Analytics Is Leaving the Dashboard. The 2030 Analyst Is Already Taking Shape.
Part 1 of 4: Where analytics goes next
Picture a stakeholder in 2030. She needs to know what a two-point swing in interest rates does to next year’s capital position. She does not open a dashboard, or email an analyst and wait for a deck. She asks an agent, software that carries out a task on its own, for forty scenarios, reads the three that look odd, and calls the analyst to ask whether the odd ones can be trusted.
The analyst’s work in that story is real work. It just looks nothing like the work most of us trained for. There is no pivot table and no formatting pass, only a judgment call about a system, made by someone who knows the business well enough to smell a wrong answer.
Analytics is in the middle of a major shift. Artificial intelligence is redefining what it means to be an analyst, how teams are structured, and what value they deliver. The first wave changed how fast an analysis gets built. The waves still coming change who asks for it, who checks it, and who answers for it. Paulo Coelho wrote, “If you think adventure is dangerous, try routine. It is lethal.” For an analytics team, the routine is the risk.
The scene above is one plausible version of 2030, not a prediction. In February 2026 I used scenes like it to walk an audience of analytics colleagues through where our work could go. This four-part series revisits those scenarios against the evidence that has arrived since. Part 1 sets the frame and checks Horizon 1 with two months of 2026 left. Parts 2 and 3 cover the next two horizons. Part 4 is the plan.
The technology is here and early value is showing, but the operating model has not caught up. Closing that gap is mostly organizational work, and it is cheapest to start now. A four-question self-check at the end of this part shows where you stand.
Approximately 1,940 words · 10 minute read
Published October 23, 2026
Plausible, Not Guaranteed
Nobody can tell you what 2030 looks like. A leader still has to decide this year who to hire and what to buy. Waiting for certainty is still a decision, and usually an expensive one.
Scenarios exist for that situation. A forecast tries to name the one future that will happen. A scenario describes a future that could happen, in enough detail that a team can ask what it would do if it arrived. I built these around four themes and three time horizons, and treated every outcome as plausible rather than certain.
The four themes:
- Landscape: What is happening around us. Market forces, artificial intelligence and analytics trends, regulatory and competitive shifts.
- Design: How we structure to respond. Team design, operating model, where analytics sits, governance, decision rights.
- Skills: Who we need to become. Analyst and leader capabilities.
- Technology: What we build on. Data platforms, semantic layers (shared business definitions), agents, tools, automation.
The three horizons:
- Horizon 1 is already happening and runs through the next two to three years, 2026 to 2028.
- Horizon 2 is emerging now and goes mainstream in three to seven years, 2028 to 2032.
- Horizon 3 is experimental today and becomes transformational in seven to fifteen years, 2032 and beyond.
The timing estimates are conservative. The technology moves fast enough that they may shift forward.
The four themes are important because they move at different speeds. Technology usually leads. Landscape follows as vendors ship it into the tools people already use. Design lags, because org charts and decision rights change slowly. Skills lag furthest, because people take years to build and months to be told they need something new. Most of the friction organizations feel with the technology sits in the gaps between those speeds.
Each horizon shows which of the four is currently the bottleneck, and where a decision made now is cheap compared with the same decision in two years.
Horizon 1: 2026 to 2028
Leading organizations already show these trends. The technology is largely proven and the skills gap is clear.
Landscape: analytics as embedded organizational intelligence. Insight stops being a destination like a dashboard. It shows up inside Outlook, pricing tools, policy admin systems, and CRM. People no longer go to analytics. Analytics comes to them.
Design: continuous intelligence operations. An always-on layer monitors the business, flags issues and opportunities as they emerge, and triggers automated or semi-automated responses. Reporting becomes a 24/7 service instead of a monthly deliverable. The question “what happened last quarter?” gives way to “what changed this morning, and what should we do about it?”
Skills: the analyst as chief curiosity officer. Execution is largely automated. Humans specialize in asking novel, high-leverage questions and in challenging what the model assumed. The scarcest skill is framing the question that creates a new option for the business, and that skill has little to do with technical depth. A model can answer a well-formed question quickly. Someone still has to notice that the question was the wrong one.
Technology: simulation-assisted decisioning. Agents orchestrate simulation platforms. Running thousands of capital, pricing, or demand scenarios becomes as routine as building a pivot table. A decision that used to rest on three scenarios chosen by a committee can rest on three thousand, with the analyst’s job shifting to deciding which of them deserve attention.
Horizon 1 lays the foundation for trustworthy analytics that agents can build on and moves teams from reporting to decision support.
