AI and the Rise of the Citizen Analyst
Unlocking the Last Mile of Business Intelligence (Finally)
Twenty years of data innovation solved the infrastructure layer brilliantly. We can store petabytes, process at lightning speed, build perfect schemas. But all that innovation primarily serves the data team.
Meanwhile, your marketing manager exports to Excel. Your ops lead maintains shadow spreadsheets. Your sales director makes million-dollar decisions from last Tuesday’s CSV.
We optimized for scale and sophistication, but forgot about the humans who need answers. AI changes this equation. When anyone can interrogate data in their own words, the infrastructure finally serves its purpose: empowering decisions, not gatekeeping them.
The New Breed of Analyst
Similar to the Citizen Developer, we think it’s time for the rise of the Citizen Analyst.
Citizen Analysts have always been around, even if we didn’t call them that. You might know them as “Power Users” or “Spreadsheet Jockeys” (a cheeky, sometimes derisive nickname). These are the people who work with data every day, though “data” isn’t normally in their job title.
The Marketing Manager
She knows this campaign is underperforming. The CMO wants numbers: “What’s the ROI by segment?” She submits a ticket. Two weeks later, the report arrives. The campaign already ended.
With instant answers? She’s adjusting spend mid-flight, testing hypotheses during the Monday standup.
The Operations Leader
Twenty years of experience says Tuesday’s delay will cascade. But spreadsheet models take days to build, and by then the damage is done. He needs to quantify the hunch NOW: which SKUs, what revenue impact, which customers. Three systems, three exports, one very late night.
With all his data in one place? He quantifies hunches in minutes, not days.
The Customer Success Manager
This enterprise account is about to churn. Support tickets up 40%. Champion left. Usage dropping. But she can’t get the data to prove it. The BI dashboard only shows quarterly aggregates, and the data team is booked solid. “Gut feel” doesn’t mobilize executive intervention. Another logo lost to preventable churn.
With cross-system visibility? She spots the warning signs and mobilizes the team while there’s still time.
The Reality of “Good Enough” Data
Here’s what data teams often miss: business users are already making critical decisions on “directionally accurate” data. That Excel export? It’s probably 95% correct. That week-old CSV? Still relevant enough for most decisions.
When I was at Heap Analytics, we discovered something telling: the vast majority of queries were on recent data, usually less than 90 days old. Yet customers insisted on retaining years of data, paying premium prices for storage they rarely touched. The allure of “Big Data” is real, but the actual usage? Intensely focused on the present.
Business users understand this intuitively. They’ll take directionally correct answers today over precisely perfect answers next quarter. They know their Excel pivot table isn’t capturing every edge case, but it’s good enough to spot the trend, make the call, and move forward.
The dirty secret of shadow IT isn’t that people are rebels. It’s that they’re pragmatists. Speed beats precision for 90% of business decisions.
Beyond SQL: AI as Your Thought Partner
These pragmatists don’t need another BI tool. They need a thought partner that speaks their language.
Asking questions in plain English is just the starting point. The real magic happens in the conversation that follows.
AI doesn’t just translate your question into SQL and serve up an answer. It suggests correlations you hadn’t considered. “You’re looking at churn by segment, but have you noticed the pattern with contract length?” Every interaction adds context. When you say “key accounts,” it knows you mean the Fortune 500 customers, not just the high-revenue ones.
It maps your data model without you having to explain it. Spots the relationships between tables that would take a data engineer hours to find. Connects your “customer_id” in one system to “account_number” in another.
This isn’t about eliminating thinking. It’s about amplifying it by 10x. When the marketing manager asks about campaign performance, the AI doesn’t just return numbers. It surfaces the seasonal pattern from last year, suggests segmenting by customer lifetime value, and excludes test accounts like she always needs.
The technical barrier that kept citizen analysts from their data? Gone. But more importantly, the discovery barrier is gone too. Every question becomes a learning opportunity. Every answer leads to smarter questions. The AI becomes a thought partner, not just a query engine.
The Fifth Question
The best insights come from the fifth question, not the first. That’s when you’ve moved past the obvious and started interrogating the data, following your curiosity wherever it leads.
Watch how a real exploration unfolds:
- “Show churn by segment.”
- “Now just enterprise accounts.”
- “Add product usage for the last 30 days.”
- “Correlate with support tickets.”
- “Which features aren’t they using?”
Five iterations. Thirty seconds. Each question building on the last, each answer sparking new curiosity. This is interrogation, not reporting. It’s the difference between reading a map and actually exploring the territory.
