<div class='bc_element' id='bc_element1' style='width:auto;padding:5px;max-height:100%;'><span><p class="isSelectedEnd">Imagine your manager sends you a spreadsheet containing three years of sales data and asks, “Why did sales fall last quarter?” A few years ago, answering that question might have meant cleaning the spreadsheet, building pivot tables, writing SQL queries, comparing regions and products, creating charts and eventually turning everything into a presentation. Today, increasingly, you can upload the data to an AI-enabled analytics tool and simply ask the question. Microsoft’s Power BI Copilot, for example, can answer natural-language questions against a data model, perform ad-hoc calculations and return the result as a visualization. That creates an uncomfortable question for anyone who has spent years learning Excel, SQL, Tableau or Power BI: if AI can increasingly do the mechanical work of analysis, does data analysis become less valuable? The evidence suggests almost the opposite. The World Economic Forum’s <em>Future of Jobs Report 2025</em> found that analytical thinking remains the most sought-after core skill among employers, considered essential by 69% of those surveyed, while AI and big data rank as the fastest-growing skills through 2030. The skill isn’t disappearing. What counts as being good at it is changing.</p><h2>The Technical Work Is Getting Cheaper</h2><p class="isSelectedEnd">For years, learning data analysis often meant learning the tools: Excel, SQL, Tableau, Power BI and, if you went deeper, Python or R. Those skills still matter, but think about what much of that technical knowledge enables you to do. You clean information, filter it, query databases, calculate metrics, compare groups, identify relationships and turn numbers into something people can understand. AI is beginning to assist with more of that process. You can ask it to write a SQL query comparing churn across subscription plans, explain an unfamiliar Excel formula, identify anomalies in a dataset, create a visualization or investigate which customer segment grew fastest. Microsoft is already moving Power BI in precisely this direction; its documentation describes Copilot as allowing users to explore data and generate insights through natural-language questions rather than manually finding or building every visualization.</p><p class="isSelectedEnd">This doesn’t eliminate the analyst. It changes where the analyst creates value. Think about what happened when spreadsheets replaced much of the arithmetic previously done by hand. Companies didn’t stop needing financial analysts because Excel could calculate thousands of cells instantly. Analysts simply stopped creating much value by being the fastest person in the room with a calculator. AI could create a similar shift in data analysis. Producing a chart, calculating a variance or writing a query becomes easier, so knowing which analysis deserves to be produced becomes more valuable. In other words, the skill starts moving away from <strong>“Can you operate the tool?”</strong> and toward <strong>“Can you investigate the problem?”</strong></p><h2>AI Can Give You the Right Number and Still Give You the Wrong Answer</h2><div><br></div><p class="isSelectedEnd">Suppose an AI analyzes your sales data and tells you that sales declined 18% last quarter and that one region accounted for most of the decline. The calculation could be completely correct, but you still don’t know why sales fell. Maybe demand weakened. Maybe prices increased. Perhaps a major customer left, inventory shortages prevented orders from being fulfilled, a competitor entered the market or the company reorganized its sales territories. It could even be something as simple as comparing a seasonal quarter against the wrong baseline. The novice analyst might ask, “What does the data say?” A stronger analyst goes one step further and asks, “What would have to be true for this pattern to exist?”</p><p class="isSelectedEnd">That second question changes the investigation completely. You might compare the affected region with others, examine product-level sales, check pricing changes and inventory availability, separate new customers from existing ones, compare year-over-year rather than quarter-over-quarter performance or speak to the sales team about what changed on the ground. At that point, you’re no longer simply producing charts; you’re investigating a business problem. This is where human judgment becomes particularly important. Microsoft’s 2026 Work Trend Index found that as AI takes on more work, the two human capabilities AI users most frequently identified as becoming more important were quality control of AI output and critical thinking. Microsoft summarizes the broader change as AI taking on more execution while humans gain more room to “direct the work, make the calls, and own the outcomes.”</p><h2>A Dashboard Was Never the Final Product</h2><div><br></div><p class="isSelectedEnd">There’s another trap in data analysis: confusing reporting with analysis. A dashboard might tell you that customer churn increased from 4.2% to 5.8%. That’s useful information, but the business doesn’t really need someone simply to announce that a number changed. It needs someone who can determine why it changed and what should happen next. Perhaps you discover that churn increased mainly among customers acquired through one marketing channel. Dig deeper and you find that those customers received a large introductory discount. Go deeper again and you discover that many cancel as soon as full pricing begins. Now you have something the business can act on: the acquisition channel may be attracting customers with weak long-term purchase intent, or the transition from promotional to full pricing may be poorly designed.</p><p class="isSelectedEnd">That is the real progression of analysis: <strong>data becomes a pattern, the pattern becomes an explanation, the explanation informs a decision, and the decision leads to action.</strong> Historically, analysts could spend a considerable amount of their time simply getting from raw data to the first couple of stages. AI can compress some of that work, giving professionals more opportunity and eventually more pressure to operate further along the chain. The interesting implication is that the future analyst may spend less time building the dashboard and more time asking what the dashboard actually means.</p><p class="isSelectedEnd"><br></p><h2>Easier Analysis Also Means We Can Produce Bad Analysis Faster</h2><div><br></div><p class="isSelectedEnd">There is a downside to making sophisticated analysis available to almost everyone. We can now produce weak conclusions much faster. Imagine an AI discovers that customers who use Feature X are twice as likely to renew their subscriptions. The obvious conclusion is that Feature X improves retention. Maybe it does, but perhaps highly engaged customers are simply more likely to discover Feature X in the first place. Their underlying engagement could be causing both higher Feature X usage and higher retention. If a company mistakes correlation for causation, it might spend six months redesigning onboarding around a feature that wasn’t actually responsible for the result.