What Business Intelligence Actually Offered

I came to Business Intelligence after the TM/1 years with a simple desire: stop drowning in reconciliations and start swimming in reasons. The promise was seductive. BI said we could wire the brain of the company directly to its senses. No more shipping the past around in spreadsheets; the present would present itself on demand.

At the coalface, the offer was concrete. Connectivity: direct lines into ERP, CRM, POS, data warehouses. Modeling: reusable semantic layers so a “margin” in sales meant the same “margin” in finance. Speed: scheduled refreshes and incremental loads that shrank yesterday into minutes. Exploration: drill‑down, drill‑up, drill‑across—time, geography, product, customer. Communication: dashboards that could travel from boardroom to branch without losing meaning.

When I sold Brio, Business Objects, and Cognos in the United States, I watched CFOs discover that a KPI isn’t a number—it’s a narrative. A red arrow didn’t simply accuse; it invited inquiry. Suddenly a week’s work for a reporting analyst could be compressed to an afternoon of thinking. Sales leakage could be traced to a price list version; a rebate rule could be separated from the fog of totals; a channel could be seen as a portfolio of micro‑behaviours rather than a single line in a P&L.

But a paradox emerged. The more power the platforms offered, the more timid the usage became. Dashboards got beautiful and the questions got small. Enterprises were buying telescopes and then using them as paperweights. The barrier wasn’t UI or data quality or compute. The barrier was the human reflex that had been trained by decades of month‑end: prove, do not probe.

BI, at its best, offered a company five structural upgrades: (1) Shared meaning through governed metrics; (2) Temporal awareness via history alongside now; (3) Comparative context—peers, regions, cohorts; (4) Scale—millions of records with sub‑second response; (5) Reusability—analytical assets as living components. The tragedy is that most organisations used (5) to automate yesterday, not to invent tomorrow. That was when I realised: we didn’t have a data problem. We had a question problem.

The Question Famine: Why Most People Don’t Know What to Ask

We hire for answers and then wonder why no one brings us questions. School drills us to converge. Corporate life rewards certainty. Compliance cultures equate doubt with danger. Over years, the muscle that generates inquiry atrophies. By the time someone is entrusted with a BI license, their curiosity has been professionalised away.

Training

Accountants are trained to close gaps, not to open them. The month‑end ritual is an exorcism of ambiguity. This conditioning is necessary for stewardship—but it is hostile to discovery. Exploration looks like error inside a calendar built from deadlines. Give that context a multidimensional tool and it will be wielded unidimensionally.

Psychological Safety

Questions are vulnerability in public. In many companies, the penalty for a failed experiment is private shame or public scapegoating. The safe path is to excel at process, not to improvise with purpose. Curiosity wilts where courage is taxed.

Biology

The brain is an energy miser. Pattern recognition is cheaper than hypothesis generation. It takes biochemical investment to hold a counterfactual in working memory and walk it through consequences. We prefer grooves to gradients. The default human setting is conservation, not curiosity.

Structure

Monolithic systems—technical and managerial—discipline thought into workflows. When all roads lead to the same committee, variance becomes a moral failing. Meanwhile entrepreneurs—who never received permission—prototype their way through the problem space, guided by feedback, not fear.

Every meaningful innovation begins life as a question that sounds naïve to the incumbent and necessary to the insurgent.

Is curiosity DNA? Partly temperamental, yes. But more often it is a cultivated stance. Entrepreneurs learn to metabolise uncertainty; corporates are medicated against it. The difference is not intellect—it is incentives. One is paid to extend the known; the other survives by trespassing into the unknown.

Enter AI. For the first time, the scarcest resource is not compute or storage but the quality of a prompt. A mediocre question produces a transcript. A precise, daring, contextualised question produces strategy. The playing field has tilted. Capital is no longer the moat; curiosity is.

