41 percent of executives surveyed in a recent McKinsey study admitted they have not yet adopted AI because they’re unsure about how it can help their organization. But 100 percent of those executives are thinking about it. It's not unexpected, it's on everyone's mind because it's in everyone's lives. Alexa, Google and Apple have invaded the home. Tesla and Uber are working on AI systems that will drive themselves. IBM’s Watson has its own office and is predicting cancer treatments for real patients.
So what is "AI"? How does it work? Can it apply to your business?
Is AI really intelligent?
Artificial Intelligence, AI, is a deep concept and means that there is an intelligent agent at work. However, whatever is in the market today, is certainly not general intelligence in any form, and is not intelligent in any meaningful way. Even Watson or various tools to apply AI to business are just Machine Learning algorithms. While ML uses neural nets and is the future potential underpinning of AI, it's not the same as AI.
How does Machine Learning work? Machine Learning is the first, important step towards AI and it’s something that can be, and is, used right now. Machine Learning is the ability to apply a multi-layered neural network for pattern recognition in data. Sounds complicated? Let me explain.
Even stupid chatbots still can help
There are two main use cases of AI adoption in business.
First, one is using prebuilt agents or platforms that allow you to apply Artificial Intelligence, Machine Learning methods to your domain. One example might be chatbots, which are domain specific and limited. Even Alexa, with its sophistication, won’t understand you, unless you ask it things in quite a specific way. Some other examples are image analysis/recognition, or sound analysis (like music).
The second main category of use is in data anlaysis. When you have a LOT of data and you want to analyze it for patterns, specifically new patterns that you are not necessarily aware of. Machine Learning can be great solution, using unsupervised learning.
Watch the video for more discussion around the two use cases.
85% of AI projects will “not deliver”
According to Gartner, risk and confusion are two main reasons for failure rates for AI projects. First, the process is not easy or quick: employees and customers alike know how your products work and what to expect from you. Investment of time, training, and money in developing successful AI projects is a big risk – most companies can’t quite justify it yet. AI adoption is very experimental in nature in terms of how you can apply it to a business. I think there are two main reasons to take the risk and decrease the confusion.
First, you want to take advantage of the hype around the topic. It is something you can start to embed in your R&D, in your technical approach.
Second, you have some specific use case that might be correctly solved by the approach of using ML or deep learning. For example, when you have a lot of streaming data, and you don’t have the manpower to process and analyze the data.
I talk more about use cases and when they are appropriate in the video.
Take the first step to adopt AI
First of all - choose one of the two contexts - either usefulness or thought leadership position. This fundamental decision will determine how much money you want to put towards the effort and what kind of team you need to assemble.
What next? Watch the video where I explain how you could approach AI adoption in your organization.
If you want to talk more on how your business can profit from adapting AI, schedule a call with me for a free one on one chat.