The Future of Analytics

From Answering Questions to Building Decision Systems

Read Free for non-members

I have always found data analysis a bit strange.

Not because it is useless. Quite the opposite. But because, in many companies, the role is not always about deep analysis. A big part of the job is simply knowing where the data is, how to access it, how to clean it, how to transform it, and how to create an output that other people can use.

• A dashboard
• A SQL query
• A weekly report
• A metric breakdown
• A slide
• A table sent in Slack

Of course, there is analysis in all of this. But there is also a lot of execution. A lot of translation between business people and databases. A lot of work that exists because the company needs information, but does not have a simple and reliable way to consume it.

This is why I think the future of analytics is not only about AI replacing analysts.

It is more interesting than that.

AI will not kill analytics. It will kill low leverage analytics workflows : dashboard factories, repetitive reporting, manual SQL extraction, and ticket based self service.

The future belongs to analysts who can combine business judgment, software practices, and data reliability.

1. The old analytics interface is disappearing

For a long time, the data analyst was the interface between the business and the data.

Someone had a question. The analyst found the right tables, wrote SQL, checked the logic, created a chart, and explained the result.

This workflow made sense because most people could not access data directly. Even when BI tools existed, they still required context. You had to know which dashboard to open, which filter to use, which metric to trust, and what the number actually meant.

Generative AI changes this interface.

Now, people can ask questions directly inside the tools they already use. They can do it in Slack, through an AI assistant, inside a product analytics tool, within a CRM workflow, or even through a weekly memo.

The dashboard is no longer the only place where analytics lives.

The dashboard was never the goal. The goal was shared understanding.

So if AI can bring the right information directly into the workflow, a lot of traditional analytics work will be compressed.

2. But easier access does not mean better decisions

This is where I think the conversation often becomes too simplistic.

Yes, AI makes data easier to access.

But access is not the same thing as trust.

If everyone can ask questions to the data, everyone can also misunderstand the answer faster. People can use the wrong metric, miss the business context, compare two numbers that should not be compared, or trust an output that looks confident but is wrong.

When everyone can generate an analysis, the differentiator is no longer access to SQL. It is judgment, context, and trust.

This is where the data team still matters.

Maybe even more than before.

The role becomes less about producing every analysis manually, and more about making sure the system behind those analyses is reliable.

Having clear metric definitions, well-structured data models, reliable documentation, strong testing practices, appropriate access controls, continuous data quality monitoring, clear lineage, good observability, review processes, and a shared understanding of the business context behind the numbers.

In other words, the future of analytics depends less on the ability to create one more chart, and more on the ability to build a trustworthy analytical environment.

3. Dashboards are not dead, but their role is changing

I do not think dashboards are dead.

Some dashboards are still useful because they create a shared ritual. You look at the same metrics every week. You know where the important chart is. You see the trend. You compare actuals with targets. You build a common visual language around the business.

But a lot of dashboards were never really dashboards.

They were debugging tools. Monitoring systems. One time analyses. Political documents. Backlog items that survived too long. Places where people stored metrics just in case something went wrong one day.

AI will probably split those use cases.

Some information will continue to be consumed through visual dashboards, while other insights will be delivered as automated alerts, summarized in AI generated memos, embedded directly into business workflows, or even consumed by software agents without any human interaction at all.

So the question is not “will dashboards disappear ?”

The better question is : “what is the right interface for this decision ?”

• Sometimes the answer is a chart.
• Sometimes it is a short memo.
• Sometimes it is an alert.
• Sometimes it is an automated action.
• Sometimes it is no interface at all, because the system can already react.

4. The analyst moves from execution to system design

This is the part that interests me the most.

If AI compresses the execution layer, the analyst has two possible futures.

The first one is passive : the analyst becomes someone who checks AI outputs, validates queries, and corrects mistakes. That future is not very exciting.

The second one is much more interesting : the analyst becomes someone who designs the data systems that allow better decisions to happen.

They will not only answer business questions. They will design the systems that make analytics trustworthy and usable. That means defining data models, metrics, and guardrails, while helping build AI assistants, automated reports, alerts, and workflows that fit naturally into how teams work. Their role will remain critical, but it will move closer to engineering.

This is why I think the software mindset becomes more important.

Not because every analyst needs to become a software engineer. But because analytics is becoming more like software.

Version control, testing, documentation, modularity, code review, deployment, observability, and ownership become essential because they are the foundations that make analytical systems reliable, maintainable, and scalable over time.

AI makes analytics faster, but software practices make it reliable.

And reliability is probably the real bottleneck.

5. The future analyst is closer to a builder

For me, this is also where the role becomes more interesting.

A data analyst with a software mindset can do more than extract data, build dashboards, or prepare slides for stakeholders. As LLM can reduce the effort required for many of these tasks, the opportunity shifts toward building systems that make data continuously useful rather than producing outputs for a single use.

This creates a role that sits between several disciplines. It is not quite data engineering, not quite software engineering, and no longer traditional business analysis either.

It is somewhere in the middle.

You understand the business problem, but you also understand that the solution is often not another manual analysis. Sometimes the solution is a reusable workflow. Sometimes it is a better model. Sometimes it is a clear metric definition. Sometimes it is a small tool that removes the need for ten future requests.

That is why I do not think analytics is becoming less technical.

I think the technical value is moving.

Less time writing the same SQL manually.

More time designing the structure that makes data usable, reliable, and action oriented.

Conclusion

Analytics is not going away.

But the interface is changing.

The analyst used to be the bridge between the business and the database. In the future, that bridge will be partly automated. People will ask more questions directly. AI will generate more answers. Dashboards will become only one interface among many.

But someone still needs to make the data trustworthy, define the metrics, understand the business context, and build the systems behind them. Trust does not appear by itself. It is built through engineering, software practices, and a deep understanding of how data moves through the organization.

So maybe the future of analytics is not about doing more analysis.

Maybe it is about building the systems that make analysis useful.

Analytics is not going away. It is becoming infrastructure.

Thanks for reading, I hope you enjoyed this article, you can find more of my writing on my page !