🫵Your #data quality as well as the #ai #ml #llms accuracy are your least problems when it comes to business decisions.
🤔As a #datascientist #dataengineer and #data guru, you may be under the impression that #machinelearning #ai and #data quality are the most important parts in building automated decisions in organizations such as #banking #insurance #financialservices and #government. You may ask, who does not want a very accurate predictive model?
⚠️If you work in an #ai research company or in a business software company where your task is to increase the accuracy of a particular approach in #ai like #transfomer #nlp, etc., that's a different story. However, to automate business decisions in #bfsi #government organizations, there is a more important problem to solve. Then, the model accuracy becomes important in that context. In fact, not all business decisions require a form of #machinelearning #nlp #transformer #predictiveanalytics techniques.
🛑In #banking #insurance #financialservices and #government organizations, your first and foremost problem is to model business decisions, then understand within the context of the business decision what techniques fit in well. Is that #ml #ai #nlp #nueralnetwork or, perhaps #optimization #rules and data #orchstration? Or, more than likely, the combination of all of them!
It depends. Does it depend on #data #ai #datasience architecture and design? BIG FAT N🚫!
It all depends on the business operation. What does this mean?
⚒️It means the team, group, or people are doing specific tasks in the business on a daily basis. So, the models #ai #ml #datasience outcomes must fit into their daily activities. What do they do?
– They do #underwriting #pricing #adjudication #claims
– They do #fruad analysis, calculating #creditscore
– They assess product eligibility and configuration
– They provide recommendations, #policy review
– They calculate #risk #rating #payment amounts based on specifics of contracts and terms
✨These are some of many business decisions they make and execute on a daily basis. Hundreds and thousands of them every day. Yes, it all depends on the business decisions.
🚀Focus on modeling business decisions #decisioncentric using a decision graph. That is a hierarchical model representing the holistic view of business decisions. It uses the decomposition technique to break down the dependencies of complex business decisions into smaller, more easy-to-understand decision units. Then, look at those units individually and identify which technique fits the best based on the nature of the decision unit. As a result, you will use a composite model that integrates multiple approaches #ai #ml #redictiveanalytics #rules #optimization to automate a business decision.
#decisionintelligence
Posted here.
Published April 24th, 2024 at 07:30 am
