How CDAOs can maximise AI adoption in the enterprise

AI falling within the purview of the CDAO makes sense on multiple levels.

---

AI adoption is at a critical point. It’s moved past the buzz and gained significant traction, with many companies already introducing AI into workflows and processes. McKinsey reports that

of participants in its studies say that their company uses AI in at least one business function. Another study found that

worldwide implemented generative AI by the end of 2024, a rise of over 30 percentage points in a single year.

But there’s still a gap between implementation, and full adoption across the organisation, and many companies are struggling to close it. A recent BCG report found that 74% of companies are failing to demonstrate value from their AI projects.

What can be done to help bridge that divide? According to a recent Gartner report, it’s all down to the chief data and analytics officers (CDAOs). The report notes that 70% of CDAOs have primary responsibility for building the AI strategy and operating model for their organisation. The same report predicts that by 2027, 75% of CDAOs who aren’t considered essential for AI success will lose their C-level positions.

AI falling within the purview of the CDAO makes sense on multiple levels. In order to be useful, AI models need to be fed with data, no matter what their use case. Additionally, AI-enhanced data analytics is one of the most impactful use cases for AI, because new AI-driven business intelligence interfaces enable even the least tech-savvy team member to hold natural language conversations with data, advancing decision intelligence.

In this respect, the CDAO’s job has always been about encouraging the use of data across the organisation, and incorporating AI into those efforts only makes them more effective.

Even before the rise of AI, though, CDAOs were over-extended. Ensuring widespread AI adoption means adding yet another ball to the many they are juggling. To maximise the chances of success, here are a few key aspects of AI adoption that data leaders need to make sure they address.

The first step in any AI project is to make sure you’re maintaining high-quality datasets which are clean, well-provenanced and reliable.

It’s vital to set up processes that preserve data integrity and ensure ownership throughout the data journey. CDAOs, with their data expertise and control over the data pipeline, are well placed to serve as transformation agents, enforcing strong governance throughout the organisation.

They can draw on tools such as Alation’s Data Intelligence Platform to automate data curation. The solution uses AI and machine learning to identify potential data stewards based on their data usage, with no-code data governance workflows and a user-friendly data governance dashboard to keep track of it all.

Full AI adoption requires all employees to be comfortable with the tech involved, but there’s still a lot of nervousness about AI, especially among older, non-digitally native generations. Basic data literacy skills are also a prerequisite for AI-powered data analytics, but many employees lack these skills.

At the same time, regulatory compliant AI use demands knowledge around ethics and transparency. Even line-of-business (LOB) users have to understand AI’s weaknesses, such as its tendencies towards hallucination and bias, so that people are careful to verify the insights in AI engine output.

CDAOs possess both the knowledge and the authority to educate employees about all these points. They can give LOB users confidence about AI tools, inform them about ethical use and prepare them to handle instances of bias.

There are serious and valid fears around AI, data privacy, and business security. “Prompt leaks” are a real danger, as data shared in AI queries can be used to train models and therefore can be revealed by clever hackers.

AI adoption has to go hand in hand with robust security protocols, such as strong access policies, regular user access reviews (UAR) and limiting AI access to sensitive and/or proprietary data. CDAOs are ideally positioned to enforce vital data security rules. They can establish compliance policies, allocate responsibility for AI governance and apply compliance frameworks to ensure best practices are followed.

For decision intelligence, Pyramid Analytics’s “generative BI” solution ensures data security by keeping business databases separate from the LLM. Instead of connecting your data source with an AI model, it creates contextual metadata and shares only that anonymised information with the LLM. Using this solution, team members can have natural language conversations with their data to generate one-off insights or even rich reports and dashboards, and the AI never gains access to anything sensitive.

Effective AI adoption requires a clear strategy, and it’s not surprising that CDAOs should be the ones to build it. They know how to identify the first and most transformative use cases that should be the focus of your AI strategy, since successful projects tend to be those that deliver quick results, as well as which order to follow in rolling it out across the organisation.

There are important decisions to be made about AI reach and use. You need to decide who will have the final say in strategic decision-making, and who has authority to sign off on/override AI-powered decisions. CDAOs have crucial insights that inform these decisions.

Your CDAO also possesses expert knowledge that equips them to clearly explain the value that AI can bring, in terms that motivate each team, department and executive. By bridging the gap between CxOs and data, analytics and AI, they can embed AI into smart products and services that deliver value for every use case.

It’s not unusual for different departments to calculate metrics and enrich data differently, giving different weight to different values, but this can cause problems for AI analysis. It means that even if you begin with a single source of truth, you can end up with inconsistencies in your input data.

There are tools that can help. For example, Informatica’s Axon is a collaborative data governance tool that helps to create a unified data governance framework by encouraging cross-team collaboration, data stewardship and data lineage mapping.

More importantly, the CDAO has a holistic view of data across the organisation, enabling them to remove silos and ensure consistency in enrichment.

Although there are many hurdles along the path to full AI adoption, the CDAO has the knowledge and abilities to help you overcome them. A combination of the right tools, data analytics expertise and an understanding of your organisation equips CDAOs to guide your company to AI nirvana.