In traditional BI, each iteration is a new ticket. A new wait. A new explanation of context. By the time you get to question five, you’ve forgotten why you asked question one.
Dashboards Are Dead. Long Live Dashboards!
Let me be clear: dashboards are perfect for what they’re designed for. Monitoring KPIs, tracking trends, keeping a pulse on the business. But we’ve made them the default answer to every question, and that’s where we went wrong.
Most questions don’t deserve a dashboard. They deserve an answer. Today’s curiosity about conversion rates by geography shouldn’t become next quarter’s ignored widget on page 47 of your BI portal.
Here’s the irony: the more ephemeral the question, the more flexibility you need to explore it. That one-off analysis about seasonal patterns? It needs more slicing, dicing, and interrogating than your permanent revenue dashboard. Yet we force these exploratory questions into rigid dashboard frameworks, then wonder why adoption is so low.
Dashboards should be like milk, not wine. Most should expire after they’ve served their purpose.
The Build-As-You-Go Philosophy
Your citizen analysts have been living this philosophy all along. Watch them work in Excel: they start with one question, build a simple pivot table, then add a lookup, then another data source, then a calculation. Their model evolves with their understanding.
Traditional BI fights this instinct. Six months of data modeling before anyone can ask a question. Perfect schemas that assume you know every question in advance.
But your citizen analysts know better. They know that Mike Tyson was right: “Everyone has a plan until they get punched in the mouth.” That quarterly forecast model? It worked great until the market shifted. That customer segmentation? Perfect until the new product launched.
They need tools that work like they do: start simple, evolve with discovery, adapt to reality. The best data model is the one that emerges from actual questions, not theoretical requirements.
The Compound Effect of Curiosity
When ServiceNow democratized workflow automation, something unexpected happened. It wasn’t just that more workflows got automated. People started solving problems that were too small for IT’s roadmap but too painful to ignore. The thousand paper cuts that slow down every organization suddenly had a cure.
The same thing happens with data.
When ten people can interrogate data, you don’t get ten times the insights. You get exponential returns. The marketing manager discovers a spike in demo requests every third Tuesday. She shares it in Slack. Sales realizes those are the days their competitor’s free trial expires. Ops starts staffing up support on Wednesdays. Product adjusts the onboarding flow to capitalize on switchers. What started as one person noticing a weird pattern becomes competitive advantage.
The real metrics aren’t about speed or user counts. They’re about decisions that couldn’t happen before:
- The regional pricing experiment that would have taken six months of analysis to approve
- The churn prevented because someone could finally correlate support tickets with renewal dates
- The campaign pivot that happened mid-week instead of post-mortem
You can’t measure the value of questions that were never asked.
What This Means for Data Teams
Data teams aren’t disappearing—they’re evolving into more strategic and impactful roles. Rather than spending time writing queries, they are:
- Teaching AI the business context and rules
- Ensuring data quality and governance
- Building advanced models for AI to leverage
- Empowering citizen analysts to ask sharper, more insightful questions
AI elevates the data team’s role from gatekeepers to enablers of innovation and insight, allowing them to focus on high-value, strategic work.
The StarLifter Difference
We built StarLifter for how citizen analysts actually work:
- Questions become conversations: Not just answers—insights, context, and suggested next steps
- All your data, one place: Connect your warehouse, SaaS tools, spreadsheets. No multi-quarter project required
- Explore without consequences: Every question builds on the last. No dashboards to break
- Your business, your language: It understands that “enterprise” means Fortune 500 to you, not just high-revenue
- Built for iteration: Start simple, get smarter with every question. No perfect schemas required
We didn’t adapt old BI tools to AI. We built analytics for people who think in business outcomes, not database schemas.
The Choice Is Now
While you’re waiting for perfect data models, your competitors are asking imperfect questions and getting directionally correct answers. Today.
Your citizen analysts are already finding insights. They’re just burning hours in Excel to get there. The marketing manager spending her Sunday building pivot tables. The ops leader merging three datasets for one question. The customer success manager with twenty browser tabs trying to connect the dots.
Your citizen analysts are your secret weapon. They know your business better than any data scientist ever could. They spot patterns your dashboards miss. They ask questions your data team wouldn’t think of.
Right now they’re heroes working with broken tools. Imagine what they could do with the right ones.
The compound effect we talked about? It’s waiting to happen. It starts the moment every question becomes a conversation instead of a project.