Where We Are: Horizon 1 With Two Months Left
Horizon 1 spans three years. With two months of 2026 remaining, one of those years is nearly spent. The evidence below comes from surveys of senior leaders by McKinsey, Deloitte, and IBM, plus Gartner forecasts. The surveying firms all sell services in this area and the surveys are self-reported, so treat them as direction.
Landscape: moving fastest at large enterprises
Gartner forecast in 2025 that 40% of enterprise applications would include task-specific agents by the end of 2026, up from under 5% in 2025. On the buyer side, McKinsey’s August 2026 survey found that 40% of respondents at large organizations are scaling agents, up from 27% a year earlier. Smaller organizations stayed flat at 22%.
Gartner’s figure counts applications, while McKinsey counts organizations, and I found no survey that checks whether the 40% was reached. They still point one way: agents are arriving inside the tools, and company size decides how fast. A large insurer or bank likely sits well into Horizon 1. A mid-sized firm may still be waiting for its vendors to ship it.
Design: early
Adoption plans are running ahead of controls. Deloitte surveyed 3,235 leaders in 24 countries and found that only 21% of organizations have a mature governance model for agents, even though 74% expect to use agents at least moderately by 2027. IBM’s survey, admittedly a bit dated now as of November 2025, of 1,700 data and analytics leaders found that 79% say they are early in defining how to scale and govern agents.
McKinsey found that about 6% of respondents qualify as high performers (McKinsey’s term for those attributing at least 5% of EBIT to artificial intelligence), the same share as last year. Nearly three-quarters of those high performers have fundamentally redesigned workflows around the technology. Only about one quarter of everyone else has. An agent layered onto an unchanged process just speeds up the old process.
Skills: too early to measure
McKinsey found a split in who feels the strain. Among mid-level managers and individual contributors, 47% report at least one negative effect, against 31% of executives and senior managers. The people closest to the work carry more of the adjustment cost, and they are the same people Horizon 1 asks to become chief curiosity officers.
No survey I found measures question framing directly, which is the skill Horizon 1 says will be scarcest. The closest signals are in hiring. IBM found that 77% of data leaders struggle to fill key data roles, and Gartner expects 75% of hiring processes to test candidates on how well they work with the technology by 2027. Employers are starting to measure skills, just not this one. A skill nobody measures is hard to build on purpose. Part 4 covers where to start.
Technology: capable, thin at the foundation
Gartner’s survey, published in 2025, found that 63% of organizations lack, or are unsure they have, the right data management practices for artificial intelligence. Gartner also predicted that through 2026, organizations would abandon 60% of projects that lack data ready for the technology. The year ends in two months, so you can check that prediction against your own portfolio. The same IBM survey found that only 26% are confident their data can support new revenue streams enabled by the technology.
Cost is now a factor (though maybe it always should have been). About 20% of McKinsey respondents say operating costs, including token costs, have limited their use of the technology. Running thousands of scenarios is cheap in concept and meaningful in a cloud bill.
Simulation-assisted decisioning is the one Horizon 1 theme where I found no survey data, so treat it as the least proven.
The value gap is still the headline
McKinsey reports that 37% of respondents attribute some EBIT (operating profit) impact to artificial intelligence, essentially unchanged from last year even as scaling grew. Individuals report real gains: 80% of respondents say the technology improved their own productivity, and about half say it helps them make better decisions.
This is my inference is that value shows up first in individual speed and decisions and reaches the P&L later. That is the Horizon 1 shift from reporting to decision support, early but real. Measurement may be part of the gap. IBM found that just 29% of data leaders have clear measures for the value of data-driven outcomes. If the P&L number stays flat for another two years, the case for the next round of spending gets harder to make.
Headcount shows a similar gap between story and result. A third of last year’s respondents expected the technology to shrink headcount. Only 14% report that it actually did. In this year’s survey, 39% expect a decline over the next year.
The scorecard
- Landscape: Moving at large organizations (40% scaling agents), uneven elsewhere.
- Design: Early, with governance lagging adoption plans.
- Skills: Demand is clear, and question framing is still unmeasured.
- Technology: Capable, with data readiness the weak point.
Score your own organization. Answer yes or no:
- Landscape: Can you name the two or three decisions where analytics should show up inside the tools your people already use?
- Design: Has at least one workflow been rebuilt for agents, instead of an agent being added to the old steps?
- Skills: Do you know which people are raising the questions that change decisions?
- Technology: Does each of your ten most important metrics have one definition and one owner?
Count the yeses. Two or fewer means the foundation work in Part 4 comes first.
What Comes Next
Horizon 1 is the one still open for decisions. The next one is where the analyst’s job changes shape, and where accountability becomes a real question. Part 2 looks at what happens when agents do the first pass and stakeholders ask for scenarios instead of decks.