How to use AI with BI
In last decade several new technologies evolved like machine learning, deep learning, Artificial intelligence, Big Data and cloud. Rapid change in solution brings IT people life in difficulty to choose right tool. I am going to focus on AI and BI option here to emphasize that when we should use which one.
Story start with running business in more intelligently, risk free, having visibility in future scope. Easy for life, brings stability and scalability in business. Mid to large size companies used to have several databases for OLTP and they develop centralised data warehouse and run their decision making in such way. Several infographics tools like excel, web browser, Tablue, Power BI and microstrategy are having stable tools to visualize abstract information from gigantic DW to small window. Business user used to make decision or perform planning.
So fundamentally BI is basics of data driven decision. Making data platform standby to measure things (called KPIs) in several arguments called dimension.
Business with Artificial intelligence
But, What happened that people must jump of AI instead of using BI. In complex business scenario business is not able to identify formula for KPIs. They are unable to forecast numbers. Either underlying formula is not easily deterministic. Here AI comes to play the role to identify formula behind based upon historical data and forecast figures. In regular BI these formula is quite simple and human being used to define to forecast and calculate. But in AI case machine used determine formula and and test and fix mode based on outcome performance.
BI- Business intelligence is solution approach available from around 20-25 years. There are several tools are already in market and they are stable well tested and easy to integrate with almost several tools. Life is quite easy for everyone because its stable and comes in affordable cost.
Solution
Smartest way is do not throw existing data platform and jump on complete AI. Just Integrate AI services with BI and keep stable running with minimum impact. Existing BI technology is well strong to process and visualise data. Question comes what if data volume in unmanageable in tradition data warehouses and architecture is not supporting existing infrastructure compute or storage capability. I will come with another solution with data lake and multiple data warehouse solution in another post.
