What is the difference between advanced analytics, machine learning, and artificial intelligence?

My favorite: if it’s written in Python, it’s probably machine learning. If it’s written in PowerPoint, it’s probably AI.

Advanced analytics generally refers to the combination of having a data platform, using machine learning for predictive analytics, and having the right talent in place. Machine learning and AI are basically interchangeable when it comes to analytics - people refer to the same thing.To become a data-driven organization, it's important to know where your organization is in terms of data capabilities - descriptive (advanced) analytics, predictive analytics, or prescriptive analytics.Descriptive just means you have some data, the ability to run reports in a static way. The reports run on a batch basis (day-to-day) and you have those reports available for users. E.g. how many users were on my website yesterday;Predictive means you have big data capabilities, real-time streaming, and you are using machine learning to 'predict' the future. E.g. you will be able to predict how many people will come to your website based on historical trends;Prescriptive means you can not only predict the future, you can tie that future to creating value for your business. You are actually influencing the output, not just looking at some reports. E.g. you will be able to recommend the right products to the traffic that is coming in based on their actions/history to increase purchases. Image courtesy DeZyreFor any data-driven organization, there are two key things to consider:Speed-to-insights: How quickly can you get to insights. Using emerging technologies, you can do real-time speed to insights. E.g. using Kafka to realtime stream and use Hadoop + Spark + in-house visualization. This reduced our speed-to-insight at Visa dramatically;Insights-to-action: What can you do with this data. It's not unusual to see reports taking 1-2 months to create based on business requirements. Even after the first process of sourcing the data, ingesting the data, ETL, data modeling, and then reporting, no one uses the reports because there is a product-market fit issue.The key to success here is to work backward from what business users are looking to consume, as opposed to working forward from the templates and reports generated by tools like MicroStrategy, etc.When it comes to machine learning and AI (which are basically interchangeable here), the difference is the platform organizations use to power their data. It used to be that organizations applied statistical modeling on small data sets to predict the future. Now with the innovation in compute, storage, and network, you can retain endless data (e.g. using Hadoop to store / Kafka to stream data in real-time) and run your models on huge data sets, going one or more years back. Now you're no longer guessing. But here are the key things to keep in mind for success:You have to ask the right questions (get the right functional people involved to define the use cases);You have to have the right platform AND access to the right data;You have to have talent that understands big data, machine learning, and are good story tellers.It took us a while to get this in place, but the benefits have been extraordinary.