That is the setting and it remains to be seen as to where the crush will end up being. This morning I was surprised by a story in CDOTrends (at https://www.cdotrends.com/story/4729/how-agentic-analytics-replacing-bi-we-know-it) where Artyom Keydunov gives us ‘How Agentic Analytics Is Replacing BI as We Know It’ this is his view and as the co-founder and CEO of cube he is talking in his own street and that is his right. The issue with the article that it is really good, but there are some issues (from my point of view). The start is (optionally) great and with “For over two decades, the business intelligence (BI) dashboard has been the primary interface between data teams and decision-makers. These visualizations, charts, and KPIs have been invaluable tools for understanding what is happening inside a business. But in 2025, the dashboard model is showing its age. In a world where data moves at the speed of cloud transactions, connected devices, and global markets, static dashboards can no longer keep up. By the time a decision-maker logs in, refreshes a dashboard, and sifts through its filters, the critical moment for action may have already passed. Business leaders want answers, not just visualizations, and they want those answers as events unfold. A new approach, driven by AI and automation, is emerging to fill this gap.” There is merely spoken truth here and he is correct, but the Dashboard was ‘thought’ of by a Business Intelligence analyst and that tends to have hidden settings as that tends to be the case and the more it is set to the BI industry it was designed for, the better that tool tends to be. So when we see “By the time a decision-maker logs in, refreshes a dashboard, and sifts through its filters, the critical moment for action may have already passed” is not incorrect, but there is a time gap, we get that and the better the tool, the smaller the gap and as the designing analyst is better the more precise the tool becomes regardless of gap. So now we get to the ‘Agentic Analytics’ of the matter. It is programmed and based on the data it is trained on. Now, if this is all in-house data, that tends to be OK, but there is still the programmer and that is the culprit of the story. You see a programer is as good as the explainer hands him his data (tends to be a sales person) and that is already the issue. Sales persons are set to the blinkers then have (like pupils shaped as dollar signs) not the most eloquent setting to begin with.
So then we get to “The static nature of dashboards has made them a bottleneck in modern analytics. They rely on the user to know what question to ask, when to ask it, and how to interpret the results. When organizations scale, the proliferation of dashboards often leads to confusion rather than clarity. A company may have hundreds of dashboards, each presenting a slightly different view of the truth, leaving teams overwhelmed and second-guessing their decisions.” This is a truth and a half no matter how you tweak it. And the stage of “proliferation of dashboards often leads to confusion rather than clarity” is set to the organiser behind this and that tends to be a salesperson, CEO or CFO, as such money is the operative word and Agentic Analytics (AA) is set to data and clarity of collected data and upgrading this won’t make the data more clear, it merely showed how the dashboard fell short of what’s needed. So when we get to the ‘good’ part with “A company may have hundreds of dashboards, each presenting a slightly different view of the truth, leaving teams overwhelmed and second-guessing their decisions” we see the gap in the entire AA setting. It isn’t less confusing, the tweaked set of data is likely misrepresenting what was needed in the first place and I will grant you that this is my view on the data. I have seen dozens of cases where that was the case and in some cases it was with people managing data the size of a Fortune 500 company. So as we get to the really good part, Artyom Keydunov tells us “The promise of agentic analytics depends on trust. Without robust data governance, AI-powered systems risk surfacing misleading or inconsistent insights — and worse, they might automate actions based on flawed assumptions.” This is a powerful statement, it is not the trust part, this is inherently drawn from the loyalty a firm instills, it is “they might automate actions based on flawed assumptions” you see, ‘flawed assumptions’ is the key here and it is with many dashboards and as such with AA solutions as well. That just gave me an idea (perhaps cube has this) there is a between setting where the app could have documentation in the ‘second tier’ a setting where a document cog could be embedded in the software solution that is merely accessible at the core company that made this setting. So where some see “growth margin per quarter” the hidden blockchain will refer to that setting and the documentation will set the parameters for inspection. It could be any kind of blockchain with the setting of corporation – application – sequential counter and that is documented. You see, it is not what is now that matter, but in 5 years the reality of any solution (or AA) will require revision and wouldn’t it be great that you are able to vet what was (correct or not). So, now go back to any dashboard that was designed over 10 years ago and still in use. How many will not be able to tell you what was?
A simple setting merely shown to you and perhaps in your own firm there are several others. So make of this what you want. The article is quite good and even as it is talking in the street of Cube, it shows some common grounds we all need to have before we all go the way of the Dodo because AI told us to do just that and we end up at the edge of a cliff like darling little lemmings and when we realise we are at a cliff, the lemming behind us its pushing us in the back making us fall over. Nice ride, don’t you agree?
So have a great day and for me a new coffeeshop open tomorrow, so another option to try pointing myself for the simple reason that only the once trusted coffeemaker knew how we wanted our coffee, just like the users of a dashboard now relying on some AA that we are supposed to do it their way (which might not be wrong).






