Leveraging AI – Best Strategies, Overcoming Challenges, and Insights for Success Keynote

With 61% of organizations evolving their data and analytics operating model because of AI technologies, according to Gartner, it’s no surprise that it was a hot conversation topic at the Spring Visions CIO Leadership Summit.

We have aging infrastructures that don’t support where we’re going with AI/ML. We have new tools like vector databases and concepts such as RAGOps and MLOps. This leads us to a lot of questions around foundational issues to address, what adoption issues could occur, and at the finer level, the social impacts of AI. When we take a look at the topic holistically, it’s clear that everyone wants and needs to contribute, but it’s not always clear where to jump in and get started.

I posed these AI questions during the Spring Visions CIO Leadership Summit keynote panel, Leveraging AI – Best Strategies, Overcoming Challenges, and Insights for Success, to gain insights from industry leaders, including:

• Adnan Hasan, Head of GTN Strategy – Data & AI, Google
• Chad Aronson, Global Head of Intelligent Automation, Uber
• Gavin Hupp, VP of IT, United Parks and Resorts

AI Implementation Foundations – Getting Started

Before even thinking about specific AI projects, organizations need to lay the proper foundations. As one panelist stated: “The question that you should always think about is, what are the business outcomes that you’re really looking at?” Having a clear strategic vision aligned with business objectives is critical.

However, the technical readiness is often lacking, according to another panelist: “We see this across the board, that most customers don’t have their foundations ready for AI ML. It has been going on for a long time.”

Data readiness is a key foundational requirement that can’t be glossed over. The panelists highlighted the importance of data quality and accessibility in successfully implementing artificial intelligence (AI) and machine learning (ML) projects. Lack of access to quality data was identified as a common challenge facing organizations looking to leverage AI and ML.

Robust data governance and privacy controls are also essential. One panelist highlighted: “Governance is going to be a priority for a lot of executives now.”

By aligning AI initiatives with the overarching business strategy and priorities from day one, the organization can understand what they’re building and why.

Challenges and Hurdles in AI Project Adoption

Many of the conversations around challenges in AI adoption are the same as 10 years ago, when we were building out the foundations of data warehouses, and later, big data platforms. Organizations often have aging, siloed data infrastructures that need modernization before they can integrate AI/ML capabilities. Having a unified data fabric/platform vision is critical, rather than looking at point solutions.

Even with a solid strategy and foundations in place, adopting AI projects faces hurdles when integrating with legacy data platforms and siloed architectures. As discussed: “Organizational data is not easy accessible because data is stored in many silos. So whoever is using it, whether it’s a data scientist, AI engineer, or analyst, they continue to struggle.”

This tech debt becomes a significant bottleneck: “We have growing business expectations and sometimes we have aging infrastructures that aren’t supporting where we’re going.” Modernizing data platforms is required but difficult alongside existing systems.

Another hurdle to address is the need to build trust and visibility with business partners in the organization. Change management is critical to drive successful AI projects, with continuous communication with end users and stakeholders. Organizational change should be handled carefully, with constant feedback involved.

An incremental approach focused on quick wins was recommended: “As we build a new service, we turn an old one off. As we build new data products in our new data environment, we decommission or turn off those legacy data sets.”

Considering the Social Impact of AI

The discussion also touched on the subject of social impact and brand protection in the context of AI/ML. Data protection and privacy were identified as crucial factors in maintaining a brand’s reputation and preventing malicious use of technology.

Robust data governance, privacy, and ethical AI principles need to be embedded in the overall AI strategy and in each AI/ML project implementation.

Protecting brand reputation from malicious AI model use or bias is paramount. Stewards, controls, and audit processes are critical components for responsible AI development.

As AI capabilities grow, panelists emphasized the need to proactively manage social impacts like privacy, ethical use, bias, and brand risks.

Overall, AI opens up amazing opportunities but only if implementation is done in a controlled, strategic manner accounting for organizational readiness and social responsibilities. As expressed by the panelists, it requires a cohesive, holistic approach beyond just the final AI use cases.

Ready to get started with your AI/ML foundation? Download our guide “Is Your Data Ready for AI/ML?” to assess your preparedness.