</p><p class="isSelectedEnd">This is why concepts that once sounded academic, like correlation versus causation, sample bias, confounding variables, statistical significance and experimental design, could become even more practical workplace skills. When analysis required significant effort, there was naturally some friction before someone produced one. AI removes much of that friction. Anyone can potentially ask a dataset a question and receive a polished answer, but not everyone knows whether that answer deserves to be believed. That makes skepticism a feature, not a weakness. A good analyst increasingly needs to ask where the data came from, whether the sample is representative, whether another variable could explain the relationship and what evidence would prove the conclusion wrong.</p><h2>So Should You Still Learn Excel and SQL?</h2><div><br></div><p class="isSelectedEnd">Absolutely, but the reason for learning them is changing. The mistake would be assuming that because AI can write SQL, understanding SQL no longer matters. If an AI-generated query joins the wrong tables, duplicates records or calculates revenue incorrectly, someone still needs enough technical understanding to notice. The same applies to spreadsheet formulas, statistical methods and visualizations. AI can produce an impressive-looking chart without understanding whether the underlying metric makes sense for the business question you’re trying to answer.</p><p class="isSelectedEnd">The goal, however, shouldn’t be to become the person who can write the longest Excel formula without assistance. Learn Excel well enough to understand how data behaves. Learn SQL well enough to understand how information is stored, filtered, joined and aggregated. Learn visualization well enough to recognize when a chart clarifies something and when it distorts it. Learn statistics well enough to challenge an attractive conclusion. Then use AI to accelerate the mechanical work. The future analyst may not be the person who remembers every syntax command; it may be the person who can direct AI toward the right investigation and recognize when it is going in the wrong direction.</p><h2>How Do You Keep Up?</h2><p class="isSelectedEnd">The best way to prepare is to stop treating data analysis as a collection of software certifications and start practicing it as a way of solving problems. Keep the technical foundation, spreadsheets, SQL, basic statistics, visualization and an understanding of how databases work, but add AI-assisted analysis on top of it. Take datasets you already understand and ask AI to analyze them. Then challenge the answer. Ask what assumptions it made, what alternative explanations could produce the same result, what additional data would strengthen the conclusion and what might prove the conclusion wrong. Microsoft’s research offers a useful principle here: 86% of AI users globally said they treat AI output as a starting point rather than the final answer and remain responsible for the thinking.</p><p class="isSelectedEnd">You should also practice starting with messy business questions rather than perfectly structured analytical exercises. Instead of asking, “Calculate customer churn,” start with, “Our retention has deteriorated for six months. What should we investigate before deciding why?” Instead of simply producing sales by region, ask, “One region is underperforming. How would we determine whether the problem is demand, pricing, sales execution or product availability?” Instead of asking AI for a chart, ask what evidence would change a decision. Those exercises force you to think about data in the way businesses actually use it. You’re no longer learning how to manipulate numbers; you’re learning how to think with them.</p><h2>The Future Analyst May Look More Like a Decision Partner</h2><div><br></div><p class="isSelectedEnd">There’s an interesting contradiction emerging. AI is making sophisticated analytical capabilities available to more people, yet employers aren’t signalling that analytical ability is becoming irrelevant. The World Economic Forum expects 39% of workers’ existing skill sets to be transformed or become outdated between 2025 and 2030, while analytical thinking remains the most valued core skill and AI and big data lead the list of fastest-growing skills. That suggests the center of gravity is simply moving. The analyst of yesterday might have been valued primarily for knowing how to extract the numbers. The analyst of tomorrow will increasingly be valued for knowing which numbers matter, what information is missing, whether a conclusion can actually be trusted, what decision the evidence should change and what question should be asked next.</p><p class="isSelectedEnd">This doesn’t mean technical ability becomes irrelevant or that every analyst suddenly turns into a strategist. Someone still needs to understand the data infrastructure, build reliable models and know when an AI-generated answer is technically wrong. What changes is the ceiling. If AI allows you to spend less time manually preparing an analysis, employers can reasonably expect more from the interpretation that follows. Microsoft’s latest research captures this broader transition well: 49% of the Microsoft 365 Copilot conversations it studied involved cognitive work such as analyzing information, solving problems, evaluating ideas and thinking creatively. AI is clearly entering analytical work. The question is no longer whether analysts will use it, but what they will do with the time and capability it gives back.</p><p class="isSelectedEnd"><br></p><h2>The WorkTote Takeaway</h2><div><br></div><p class="isSelectedEnd">This shift also changes how professionals should demonstrate data-analysis ability. Writing <strong>“Excel, SQL, Power BI, Data Analysis”</strong> in the skills section of a resume tells an employer which tools you’ve encountered, but it doesn’t show whether you can use data to solve a problem. The stronger evidence might be a dashboard that exposed an operational bottleneck, a pricing analysis that changed a recommendation, a forecasting model that improved planning, a customer analysis that revealed why people were leaving or a presentation that turned thousands of rows of data into three decisions leadership could actually act on.</p><p>That’s where WorkTote’s approach fits naturally. <strong>Content</strong> lets professionals bring together the projects, presentations, reports, case studies, certifications and other work that demonstrates what they have actually done. That evidence can contribute to a richer <strong>Digital Twin</strong> of their professional experience, while <strong>Career Trajectory</strong> can help connect existing capabilities with where those capabilities might take them next. As AI makes analytical tools easier to use, simply saying <strong>“I know data analysis”</strong> will tell employers less and less. The more revealing question will be: <strong>What problems have you actually solved with it?<br></strong></p> <span></div>