customer value proposition influences on key business metrics.
3. Categorising. Sorting, identifying and determining the difference between a dog and a cupcake. Techniques such as clustering can uncover natural segments in your data. Anomaly detection can identify outliers in your data.

Rob May explores this idea in more detail in this very informative post.

Weak vs Strong AI

We have built small scale AI implementations that align to the above capabilities, some of which I will share in the next 3 articles. Overall we are working with weak AI — systems that have the potential to bring efficiencies to specific business situations, through the use of technology in new and innovative ways.

Strong AI is more Hollywood, generalised intelligence, machines with consciousness, sentience and mind which is not a present day reality. What has really changed recently within the AI world is the enhanced ability of systems to self-learn and improve through the information that they process i.e. through machine learning. This is where the real power in our AI solutions lie.

The job stealing, self-flying saucer with optional neural network. Lucy Adelaide

It is worth calling out again that underlying AI technology is still not as smart as your attic possum if you think about true generalised intelligence. It’s very easy to fall into the trap of seeing something that seems extraordinary, such as a group of autonomous drones moving in unison and explain it in anthropomorphic terms. Just because it seems like human intelligence does not make it so. Just because a system like Siri can respond to natural language does not infer any understanding. Deep red unfortunately has no idea it is playing chess against our augmented possum.

AI as a Platform Layer

How can I directly utilise the AI power available today to improve my business and drive efficiencies? As consumers we all use AI elements within our phones, our apps and when being recommended movies on Netflix. As business leaders what are we doing?

When determining the business relevance of AI, I am considering it as a technology layer, like social, mobile, cloud, self-service in all present and future platforms. I am also asking what my vendors and implementation partners are doing in this space. We are building competence with the AI capabilities on Microsoft Azure, our preferred platform, and others such as IBM Watson, Dialogflow, Tensorflow. These small steps have helped us identify practical use cases.

Checkmate

The alternative to these small steps is to be checkmated by competitors who combine the right talent, tools and investment in more creative ways. Sitting around talking about AI and not building will also ensure we cannot see or remove the threat until it is too late. The small steps we are taking bring clarity.

I’ll share some of the basics of what we have learned in the next 3 articles by providing an overview of some accessible platforms you can easily use today.

This is part 2 of a 6 part series of articles exploring practical business applications of artificial intelligence.

Thanks to Lucy Adelaide for working on the illustrations above.

Originally published at https://www.linkedin.com on October 15, 2017.

Part 1 — Charlatans, Sharks and Dreamers — Navigating the AI Noise