The Evolution of Analytics Execution: From BI to AI Agents

Driving the Analytics of 2030, not 2020

Part of the Driving the Analytics of 2030, Not 2020 series.

Having covered how analytics permeates every function, the question becomes: how do organizations execute analytics work, and how is that changing? This is the Art of Analytics. The journey looks like this:

1. Enterprise Reporting
2. Self-Service Analytics
3. AI-Assisted Analytics
4. Decision Support & Agentic AI

Here’s how the evolution breaks down.

Enterprise Reporting: The Starting Point (Yesterday and Today)

In many organizations, analytics started (and sometimes still ends) with Enterprise Reporting: a centralized team (often in IT or Finance) producing standard reports and dashboards. This is the era of static monthly reports, PDFs emailed around, or big Business Intelligence platforms managed by a few analysts. It’s valuable, but has limitations: it’s slow, not customized to each decision-maker, and can become a crutch (everyone waits for “the report” instead of exploring data themselves).

Self-Service Analytics

Self-Service Analytics emerged to address this. Tools like Tableau, Power BI, and Qlik democratized data visualization by enabling non-technical users to drag and drop their way to insights. The idea was to empower the masses to create their own charts and queries without always going to IT. Culturally, this was a shift: business users taking analytics into their own hands. By 2025, Gartner estimated 90% of analytics consumers will become creators thanks to AI. Self-Service Analytics has been mainstream since 2010, and by 2030 it will be expected table stakes, similar to how everyone today uses office productivity software.

Self-service alone isn’t enough if people don’t know what to ask or how to interpret data. That’s where the next stage comes in.

A Career Spanning the Full Arc

My career has tracked this entire evolution firsthand. In 1996, I started at CERN writing neural network ensembles for particle identification — enterprise reporting was all we had. At eBay (2004–2012), I lived the transition to self-service analytics, building the Flight Deck reporting system and democratizing data across nine roles over eight years. At Google/Motorola, I built MotoInsights, a Google Drive-based analytics tool that hit 25% monthly adoption — self-service in practice. At Change Healthcare (2020–2023), I launched the Analytics Academy that took business teams from filing tickets to querying data directly — the bridge to AI-assisted analytics. At Optum/UHG (2023–2025), I fused LangChain, GPT-4, and data platforms into the Single Pane of Glass, enabling executives to ask questions in natural language — AI-assisted analytics deployed. Now at Blue Fermion, I’m shipping AgentResume.ai, Demeterics (LLM gateway for 8+ providers), and BlueVideo.ai (autonomous video generation) — agentic AI in production. This progression from reporting to agents isn’t theoretical; it’s the path organizations are on, compressed into one career.

Next in series: AI-Assisted Analytics: BI Meets AI (Today and Tomorrow)