Are we having a BI revival?

My colleague at Advancing Analytics, Ust Oldfield has a recurring joke he makes that “everything is BI”.

Everything is BI

Data Engineering? That’s just what BI developers used to do with SQL Server Integration Services.1

Data Science? That’s just advanced applications of BI techniques2.

Data Storytelling? BI developers have been producing dashboards for donkeys.3

We even got t-shirts made of this recurring in-joke.

Take any modern data and analytics discipline and you’ll probably find it has its roots in the work that has been historically carried out by Business Intelligence developers, the OG jack-of-all-trades of the data industry.

Cut me and I bleed BI

For a very long time, Business Intelligence was how I self-identified. What did I do for a living? I was a Business Intelligence professional.

But Business Intelligence as a discipline seemed to fall out of favour for a while. I was no longer being offered jobs in “the business intelligence team”. It was doing “analytics” or “data and insights”. The work hadn’t really changed much, but the nomenclature had moved on.

This re-badging of the role wasn’t a new thing. Any one remember it being called Decision Support Systems?

I oft wondered if Microsoft might even consider rebranding my favourite tool, Power BI, and label it with a more modern product name (and it’s not like Microsoft Don’t ever rename their products, right?! ).4

I have to admit, even by now, I feel like I’ve trained my brain out of using the term Business Intelligence. Ask me what I do now, and I’ll tell you I’m a Data & Analytics Consultant.

Emerging Disciplines

Why did this happen? In some ways, I’d cite the emergence of more specialist disciplines. Did we even have Data Engineers 20 years ago? It wasn’t a job title in the organisations I was working in. But we’ve seen so much innovation and progression in the area of industrialised ETL that it’s now a job in it’s own right (and in all honesty, probably for the better. The rigour and sophistication in the way we build pipelines these days is definitely an upgrade on me dragging and dropping boxes around a clunky UI).

We’ve also seen the rise of the Data Scientist and Analytics Engineer (though my personal definition of an Analytics Engineer is that it’s just a Hipster BI Developer).5

In parallel, organisations have attempted to move towards self-service models for a lot of reporting requirements. Tools like Power BI have attempted to democratise data and put the capability and responsibility for reporting requirements in the hands of business users.

How does BI stay relevant?

Did Business Intelligence become obsolete?

Well no. I’m still here aren’t I?

A decade ago, I wondered to myself what I would need to do to stay relevant in my data career. Amongst the back drop of the late-noughties big data explosion, I figured my career path would likely lead me to train as a Data Scientist.

Yet here I am a 10 years later, and the role I perform is still fundamentally the same. I help identify business problems, find the data that could help us answer the questions those problems pose, restructure that data in to something that helps those questions be answered in a more straight forward manner, and then surface that data up in a way that allows businesses to take action on it (sometimes that might be an enabling role that provides curated datasets for analysts to interrogate, though I do still sometimes scratch my dashboarding itch).

The truth is, BI never actually went away.

The revival?

So why am I harping on about a revival? Because using the term BI is BACK!

At their recent Data and AI Summit in San Francisco, Databricks announced AI/BI, their new AI powered business intelligence offering.

We’re also seeing more and more people using the term GenBI, a category of BI tool (in which AI/BI fits) that combines all the recent progress made with Large Language Models with reporting tools to help generate conversational BI.

Even that concept isn’t new. The Q&A feature in Power BI did this to some extent, and I’d argue that Thoughspot niched down on the feature and made it even better.

However, riding in on the coattails of all the AI hype, we’re seeing renewed interest in the idea, with more sophisticated technologies now underpinning it.

So we’re going to replace BI with AI?

The narrative this gives could well lead to arms-in-the-air reactions from BI professionals around the world. AI is coming for our jobs!

But having a solid understanding of BI concepts is going to be a key enabler for this new breed of BI.

Earlier this year, Microsoft released a set of guidelines for getting the most out of Copilot for Power BI. The irony with this list of recommendations, is that these are the very same recommendations that Power BI community experts have been espousing for years!

For the new Databricks AI/BI feature, similar will apply, making sure you add relevant metadata in Unity Catalog and define your relationships appropriately being an important enabler for the AI to work effectively.

It feels somewhat grotesque that people are rushing to support these practices now, as it makes life easier for machines, and weren’t previously so keen to show the same courtesy to human data consumers. But regardless of motivation, I definitely welcome it.

So BI IS back!

Let’s quickly revisit how I described my role a few paragraphs earlier:

“I help identify business problems, find the data that could help us answer the questions those problems pose, restructure that data in to something that helps those questions be answered in a more straight forward manner, and then surface that data up in a way that allows businesses to take action on it”

Does the above still apply in the era of GenBI? Absolutely yes. And the name of the game is still the same: Business Intelligence