The journey to AI part II Nicolas Babin IBM Blog
Written by Nicolas Babin

Data relevance

In the early 50s, the power of data started to become clear. We could sense that data was something of value, but nobody knew how to use it and what kind of information could be valuable.

In 2012, according to the Guardian, only 0.5% of all data was analyzed. In 2017, the Economist claimed that data would replace oil as the world’s most valuable source. Even if many people would compare data and oil, they did not mention the fact that data was very different from oil. In fact, data can be extracted, analyzed, shared, stored with endless means. Moreover, anyone can use data endless times and get very different type of insights from it.In 2012 a report from IDC about big data statistics claimed that only 22% of all the data had the potential for analysis. It also claimed that by 2020, the percentage of useful data (that can be analyzed, shared, stored and that has the potential for analysis) would climb to 37%.

Many people will ask why AI is now so popular. AI will allow data to be optimized and used in a way that you probably have not expected yet. Thanks to AI you will be able to predict and organize models that will then give you more potential. Thanks to AI you will think of disruptive ways of doing business. You will optimize your resources, and finally ensure that processes, experiences and decisions are automatized based on intelligent models.

Today with the rise of Chief Data Officers, data scientists, and big data technology, there are vendors capable of providing solutions that allow easy and efficient access to data.

The journey to AI is challenging, but it will ensure business success .

A good approach is to use 4 steps that allow an easy understanding of what is required in an efficient, highly transparent process.

As a first step you will need to collect data. You probably have a lot of data available through your company. You need to audit what is available and ensure you can collect new types of data while combining data already stored. You also need to ensure all data remains with the same architecture and that all data can be used.

Once data has been collected, it is critical to get it organized. Using big data architecture and databases allowing the 5 Vs (Velocity, Volume, Value, Variety and Veracity), you need to ensure the data is prepared for analysis and organized in an intelligent way so data specialists can get their algorithms optimized.

Analyzing data is the next step. Data scientists will be able to determine what kind of data is needed and how to analyze it. Once the infrastructure and the architecture are in place, data scientists can operate their magic and work with you to ensure data analysis optimization.

Once data has been collected, organized and analyzed, you will need to get it Infused. This means build trust, transparency, compliancy and ensuring model based decision making and explanations are in place. Bias is a challenge for all AI algorithms. What kind of data should be used; how should we use them; could just one data scientist make these decisions; would bias use of data create some issues with AI algorithms?

Data relevance is key to all AI projects. The right data in any format with the right architecture and analyzed with the right data scientists will ensure success in your project.

According to New Vantage, 97.2% of organizations are investing in big data and AI today. In the same report from New Vantage, you can read that 62.5% of survey participants said their organization appointed a Chief Data Officer (CDO) with data scientists recruited as well.

Companies are transforming their business; it is very important everyone thinks about a new strategy to ensure your company survives the next 5 years.

 

 

Back to top