Ai and Analytics Blog

By; Andrew Tumwesigye
Data and Business Intelligence Analyst, Dimension Metrics

Feb 25th 2025

When I became an analyst, the AI revolution was already in motion. A common narrative, especially on social media, suggested that certain professions—such as data analysis, data science, coding, and app development—were at risk of being replaced by AI. Having completed several certifications and built a promising portfolio, I was confronted with claims that AI could do an analyst's job faster, better, and at no cost. Naturally, this was unsettling.

At first, I was concerned. I had invested significant time, money, and energy into my career, only to hear that my efforts might be in vain. However, rather than succumb to fear, I chose to approach the situation with an open mind. While some argued that AI would replace analysts, others believed that AI was merely a tool that still required human oversight. After all, could AI truly function independently? It needed someone to prompt it, interpret its results, and ensure accuracy.

To objectively assess AI’s capabilities, I decided to compare data analysis with and without AI. Various AI tools, such as ChatGPT, Julius AI, and DeepSeek, are available, but I opted for ChatGPT due to my familiarity with its interface.

AI-Assisted Analysis

Using AI for data analysis proved significantly faster. A simple prompt like “What is the average income by gender?” would generate code and output within seconds, complete with an explanation of the logic behind the code. However, a key limitation became apparent—ChatGPT, despite its accuracy, is a trained model that learns over time.

As a trained analyst, I cross-verified AI-generated code using tools like R Studio or MySQL. On multiple occasions, I found errors in the AI-generated queries—incorrect column selections, misplaced filters, or inaccurate grouping variables. Because I understood data analysis, I could detect and correct these mistakes by refining my prompts. However, an untrained analyst might accept AI’s output without question, potentially leading to flawed business decisions based on inaccurate insights.

Analysis Without AI

Performing data analysis manually, particularly complex queries involving CTEs and joins, took longer—but not drastically so. To validate my work, I would paste my queries into ChatGPT, which helped identify minor errors or optimization issues. While experience had taught me to identify mistakes, AI provided an additional layer of validation. For example, when asked to calculate sales rate by city, I might mistakenly sum sales by city instead of calculating the percentage sales by city. AI could catch such minor errors and guide me in correcting them.

AI Dependency: Potential Pitfalls

Despite AI’s benefits, overreliance can be risky. What happens when the internet is down, electricity is out, or AI servers are overloaded? If an analyst lacks foundational knowledge and depends entirely on AI, they may be unable to complete urgent assignments when AI access is unavailable.

Another limitation is that free AI versions may have outdated information. For instance, if a tool was developed after the AI model's last training update, the AI might not provide an accurate response. Additionally, while AI tools might be able to generate queries and results, analysts must still know how to visualize the results using tools like Tableau, Excel, R, or Python.

The Future of Analysts in an AI-Powered World

I do not believe AI will replace analysts—at least not yet. AI cannot function independently and requires expert prompting. However, analysts who integrate AI into their workflow will outperform those who do not. The real shift will be in efficiency: AI-assisted analysts will deliver faster and more accurate insights.

One valid concern is whether AI will impact salary negotiations. If junior analysts can perform tasks traditionally handled by senior analysts, companies may question the need for higher salaries. While this may reduce senior analysts' bargaining power, it reinforces the importance of continuous learning and adapting to AI-driven advancements.

Conclusion

AI is not the death of analysts—it is a tool that enhances efficiency. Analysts who learn how to leverage AI effectively will have a competitive advantage, while those who resist adaptation may struggle. Rather than replacing analysts, AI will reshape the profession, making it more dynamic and data-driven.