I think the bigger shift is happening in the analytics architecture itself. As AI becomes the interface for querying data, the bottleneck moves from accessing information to establishing whether the answer is trustworthy.

An AI-generated insight is only as reliable as the metric definitions, joins, filters, and aggregation logic behind it. If those aren’t governed consistently, two agents can answer the same business question differently, even when both queries execute successfully.

Dashboards can serve as a validation surface, but I’d go a step further: connect AI-generated answers to a governed semantic layer, executable query lineage, and the underlying records or aggregates. That gives analysts a way to reproduce the result instead of relying on a visual cross-check alone.

I also don’t think every question needs a dashboard anymore. For straightforward queries, a direct answer with its source and assumptions may be enough. Dashboards become more valuable when investigating anomalies, comparing segments, or exploring relationships that weren’t part of the original question.

The real evolution is from visualising data to making analytical answers reproducible, traceable, and trustworthy.

I think we’re moving toward a model where AI handles the query layer, while dashboards provide the verification layer.

Instead of navigating multiple reports to understand why revenue dropped, an analyst can ask an AI agent to investigate the change, identify contributing segments, and surface anomalies. But the answer is only useful if it can be traced back to the underlying data.

That’s where dashboards still have a technical advantage. They expose metric definitions, filters, time ranges, and relationships that can be difficult to capture in a generated response. The real challenge is ensuring the AI and dashboard operate on the same semantic layer, so a metric like revenue or churn has a consistent definition across both.

I also think dashboards will evolve beyond static visualisations into interactive evidence layers, where users can inspect the data behind an AI-generated insight, reproduce the query, and validate the assumptions.

The future isn’t dashboards disappearing. It’s moving from manually searching for insights to asking better questions and verifying the answers.

I think dashboards are gradually shifting from being the primary place where people discover insights to becoming a place where they validate what AI tells them.

AI makes it much easier to ask questions in natural language and get an immediate answer. But the answer is only as reliable as the underlying data, definitions, and context. That’s where dashboards still matter. They let people drill down into the numbers, compare trends, check anomalies, and understand what’s actually driving a result.

The interesting shift is that dashboards may no longer need to be the starting point for every analysis. AI can handle the initial exploration, while dashboards provide the transparency and evidence needed to trust the conclusion.

The technical challenge is making sure both layers use consistent metrics, reliable data pipelines, and shared business definitions. Otherwise, AI-generated answers and dashboard figures can contradict each other.

I see the future as less about dashboards versus AI and more about how the two work together: AI helps you find the answer; dashboards help you verify whether it holds up.