If you are not leveraging Artificial Intelligence (AI), then you must read this. Eventually most, if not all, organizations will adopt some form of advanced analytics powered by big data and statistical models. The question is whether you will wait until it’s almost too late and potentially not survive changes in market sentiment, or whether you will adopt it early and command market sentiment. Analysts predict that consumers will soon prioritize solutions which offer hyper-personalization. This means, not only do you need to know more about your customers, but you need to deliver solutions through a set of experiences which predict what they need. Granted, this relates to consumer facing solutions, but consumers will also prefer solutions which use AI as a form of sustaining innovation to reduce cost or increase capabilities.
Many times, organizations do not know what is possible with AI. Worse, some organizations assume that an AI can be built with traditional staff. Historically, large ERP projects tend to be a career ending event for those involved in ushering a complex ERP transition ineffectively. Likewise, organizations which use unverified models live in their enterprise risk making mistakes at a large scale. For this reason, you should only be using models which you can statistically prove or models you know have been statistically proven by data scientists. In our firm, we consider a data scientist as someone with a PhD and a background in statistics and machine learning. In more and more cases, organizations are using ‘citizen’ data scientists (those who do not have PhD level credentialing) to consume components authored by data scientists. Some components are now modular enough that they can be consumed by developers such as Azure’s Cognitive Services APIs which provide text analytics, image recognition, etc. This can be an effective alternative for light weight advanced analytics in some applications.
What is possible with AI is best understood when you consider what AI actually does. At the end of the day, AI automates intuition. In fact, AI attempts to automate human intuition. Human intuition is extremely powerful, but has challenges both in understanding large volumes of interdependent data or any use of statistics. Machine Learning (ML) solves this by using new technologies to apply statistical models to big data. It’s true that big data has been around for at least a decade now, but the reason AI is gaining so much publicity now is because the rate of innovation has created more effective tools for scientists combined with more data from IoT and other sources.
To give you an idea of the kinds of solutions possible, consider a few of the types of projects our customers have asked us to build:
Predict when the price of certain commodities, the stock price of ten of our customers, influences the amount of textile inventory in one of our warehouses.
Automatically predict the complexity of a new part based on detected similar geometry of that part to all parts we have ever created in the past and the thousands of different staff it took to create the part.
Automatically detect new vendors we should be using based on similarities the new vendors have to our existing vendors by correlating every single word ever spoken by persons of leadership in the public domain by all vendors in the market.
- Predict whether our B2B partner’s orders are accurate based on every single order and revision ever made.
If you are not leveraging AI yet, you are already behind. Gartner predicts that by next year, about a third of brand leaders will have their revenue adversely impacted by AI capabilities others have introduced. Advances in AI tools make existing scientists more effective at driving value in your organization, so it is important to get the right tools to the right analysts with the right strategy as soon as possible to start delivering value to consumers now.