Your BI Tool is Dying: Your Career Doesn’t Have to

GoodData’s Ryan Dolley offers commentary on why your BI tool is dying, but your career doesn’t have to. This article originally appeared in Insight Jam, an enterprise IT community that enables human conversation on AI.

You’re scrolling LinkedIn and another post appears; someone building real-time analytics in the latest ‘cool’ BI tool, someone else talking about AI native business intelligence stacks. It all seems so futuristic. Meanwhile you maintain your old school dashboards and pipelines that break every time someone changes a column name and wonder what the hell an ‘ontology’ even is.

The industry talks endlessly about Iceberg, you feel like the Titanic. And with AI looming on the horizon, you’re falling further and further behind. 

Your fear is justified. The days of being just a dashboard monkey are numbered.

In the age of AI, anyone can generate a chart. What separates people who get promoted from people who get automated is knowing which problems are worth solving.

This skill transcends tools and trends and it’s something you can build despite whatever legacy tech stack you are saddled with today. Here are four steps to transcend tooling that will set you up for the next stage of your career.

The first is the most important and it has nothing to do with technology.

Focus on People, not Technical Problems

Most BI and analytics professionals suffer from what I call the ‘builder mindset.’ Like the medieval stonemason, the builder is concerned with how. How am I going to set the keystone in the cathedral arch; how am I going to visualize this data? The builder has the secret knowledge and skills to make the vision of others a reality.

Being a data builder is a satisfying job that has paid very well. But there’s a big problem. Increasingly, how will be handled by AI. The skill of knowing every menu and button in Tableau or Power BI was already commodified, soon it will be the domain of machines.

You need to move from how to what, why and for whom. These are the critical gaps that you can fill TODAY regardless of what BI tool you are using. So how?

Stop obsessing over the data and start obsessing over the people. Value doesn’t exist in clean data models or optimized queries, it exists in the real world, in the feelings, thoughts, goals, and fears of your users, executives, and customers.

Here’s my success metric: I want someone to tell me, ‘The data you gave me helped me get a promotion.’ That’s the bar. Not ‘nice dashboard,’ not ‘this was helpful.’ I want to know I changed their career trajectory.

Start there. Not with questions about what reports they need or what metrics they track. First, understand their goals in human terms. Then translate to metrics and reports. Then build.

When you succeed, you dramatically improve the quality of your output without modernizing your technology at all.

Move Fast Despite Bureaucracy 

Legacy BI teams often have governance-focused, default to ‘no’ cultures. As frustrating as it may be, that culture has a purpose; to protect the people who put it in place by eliminating risk and externalizing blame. I ran headlong into this culture early in my career.

I was on an enterprise BI/DW team with a seventeen step SDLC that involved multiple gates, approvals and migration paths for literally every change in production – and this was after we transitioned to ‘agile.’ No matter how hard I argued against it, changing things was viewed as just too risky.

So I de-risked it. I identified a class of changes that were perceived as low risk, mostly focused around display elements like colors and chart types, but also more meaningful changes like adding metrics that already exist in the semantic model to non-regulatory or financial reports.

I then ran the analysis and showed that these requests took up roughly 30% of our total development with a truly horrible ratio of process time to actual development – something like 5 to 1. This meant that 25% of our total department work effort was going to bureaucracy for these low risk requests.

Armed with this, I designed a ‘quick access workflow’ that had 5 steps with just 1 gate. Because I understood that fear of risk was the number one motivation for my department leaders. I argued that the risk of burning so much effort on busywork was greater than allowing this well defined set of tasks to proceed quickly, and I eventually won out.

You may be in a similar situation. Identify what drives that culture and find concrete, low risk ways to start pushing against it.

When you succeed, you dramatically improve the velocity of your output without modernizing your technology at all, and over time can expand the scope of requests that can be done quickly.

But you’ve also done something very sneaky – you’ve planted the seeds of a rapid, AI driven workflow in your organization. It’s just nobody knows it but you.

Create Your Personal AI Toolkit

Is there a huge AI-native data stack you need to start doing modern BI? No. Forget semantic layers, ontologies, AI chatbots for now. Remember, you are stuck in legacy land. Don’t wait for your boss, start where you can – your own work.

Odds are you have a massive amount of metadata stored in your head about how to get shit done at your organization. You need to get it out and into a format AI can interpret to help you build.

Pick a domain or business area you know extremely well, pop open the model you have access to and input the following prompt.

I am setting up the necessary documentation for you to help me create dashboards, reports and ad-hoc answers for end users in [[domain]]. The final output will be in one or more files for you to interpret. What questions do you need answered to begin helping me with this.

You don’t need to be an expert in interacting with AI to start. Let AI guide you and learn as you go. It will probably ask about data sources, metrics and calculations, semantics, personas, visualization standards. If it leaves one of these out, prompt it in eventually.

Then dump everything to Google Docs, YAML, whatever. Feed it back in every time you start a new project.

This is the basis for your one person AI analytics department. The important thing is that you created the personal AI development infrastructure to accelerate your work and teach you the basics of the technology.

Get AI on Your Resume

Now is the time to take the step beyond legacy to become an AI driven BI developer. You developed the right mindset. You created options for fast moving, modern development styles. You built your personal AI skills and have an assistant ready to go.

Put them all together and identify a low risk BI output that you can deliver end to endwith AI. Here’s what this looks like in practice:

A stakeholder asks for a ‘sales performance dashboard.’ Instead of jumping into Power BI:

1. Use AI to research: ‘What are common sales performance metrics in [industry] that we don’t use yet? What decisions do sales leaders typically need to make?’ This gives you intelligent questions to ask before the meeting.
2. Come prepared: Instead of ‘what do you want to see?’, walk in saying ‘I understand these are common concerns in your role,which resonate with you?’ Suddenly you’re a strategic partner, not an order-taker.
3. Use AI to prototype and iterate: Have AI generate sample metric definitions and visualization specs before you touch the BI tool. Iterate in minutes instead of days.
4. Build in legacy tool: Go ahead and click the buttons in Power BI, you’ve already done the strategic stuff much faster thanks to AI.

When you’ve done this a few times one of two things will happen.

Your velocity and work quality will improve so much that your employer loosens up and embraces the future.

Or you’ll be able to walk into your next interview and say: ‘I used AI to accelerate my analytics workflow, improved delivery speed by X%, and drove strategic outcomes in a legacy environment.’ That’s a compelling story.

Evolve or Die, BI Style

You’re 35. You’ve got 10-15 years before you’re either leading data teams or explaining why you’re still a senior BI developer. The people who make the leap aren’t the ones who got lucky with modern tools, they’re the ones who accelerated past their constraints to embrace the future their employer wasn’t ready for.

The Brutal Truth: Your company will eventually replace your legacy BI tool. The question is: will you be part of the modernization, or will you be the person they replace alongside the tool you’ve been maintaining?

Ryan Dolley is a data VP and host of the popular Super Data Brothers Podcast and author of the Super Data Blog Substack.