We Have Power BI, But Everyone Still Exports to Excel – Here's Why

By the team at Human Ready · Updated July 2026

When a company has Power BI but everyone still exports to Excel, the cause is usually not training, dashboard design, or user stubbornness – it's that dashboards answer the questions someone already anticipated, while the questions people actually have keep changing. The export to Excel is the moment a user hits a question the dashboard wasn't built for and reaches for the only tool that lets them ask it. That makes Excel exports an adoption signal, not a formatting preference – and this page explains what the signal means, why BI adoption stalls at this exact point, and what real self-service analytics requires.

Why does everyone still export to Excel when we have Power BI?

People export to Excel because the dashboard shows them a fixed view, and their next question doesn't fit it – Excel is where they can keep asking. You can see the pattern in the BI community's own words: Power BI practitioners on Reddit ask "Why are Excel report requests so common?" and "Why Power BI? when all you need is a report in Excel" – threads full of BI developers whose users have one dashboard and one request: could you send me this data in Excel?

Finance leaders feel it from the other side. One group CFO – relayed verbatim by a finance lead who works for him – put it like this:

"Isto aqui fala tudo de AI, fala tudo de AI, mas eu continuo a ver toda a gente entrar no Excel e eu fico maluco com isto. E estava à espera que alguém me dê uma solução." Everything here talks about AI – everything talks about AI – but I still see everyone going into Excel, and it drives me crazy. And I was waiting for someone to give me a solution.

The scale of the phenomenon is well documented:

  • In the 2025 AFP FP&A Benchmarking Survey (362 practitioners), 96% of FP&A professionals use spreadsheets for planning and 93% use them for reporting on a daily or weekly basis – numbers that AFP notes hold across company size, geography, and seniority.
  • The 2024 FP&A Trends Survey found 52% of FP&A teams still use Excel as their planning tool, years into the modern BI era.

Spreadsheets survive every BI rollout for one reason: Excel is the only layer in the stack where the user can ask a question nobody predicted.

Is low BI adoption normal?

Yes – stalled BI adoption is the industry norm, not a local failure. BARC's multi-year adoption research finds the average adoption rate of BI and analytics tools has been "stuck around 20% for many years," with the share of employees actively using BI currently about 25% on average – minimal growth in the seven years BARC has tracked it. Gartner data reported by IBM tells the same story: analytics and BI usage grew in 87% of surveyed organisations, yet is still used by only 29% of employees on average.

Sit with that: after two decades of BI investment, roughly three out of four employees don't actively use the tools. The dominant response – build more dashboards, run more training – hasn't moved the number, because it treats the export as a usage problem. It isn't. It's a question problem.

What is the Excel export actually telling you?

The export is telling you that your users have tier-three questions and tier-one tools. A finance-and-strategy executive who has run controlling across several groups gave us the sharpest version of this diagnosis:

"As pessoas querem usar dashboards para fazer análise — mas os dashboards são o que já conhecemos. Análise pressupõe novas perguntas que ainda não estavam pensadas." People want to use dashboards to do analysis – but dashboards are what we already know. Analysis assumes new questions that weren't thought of yet.

In her framing there are three tiers of data consumption: the daily operational view, the periodic monitoring of known variables, and ad-hoc analysis – the new question no view was built for. Dashboards excel at tiers one and two. The Excel export marks the exact border where tier three begins.

Two more voices from our buyer conversations complete the picture. The head of FP&A at a multi-country healthcare group – a Power BI shop – described how the actual analysis gets done: "This insight today is very manual. It's sitting every Monday with five analysts... doing queries, having hypotheses, drawing conclusions. There are questions I just have no answers to." The dashboards run; the analysis happens next to them, by hand. A CFO at a multinational healthcare provider, on the output of his internal BI department, said: "trata muito de questões históricas, olha para o histórico e não projeta futuro, nem dá pistas para melhorias" – it deals mostly with historical questions; it looks at the past, doesn't project the future, and doesn't offer clues for improvement.

That's the ceiling in one sentence. Dashboards answer yesterday's questions. The business runs on tomorrow's.

What does real self-service analytics require?

Real self-service requires that a business user can ask a new question of governed data and get a trustworthy, finished analysis – not a data dump they must finish themselves in Excel. Concretely, four requirements, and most "self-service BI" deployments meet none of them:

  1. Questions in plain language, not navigation. If self-service means "find the right dashboard among 40, then filter," users will export. The interaction has to start from the question, not the view. (This is the search-not-browse shift – the full argument is in the hub essay: why every new finance question takes a week.)
  2. Analysis as the output, not rows. An export happens because the dashboard gives data when the user needs an answer – a comparison, a driver breakdown, a variance explained. Self-service that stops at the extract just relocates the analyst work to the user.
  3. Every number traceable. Finance users re-do analysis in Excel partly because there they can see every formula. Any layer that answers new questions automatically must offer the same: a visible path from answer back to source data. Without it, trust – and the user – goes back to the spreadsheet. What that looks like in practice: how AI analytics can show exactly how it got the number.
  4. Power BI stays. The goal is not to replace the cockpit – dashboards are the right tool for the daily and monthly views. The goal is to stop asking the cockpit to do analysis. An IT director at an industrial company, mid-rollout on his own modern data platform, framed the addition not as replacing his warehouse or BI but as the next presentation layer – "o BI do futuro," the BI of the future.

Get those four right and the Excel export loses its job: the new question gets answered where the data lives, at the speed the meeting requires.

Where does Human Ready Advisor fit?

Human Ready Advisor is the tier-three layer described above: an AI-native advisory platform that sits on top of your existing stack – Power BI included – and answers new finance questions in plain language, with every number traceable back to how it was produced. Your dashboards keep the cockpit job; Advisor takes the question that was never going to be on a dashboard. If your team ships beautiful dashboards and still watches everyone go into Excel, that's the gap we work on – humanready.io.

Related reading in this series:

Buyer quotes come from Human Ready's conversations with finance leaders at European mid-market and enterprise companies, anonymised to role and company profile. Portuguese quotes are translated faithfully.