In my consulting engagements, I’ve used generative AI for market sizing, business modeling, product naming, and growth forecasting. I can’t think of an analytics application more naturally suited to AI than customer segmentation.
The latest AI systems automatically coordinate multiple specialized models to tackle complex problems, combining best-of-breed language models with specialized tools for search, analysis, coding, and reasoning. The industry has realized that no single AI excels at everything. The same concept applies to the gen AI tools themselves. Here are three reasons for making a practice of using more than one generative AI tool.
For the past few decades, SAS, Python and R have dominated exploratory analytics and modeling, supplemented by a progression of interactive tools, starting with OLAP in the 1990’s and progressing to visual tools like Tableau in recent years. The most productive versions of these solutions still required weeks of researching data sources, downloading them (often for a fee), multiple trial-and-error cycles on quality, and wrestling with mismatched keys across datasets. Once the data was ready, producing analysis required laborious hand-coding to produce and present insights.
Since ChatGPT launched just 30 months ago, leaders have been asking if it’s fundamentally changing how customer analytics works. The short answer is: yes. While analytics has used AI for decades in the form of regression models and predictive algorithms, what I’ve found through recent client work is striking: when wielded by skilled practitioners, generative language models don’t replace human insight – they amplify it. These tools apply in numerous and rapidly growing ways that aren’t immediately obvious to everyone. The key difference? Gen AI dramatically shortens the path from question to insight, allowing us to focus more on the “so what” than the “how to.”
Gen AI tools can make things up with stunning confidence. In researching a book I’m working on right now, I recently prompted an AI engine to help me find some supporting points. I specifically asked it to use only credible sources and provide citations. I was delighted to receive 5 relevant article titles, cited from publications that included WSJ, New York Times, and Forbes. Unfortunately, none of the stories were real. AI made up the stories, the titles, and even the hyperlinks.
I used to track nutrition with MyFitnessPal, but recently switched to a dedicated project in Claude. Typing in foods is faster than searching a database in an app, but even that feels like work. So I snapped this quick picture of my meal to see how well AI tools could itemize what I ate.
This experiment compares ChatGPT-4 and Claude 3.5 Sonnet to analyze personal fitness journey data. It evaluates their strengths in data visualization and question formulation, proposing that their combined capabilities could enhance data analysis. The findings offer valuable insights for data professionals, business leaders, and AI enthusiasts.
Generative AI keeps accelerating. You ask a question, within seconds you have an answer. It’s based on the worlds collective knowledge. New insight sparks new lines of thinking. You prompt for even more context. With answers in hand you set an action plan.
