Why Most “Data Driven” Programs Fail Before They Start
Everyone wants AI… but no one wants to fix the data first.
Over lunch this week with a Head of Technology, we ended up having a really interesting discussion about AI.
I’ve known him for more than a decade and he recently stepped into a new Head of Technology role. Naturally the conversation turned to the organisation’s technology roadmap.
He mentioned the CEO and Board are pushing hard to move toward AI driven capabilities.
Then he said something that really stuck with me.
“We keep talking about AI, but the reality is our data foundation is nowhere near ready.”
I see this mistake constantly across Brisbane.
A department or business decides they want to become “data driven.”
They launch a major analytics or reporting program.
Then they hire a junior analyst or junior data engineer to lead the work.
And within 6 to 12 months the project stalls.
Not because the person is not capable.
But because the role was never truly junior.
As he put it during our conversation:
“Everyone wants AI outcomes, but very few organisations want to invest in the data architecture that actually makes it possible.”
In my experience, data programs tend to struggle when organisations hire too junior for three reasons.
1. Data requires architecture, not just reporting
A Power BI report is easy.
Building the data pipeline, governance model, security, performance and data modelling behind it is where the complexity lives.
Junior hires can absolutely support that work.
But they cannot design it.
Recommended by LinkedIn
2. Poor foundations create messy and unusable data
When the underlying architecture is not designed properly, organisations quickly end up with:
• Multiple sources of truth • Broken or inconsistent dashboards • Different definitions of the same metrics
Eventually, the business loses trust in the data.
And once that trust is gone, it is incredibly difficult to rebuild.
3. The business expects insights, but funds support
Most organisations say they want strategic decision making capability.
But the roles they hire for are focused on operational reporting and dashboard maintenance.
As he summed it up perfectly:
“The expectation is strategic insight, but the hiring decision is often tactical.”
The fix is actually quite simple.
If the program is strategic, hire strategically.
Bring in a Data Architect, Senior Data Engineer or Data Governance Lead early.
Then build the analysts and engineers underneath them.
Data transformation is not a junior problem.
It is a senior foundation role first, then execution.
This is something I’m seeing more and more across the Brisbane technology market.
Curious to hear from others.
What is the biggest mistake you have seen organisations make when building their data capability?
#AI #DataAnalytics #DataGovernance #Hiring #DigitalTransformation #BrisbaneTech
Everyone wants AI but far fewer want to fix the data first. The biggest mistake I see is treating data capability as disconnected initiatives rather than a cohesive system. Organisations invest in data strategy, governance, master data, and data quality but often in isolation, with unclear ownership and weak links to outcomes. Some even distrust their internal teams, bringing in consultants who shift the goalposts on tools and processes often to create more work and justify their own engagement. The result: • Strategy becomes a document • Governance adds control, not impact • Master data stays fragmented • Data quality lacks accountability • AI amplifies inconsistency instead of insight Real capability comes from clear ownership, alignment to outcomes, and accountability at the source. Until then, AI sits on unstable foundations
Data foundations absolutely make or break AI initiatives. Worth the investment upfront.
Very common in the industry, the data basics are not there. The market needs AI ready data, but we have been as an industry building dashboards and hiding the data quality issues, lack of business consistent data understanding through our development practice. I would also add we tend to play in safe data domains finance, hr, procurement etc but business efficiencies are in operational data across business process optimisation. Current data practices are hard to scale out. Most of my work now are resetting the data.practices up in companies! Fun times.