Corporate Mindset vs Entrepreneurial Instinct

The Corporate Mandate: Minimise Variance

Industrial logic built corporate power: replicate, standardise, control. In that world, the champion operator is the one who eliminates surprise. Information technology joined the assembly line and promised the final victory over noise. Data warehouses, ERPs, and BI formed a panopticon of performance. But surveillance is not the same as sense‑making. You can measure a million KPIs and still miss the causal hinge.

Inside these architectures, people become guardians of cadence. Schedules, not hypotheses, govern the week. Curiosity is a cost centre. The safest employee is the one who never trips the alarm.

The Entrepreneur’s Contract: Maximise Learning

Entrepreneurs operate under different physics: speed of feedback beats size of budget. They practice empirical curiosity—cheap trials, rapid inference, modular bets. Failure is tuition, not termination. Their systems resemble Lego: small blocks, clean interfaces, constant recombination. Replace a piece, keep the whole.

From Monoliths to Modules

In the 2000s, banks erected monoliths—one system to span the enterprise. Change requests were surgical operations. Startups built constellations—microservices stitched with APIs. When a block underperformed, it was swapped. When customers asked for something weird, a new block appeared. Modularity outran magnitude. By the time the fortress noticed, the field had moved.

The Neuroscience of Curiosity at Work

Dopamine rewards the hunt. Cortisol punishes exposure. Org cultures choose the hormone they issue most. Where questions are taxed, cortisol floods and the prefrontal cortex narrows options. Where questions are welcomed, dopamine sponsors exploration and the map expands. Strategy is chemistry in a suit.

AI as Equaliser

Artificial Intelligence is insensitive to status. It pays only for precision of intent. A two‑person shop with clear prompts and messy data will often out‑perform a thousand‑person division with immaculate data and vague intent. The moat moves from infrastructure to interrogation.

From Pivot Tables to Prompts — Turning Questions into Capital

The pivot table was our apprenticeship. It taught us to rotate a fact and see its faces. Useful, yes—but planar. We were still confined to the schema we inherited. Then the interface changed. We stopped dragging fields and started drafting intents.

Dialogue, Not Manipulation

BI made us operators. AI makes us partners. The work is no longer “Rows, Columns, Values” but “Explore the interaction between discount depth and customer lifetime value; include seasonality and cross‑category cannibalisation; return three counterintuitive hypotheses.” Machines now propose, not merely expose.

Prompt Craft = Thought Craft

A competent prompt encodes five elements: context (where we are), intent (what we seek), constraints (what to ignore), criteria (how we’ll judge), and continuations (what to try next). Teach a team those five and you upgrade the company’s collective IQ without buying a single new server.

The Lego Continuum

Modern analytics is a kit: LLM for language, classical ML for prediction, optimiser for allocation, rules engine for guardrails, vector store for memory. Each block is modest; the composition is magic. Questions become workflows; workflows become advantages.

Questions on the Balance Sheet

Track them. Question velocity (how many distinct hypotheses explored per week). Time to insight (from prompt to decision). Hit rate (experiments adopted). Idea diversity (functions involved). Curiosity becomes accountable—and therefore fundable.

Thought Governance

Data governance was table stakes. Now we need thought governance: an indexed library of prompts, decisions, and outcomes. Over months it forms a question graph—a living memory of how the firm learns. New hires onboard into the mind of the organisation, not just its drives.

The Return of Judgment

As automation blooms, the scarcest asset is discernment. AI can enumerate possibilities; only people can choose which are worth becoming. The work of leadership shifts from requesting reports to orchestrating inquiry.

Practical Pattern: From Report Factory to Curiosity Studio

Week 1: sunset unused dashboards; create a questions backlog. Week 2: establish a prompt guild; publish the five‑element template. Week 3: institute “Friday experiments” with tiny budgets. Week 4: convert one recurring report into a living conversation (“explain, simulate, recommend, challenge”). Quarter 2: measure question velocity; fund the teams that learn fastest.

Closing

We thought data would save us. It didn’t. Questions will. The tools have finally caught up with the mind that dares to use them. From TM/1 to BI to AI, the arc is simple: reconcile → reveal → rethink. Those who master the last verb will own the decade.