Self-service analytics is perhaps one of the most powerful tools organizations have to enable widespread data-driven decision making. The goal of this business intelligence (BI) branch is simple – enable line-of-business professionals with query and reporting capabilities to perform citizen development and analytics without significant support from IT (Gartner, n.d.). The first wave of self-service analytics relied on simple-to-use BI tools such as PowerBI, Tableau and Google Data Studio (now Looker), enabling business users to connect to and access organizational data, explore relationships within and across different datasets and pull together simple dashboards to visualize key performance indicators (Shea, 2023). With time, the evolution of such tools has become imminent as descriptive analytics has become second nature for well-nurtured citizen analysts. Line-of-business professionals will most likely never run out of variations of the “what happened?” question but it is only natural that more advanced questions – requiring equally as advanced analytics techniques to be answered – will start being posed.
Although drag-and-drop tools to deploy citizen data science exist, there is a new kid in town. Generative AI tools have the ability to help parse meaning from vast amounts of data at scale, generate bespoke codes (Sankaran, 2023) and help alleviate the burden of repetitive or mundane tasks in the analytics lifecycle (Porter, 2023). The novelty of the technology and its natural language processing interfaces only makes the imaginable impact potential cosmic. Examples of how generative AI is adding value to analytics processes are already widespread. For instance, GitHub reports that their copilot solution has been signaled to aid developers to perform tasks around 55% faster than without the solution (Kalliamvakou, 2022). One can further speculate what this number would be if accurate and user-friendly versions of generative AI analytics tools would land in the hands of less data-proficient user bases. With a positive outlook, I like to imagine the number could reach high levels.
The potential of implementing generative AI tool in self-service analytics is, thus, great. That being said, for anyone working in analytics and BI branches of IT, empowering the entire workforce with accessibility to build advanced analytics solutions with generative AI, sounds like a nightmare. As highlighted by Shea (2023), such a system “(…) without the right tools, technology, and processes, it is scary and can devolve quickly into data chaos”. Consequently, governance, becomes one of the big concerns for organizations worldwide. In fact, in a survey performed by Evanta, 46% of Chief Security Information Officers report that their biggest concern is either data privacy, or secure deployment of solutions (Hiestand, 2023). Alongside governance, other challenges overshadow potential large-scale applications of generative AI in citizen analytics practices. The accuracy and reliability of outputs remains a concern as evidence from practical use showcases that solutions incorrectly construct outputs, also referred to as hallucinations (Porter, 2023). Most students who have tried to generate references using Chat GPT, for instance, will recognize this as a concern. Such hallucinations can have significant impact on the accuracy and potentially mislead line-of-business professionals in conclusions. There is therefore a real need to develop approaches to safeguard utilization of such tools in business intelligence.
Personally, I believe that generative AI is still far from ready to be scaled in organizations. I do, however, feel immensely optimistic that businesses who are able to deploy small scale applications with proper governance and guardrails in place will have superior advantages. This, especially in building systems to support every-day use in self-service analytics.
References
Gartner. (n.d.) Self-service Analytics. Gartner Retrieved October 14, 2023 from: https://www.gartner.com/en/information-technology/glossary/self-service-analytics#:~:text=Self%2DService%20Analytics%20is%20a,own%2C%20with%20nominal%20IT%20support.
Hiestand, L. (2023). The Current CISO Outlook on Generative AI. Evanta. Retrieved October 14, 2023 from: https://www.evanta.com/resources/ciso/blog/the-current-ciso-outlook-on-generative-ai
Kalliamvakou, E. (2022). Research: quantifying GitHub Copilot’s impact on developer productivity and happiness. GitHub. Retrieved October 15, 2023 from: https://github.blog/2022-09-07-research-quantifying-github-copilots-impact-on-developer-productivity-and-happiness/
Porter, T. (2023). Generative AI for Data Analytics: A Force Multiplier. Alteryx. Retrieved October 15, 2023 from: https://www.alteryx.com/blog/generative-ai-for-data-analytics-a-force-multiplier
Sankaran, A. (2023). How Could Generative AI Impact The Data Analytics Landscape? Forbes. Retrieved October 15, 2023 from: https://www.forbes.com/sites/forbesbusinesscouncil/2023/05/24/how-could-generative-ai-impact-the-data-analytics-landscape/
Shea, B. (2023). Everyday AI: Next-Generation Self-Service Analytics. Capgemini. Retrieved October 15, 2023 from: https://www.capgemini.com/insights/expert-perspectives/everyday-ai-next-generation-self-service-analytics/
What a well-balanced article! I personally find it insightful of the potential and challenges associated with integrating generative AI into self-service analytics. I specifically want to thank you for effectively stressing the transformative capabilities of generative AI, including its ability to handle vast amounts of data and streamline analytics processes. However, the author also underlines the importance of governance and the potential risks of inaccuracies in AI-generated outputs. Other than that, I find it professional that the author cautiously expressed the current state of AI and its potential impact on organisations.