Artificial Intelligence (AI) is in the running for the most overused and overhyped term in business, IT, and popular culture. With the term AI used in so many different ways and contexts, you may easily think AI is not something you can apply quickly to your business operations.
You are wrong.
The algorithms and statistical underpinnings of AI as applied to business are old and have been used for decades. Indeed, new technologies are making AI more powerful and expanding its potential uses. However, using machines to analyze and predict what might happen from historical data is not anything new.
The biggest impediment is bridging the gap between what AI can do and how you can use that to change business processes. For that bridge, you need focus.
When you think about AI from that perspective, hopefully you see it can have some immediate applications to your organization’s operations. Here are three targeted, focused, and quickly achievable common uses of AI.
1. Chatbots
Chatbots are becoming common on websites for customer service. They are increasingly being deployed for internal purposes as well. The obvious benefit is that no person needs to be actively on the company’s side of the conversation (until the conversation gets beyond the point the chatbot has been trained to handle).
Many providers offer chatbot services. Most work by ingesting a set of documents, and then extracting the potential words and questions that could be asked. This is the AI part of a chatbot—not having to program every response explicitly.
2. Forward Predictions
Probably the oldest use of AI is best thought of as applied statistics. For a long, long time, organizations have sought to predict future events or outcomes—from financial projections to sales to production demand and so on. Most of the time, this work is still done using spreadsheets and educated guesses.
AI brings discipline to the process, and not in some mystical way. The algorithms range from time-tested statistical predictions to cutting-edge neural networks. Regardless of the algorithm choice, when used against historical data to make predictions about future events, that is AI.
Taking the forward-looking predictive process from scattered spreadsheets to centralized AI processes is one of the easiest and quickest AI wins an organization can have.
3. Correlations—Now and Trends
Another common use of AI is to show how different factors affect outcomes. For example, what are the factors that suggest a customer will buy from you? Looking at historical data using AI algorithms highlights which factors are more critical to getting that customer. A similar analysis can be done for cost factors, employee turnover, and many other quantifiable parts of your organization’s operations.
Correlations are useful for informing people about key factors that influence an outcome. Knowing what is important and what is not is extremely useful for focusing time and resources on things that matter.
A highly informative extension of AI for determining correlations is to track the effect of different factors over time. Seeing how influences are changing is a great way AI can help to guide organizations to improve.
What Not to Do
Be careful about setting the performance bar too high with AI projects. For some areas, like clinical medicine, the risk of a wrong AI outcome is catastrophic. Those situations require rigorous testing and evaluation, leading to long development periods. Most AI uses for most organizations do not need that level of certainty and precision.
Thinking AI systems must be perfect to be useful is like thinking every employee will perform flawlessly in every situation. That is simply not going to happen. Good enough is good enough for the AI system, with the understanding that it will get even better in the future.