Is Data Quality the CIO's AI Dilemma?

I’ve seen and read so many articles on data quality over the last year or so, and of course this is down to the introduction of AI specifically LLMs etc., which makes data quality, suddenly the hot topic that everyone is moaning, I mean discussing! I’ve seen a lot of talk on LinkedIn about how many companies have stood up Lakehouses over the years and are now in the pursuit of clean data for AI tools.

I’ve also seen the conversations of late on whether the “data” problem that has been created needs to be put back under the CIO. So, it was very interesting for me as I stared to write this article over the weekend to see the Gartner “CIO Report” drop into my email this morning.

This was very timely and there were two Gartner stats released today:

• 92% of CIOs believe AI will be implemented in their organisations by 2025 (more than any other technology), and
• 61% of organisations have assigned AI teams and accountability directly to the CIO.

Well looking at these stats, you would think CIOs are dancing from the rooftops, maybe they are, but I’m equally sure that some of their data colleagues aren’t! The stats are clear, the CIO will take the reins,? but I think this also brings a new kind of pressure for the CIO. The weight that no one in the C-suite is entirely ready to shoulder. The pressure I’m talking about is the CIO being pulled into a war zone, and the battles will revolve around data quality, AI strategy, and who really owns this domain.

As ever, I do need to point out that it’s not just about technology; it's about navigating the murky waters of organisational politics, particularly when it comes to roles like the Chief Data Officer (CDO), Chief Data & Analytics Officer (CDAO), and the emerging Chief AI Officer (CAIO). Ouch!

The Elephant in the Room

If you're a CIO, it's likely you have already clashed, or will clash (if you subscribe to Gartner), with the CDO or CDAO on data-related responsibilities, and quite frankly, this is where a lot of the problems begin. These roles are designed to be partners in strategy but often end up competitors in execution. So, while the CIO is expected to spearhead AI initiatives, it's not without resistance or confusion over who actually owns the data quality, data governance, and AI rollout strategy.

The fact that 61% of organisations are throwing AI accountability onto the CIO shows that the pendulum has swung. Many organisations have realised that data, AI, and digital transformation are intertwined (I hope), and when things go wrong, it’s typically the CIO who must answer for it. I say this as there is now a trend for some of those data roles to be reporting into the CIO.

But here's the catch. Just how many CIOs are truly set up to deliver high-quality data, sophisticated AI tools, and a seamless data & AI operating model for the entire organisation?

Don’t Demote Data Quality to a Check Box Activity

One of the biggest misconceptions in the AI world is that once you have clean data, you're ready to go. It’s more complex than that. AI doesn’t run on just any data, it requires high-quality data, and that's where most organisations fall short. Data lakes, warehouses, and now Lakehouses often contain a mix of structured and unstructured data, much of which is incomplete or unclean. Some even call them swamps, hopefully there aren’t alligators lurking or is it the crocodile, I always get confused about those ferocious creatures! Either way don’t get caught in those jaws!

If your data isn’t up to scratch, you’re setting yourself up for failure, regardless of the tools you deploy. AI amplifies errors, biases, and inconsistencies. Bad data leads to bad decisions, and if AI is making those decisions, the scale of the problem increases exponentially.

This is where the CIO needs to get serious about air cover. If you're a CIO reading this, you need to make it abundantly clear to your board that you can't be successful with AI if you're not being given the organisational support to clean up your data. Sure, the CDO might claim ownership over data governance, but without strong backing from you as the CIO, the data strategy will remain fragmented.

Why the CDO Role Hasn't Delivered

I’m not going to sugar coat this and interestingly enough, there are a few CDOs in the past week who have written the same! A big part of the issue is the CDO role itself. It was introduced to fill a void in data governance and data management, but in many cases, it has muddied the waters rather than clearing them. How many organisations have CDOs or CDAOs who are truly delivering business value from data?

The answer isn’t encouraging. Many of these roles have focused too much on technology and large protracted data governance programmes spanning years with no fruit to bear, forgetting that these initiatives must directly tie to business value.

A CDO who can’t deliver a clean, usable data sets is essentially handing the CIO a loaded gun with no bullets. The blame for failed AI initiatives then often falls at the CIO's feet, even though they never had full control of the underlying data quality in the first place.

The time has come for CIOs to stop assuming that the CDO is their safety net. In many cases, the CDO is struggling with their own remit, and you as the CIO need to be prepared to take ownership of data quality and AI strategy if you want to succeed. Did I just say that? Yep, I did, because I think it’s time that we start throwing weight behind the CIO who needs to come in from the cold.

The Technical Challenges Around Data Quality

Getting back to the Lakehouses for a moment, and I must tell you I am not a technical specialist, but have learned through many data strategy initiatives the true reality. They promise to bring together the best of both worlds with the flexibility of a data lake with the structure of a warehouse. But I think they have also introduced a new complexity, and this in my opinion is managing the flood of structured, semi-structured, and unstructured data, to ensure it’s usable for AI tools.

This overall problem has not just caused an IT problem; it’s had a knock on effect onto the organisation. This is where we see the usual problems, such as data silos, lack of standardisation, no data modelling, and poor data governance which all contribute to a tangled mess that AI tools simply can’t work with effectively.

As a CIO, your focus now has to be on breaking down these barriers and ensuring that your Lakehouse is more than just a glorified data dumping ground.

What’s worse is that many organisations don't have the right processes in place to monitor and improve data quality over time. And without continuous oversight, even the cleanest datasets will degrade. If you’re investing in AI, you have to invest equally in long-term data quality monitoring, which is something that often gets side-lined in the rush to implement the next shiny tool. I was only speaking to a CDO that I’m working with for a large payment processor in Latin America, and this is the true reality of the problems they are now having!

Is there a way forward for the CIO?

Organisations right now are haemorrhaging cash into new shiny toys, and that means it is now time to talk about taking control of the spiralling situation. If you're a CIO and 61% of organisations are now looking to you to lead their AI initiatives, you need to own it.

So, as ever, here are my thoughts on what this means for the CIO and what can be done:

1. Define Data Quality Accountability: Push for clear ownership over data quality. Make sure DQ initiatives are related to specific use cases that are related to business value. If the CDO isn’t delivering, don’t hesitate to take control.
2. Fight for Organisational Alignment: Ensure that your AI and data strategies are fully integrated into the wider business strategy. Stand up the right overarching data and AI operating model which means that AI isn’t a standalone project, and becomes part of the DNA and is tied into value creation.
3. Get Executive Air Cover: Make it clear to your CEO and board that AI success depends on clean, high-quality data. You can’t deliver AI without the budget, and mandate to ensure data is clean from the ground up.
4. Push for Data Monitoring: Don’t just focus on getting the data right today; ensure that it stays clean with the right governance, monitoring, and processes in place for the long haul. As AI becomes more prolific and the competition hot up, you need to be in the right position to move against your competitors.

The data and AI landscape is evolving fast, and if you're a CIO, you need to recognise the massive shift in expectations. With 92% of CIOs believing AI will be implemented in their organisations by 2025, the pressure is real. Well, at least that’s what Gartner tells us today!

But here’s the catch!

Don't get caught in the trap of thinking that someone else is handling the data problem. The success or failure of your AI initiatives will rest on the quality of your data, and if you’re not owning that, who is?

?