As businesses adapt technology-driven models, experts believe that using artificial intelligence (AI) in business intelligence (BI) tools can automate many of the manual tasks involved including data preparation, cleansing, and analysis, among others. This can free up BI professionals to focus on more strategic tasks, such as developing new insights and recommendations, among others.
About 80% of enterprises reported that demand for AI and machine learning skills will increase as the use of AI proliferates further, as per insights from Statista, a market research platform. It is believed that businesses adopting AI in 2024 can expect to improve efficiency, save time and decrease costs, transforming businesses and the wider economy, among others. In conversation with FE-TransformX, Anurag Sanghai, principal solution architect, Intellicus Technologies on how the collaborative effort of AI and BI can upgrade businesses. (Edited excerpts)
What are the primary BI tools you’ve worked with extensively?
Intellicus itself is a BI tool and a complete platform which covers the full spectrum. So, as a business platform, we do not need any additional platform for BI parts. Everything gets sufficed by Intellicus itself. It has an inbuilt extract, transform, and load (ETL) module and an inbuilt scheduler. Additionally, Intellicus has an inbuilt pixel-perfect reporting module. So, everything is on a single platform.
How do you integrate data from disparate sources into a unified data warehouse for BI purposes?
I have started seeing a trend in the industry where large enterprises tend to have a unified solution or a platform rather than orchestrating between or trying to stitch them to create a solution. So, that becomes very cumbersome for them. So, rather than that, they’re looking or they tend towards going for a unified platform. And that’s what we offer.
With AI how do you ensure data quality and consistency within a BI environment?
Data quality and consistency are paramount for every organisation. There are several ways to ensure that the data quality is consistent. Intellicus has a quality assessment module and a mechanism that can do data profiling and quality assessment whenever we create any ETL flow. It also has data standardisation rules which are validation rules.
I believe AI can upgrade ETL module and help in connecting to different sources while processing the data, at every layer, with a logging mechanism where we ensure that there is no data loss. These rules help to create a detailed log. Apart from this whenever there is any discrepancy, an alert will be automatically generated.
How machine learning (ML) algorithms are used within BI projects?
Intellicus has inbuilt ML capabilities as well. Now, we use ML and advanced analytics in multiple phases. One is for doing augmented analytics, predictions, and forecasting. Additionally, Intellicus has an inbuilt data science module or a data science step in our data preparation layer which helps in data cleansing and executing a missing data treatment. So, if there are some missing data points, it matches using a machine learning model.
Furthermore, ML can help in intelligent character recognition (ICR), optical character recognition (OCR) and extracting data from images and PDFs. In addition to this AI can be used to create BI dashboards that are more personalised and interactive. AI can also be used to provide users with real-time insights and recommendations. Additionally, AI can make BI more accessible to business users who do not have any coding or technical expertise.
How do you see the intersection of BI and AI shaping the future of data-driven decision-making?
I foresee predictive analysis and augmented analysis can be enhanced further and become more sophisticated. Nowadays what we see are the standard predictions which are time series-based. But what if it could also find out a correlated prediction? For example, we see sales or are predicting sales of an organisation which is increasing by 8-10% every year. Now, a standard machine learning algorithm will do training on the historical data and give a forecast. So, what we foresee is on a similar line this would be a sales. Now, what if the system could also tell you that your sales will increase by 8 to 10% but your profit will go down? So, with the power of the large language models (LLM) a lot of these correlated analytics is possible.
Another is the recommendations part, by doing a cognitive, prescriptive and the recommended analytics. With natural language-based analytics the integration of ChatGPT with different platforms will be possible. So, these are the areas primarily where I see the future with the integration of AI with BI.
When we talk about automation the first thing that comes is security. What can be the loopholes in automation?
On the positive side with robotic process automation and machine learning (ML) end-to-end processes could be automated. There can be a risk because you cannot rely 100% on machines. But by saying that, that does not mean that it’s too risky there. There are several ways by which even a fully automated system could be fail-safe. For example, by having a planned master data management (MDM) businesses can ensure the consistency of using AI across the processes and organisations, without any duplicate information or repeated tasks. So, whenever there is an anomaly it should quickly point it out and generate an alert.
