Rajavel Selvaraj Ganesan on Generative AI in Analytics

Artificial intelligence (AI) has been all the rage lately with OpenAI’s evolution and consumer adoption of various use cases. In Rajavel Selvaraj Ganesan’s, or Vels as he is known to many, perspective, AI is going to be powerful in the “analytics” space due to the way data, data, data everywhere!! So, being able to intelligently indulge AI on top of the enterprise data warehouse system is like mining gold against it. 

Vels is a solutions architect with a bachelor’s degree in engineering focusing on computer science and engineering from India’s Anna University and a master’s in business administration with a concentration in project management from Ulyanovsk State University in Russia. He helps companies design management technology that enables executives to organize data and other forms of information to render their entities more productive and orderly. 

There are four types of data analytics: 

1. Descriptive analytics (what happened?)
2. Diagnostic analytics (why did this happen?)
3. Predictive analytics (what will happen?)
4. Prescriptive analytics (what should we do?)

Understanding how Generative AI can help in this space would be interesting. Here, Vels shares his research and view of how AI helps on top of these four categories and how they adopt AI into their products for analytics use cases.

Descriptive Analytics

Descriptive analytics can show “what happened” and is the foundation of data insights. It looks into past data to understand what happened and find relative patterns and trends. Vels likens this to “data augmentation,” meaning to generate synthetic data that can help people understand various scenarios of business cases. In “data augmentation,” generative AI can help produce or generate synthetic data that can augment existing datasets, allowing for a more comprehensive analysis.  

Generative AI can generate synthetic data that mimics real data without the challenges of collecting it.

MicroStrategy

Vels explains that it’s possible to look at how this is being applied or planned to be utilized in MicroStrategy, a leading provider of business intelligence software and cloud-based services. MicroStrategy shared a project called “Project Data Forge” at the MicroStrategy World 2023 Conference. 

It utilizes “generative AI” for data stewardship or analysts to generate dummy augmented data for data analysis. Isn’t that awesome!!? Looking back to 2000–2020, if we needed to load data in an enterprise data warehouse in our Development or QA region, it was a long process (such as working with DBA & performing data masking or figuring out a strategy for the development team to proceed). Another scenario is working with the ETL team to understand source systems, developing a pipeline, and then loading the data into data warehouse systems (database) for the Business Intelligence (BI) team to build reports or dashboards for customers. This “Project Data Forge” concept by MicroStrategy will be a game changer in generating dummy data based on the business need to continue developing reports or dashboards for customers. The ETL team can process the data from the source system and develop the pipelines to get it loaded into the target system (EDW or Enterprise Data Warehouse) for the BI team to process later. 

Diagnostic Analytics

Diagnostic analytics addresses “why things happened.” Common diagnostic analytics techniques/insights include drill-down, data discovery, data mining, and correlations. 

Tableau

This can be explained using the functionality within “Tableau,” a business intelligence analytics product serving various customers across various domains. For this analytics stream, Tableau developed an in-built AI functionality named “Tableau Explain Data.”This automatically provides AI-driven explanations for the value of a data point with a single click. It provides a deeper data exploration with statistical models applied behind the scenes to explain the details of the data for easier understanding for business users or analysts studying their data.

Another functionality that Tableau added to their product is “Data Stories,” utilizing AI built into it. Data Stories adds automated plain-language explanations to your dashboards in seconds. Isn’t it amazing to utilize AI to provide information about dashboards based on your own data to users in a simple language? 

Predictive Analytics

Predictive analytics helps businesses to “see the future” and predict “what is likely to happen.” For example, let’s say we have seen the last five years’ sales for an eCommerce giant in our database (Enterprise Data Warehouse). With predictive analytics, we can predict future sales based on historical patterns and trends seen in the business data. Below are the common processes in predictive analytics. 

1. decision trees
2. neural networks
3. regression models 

Generative AI is different from predictive AI. Generative AI can help design product features, while predictive AI can forecast consumer demand or market response for these features. In my experience, predictive analytics is an AI-based strategy. This means that it is called predictive AI by default when we do deeper analyses with historical data and predict the future using the models. 

Tableau

Tableau introduced an AI feature focused on predictive analytics & it is called “Einstein Discovery.” With Einstein Discovery, you can create trusted, predictive ML (machine learning) models without writing a single line of code. Usually, organizations tend to utilize a data science team to work on ML models and then infuse the data into the ML models tuned by data scientists to derive future predictions of your business. With Einstein Discovery, it is handled by AI instead of data scientists. 

MicroStrategy

MicroStrategy incorporates “advanced analytics” with AI. It automatically provides precise forecasting, allowing users to better anticipate market trends and optimize future strategy. It uses AI to predict based on historical data trends/patterns and provides business users or analysts with the right forecasting recommendations within the MicroStrategy application. 

Prescriptive Analytics

Prescriptive analytics, analytics driven by AI systems, helps companies make decisions and determine “what they should do next” based on the predictive analytics outcome. For example, rather than describing or diagnosing something that’s happened or predicting what could happen, prescriptive analytics can tell us what we should do about it. It is currently closer in reach via AI based on the growth of generative AI and large language models. 

In my understanding, I haven’t seen any Business Intelligence tools that have reached this stage of prescriptive analytics yet, and I hope we get there to that stage by 2030! 

A few great examples are shared from Bernard Marr’s site below: 

1. Customer loss prevention
   1. Customer churn score can be derived by predictive analytics, i.e., how likely are they to stop being a customer?
   2. Proactively advise on steps that could be taken to reduce the likelihood of the customer leaving via prescriptive analytics. 
2. Delivery logistics  
   1. How long will it take drivers on each route to complete their deliveries? 
   2. What is the most efficient route that each driver should take in order to complete their deliveries quickly and safely? 

Security & Governance with AI

Generative AI attracted many business use cases when ChatGPT became popular during its initial release. An organization needs to consider implementing the right Generative AI Governance to be successful. Generative AI Governance refers to the principles, policies, and practices designed to ensure the responsible and ethical use of Generative AI technologies. 

MicroStrategy

One good reference to it is with analytics/business intelligence tools like MicroStrategy, which is fundamentally defined from its metadata-based model approach. MicroStrategy has security and governance built within the application at various levels, such as data, object, user, security roles, security filter, DB connection mapping, project access, etc. 

With the above layers of security applied to the data being accessed by business users, generative AI adds better governance and security to customer requests. Below are the references for the architecture of MicroStrategy security and governance for Generative AI. 

Tableau

Tableau AI is built on top of the Einstein Trust Layer and inherits all of the security, governance, and trust capabilities. 

• Zero-data retention policy
• Dynamic grounding with secure data retrieval
• Prompt defense
• Data masking
• Toxicity scoring
• Audit

Rajavel Selvaraj Ganesan hopes this helps as a reference for people aspiring to utilize generative AI in the analytics/business intelligence space. 

To follow Vels on social media, please visit X and LinkedIn