What it takes to succeed with AI

In a landscape where artificial intelligence (AI) is rapidly evolving, enterprises are struggling to harness its full potential, risking a decline in employee productivity. Vini Cardoso, Cloudera’s chief technology officer for Australia and New Zealand, emphasizes the need for organizations to transition from fragmented AI initiatives to a cohesive, organization-wide platform approach that integrates AI models directly with their data. This shift is not merely a technological upgrade; it represents a fundamental change in how businesses can leverage AI to drive measurable outcomes.

Cardoso's insights, shared in an interview with Computer Weekly, underscore the importance of aligning AI projects with tangible business value. He points out that while initial estimates for AI project approvals may be encouraging, it is crucial for organizations to track and measure the actual value generated during implementation. For instance, Cardoso cites a case involving a regional bank that successfully created A$150 million in annual value through AI applications spanning both generative AI and traditional machine learning. This achievement was made possible by carefully selecting use cases that aligned with the bank's strategic goals and demonstrated significant value.

The Role of Leadership in AI Adoption

A critical element in the successful adoption of AI is the involvement of business leaders. Cardoso argues that leaders must embrace the changes brought about by AI, which can only happen if they understand the positive outcomes that AI can deliver. He highlights the necessity for leaders to set an example by fostering a culture that encourages experimentation and learning among teams, all while maintaining focus on transformation objectives. This leadership approach not only empowers employees but also enhances job satisfaction by allowing them to see the direct impact of their work on business decisions.

For instance, a data analytics team could analyze quarterly business reviews and suggest targeted investments based on their findings. This strategic guidance can lead to more informed decision-making, moving away from a scattergun approach to a more focused strategy that maximizes potential returns. Cardoso notes that such involvement transforms the role of employees from mere software operators to influential decision-makers, enhancing their job satisfaction and engagement.

However, the transition to AI is not without its challenges. A recent survey conducted by Sapio Research for Foxit Software revealed that the initial use of AI might temporarily hinder productivity. Among 1,400 desk workers and executives surveyed in the US and UK, executives reported saving only 16 minutes per week, while desk workers experienced a loss of 14 minutes due to the time required to validate AI outputs. This discrepancy highlights a significant trust gap, with only one in four executives expressing strong confidence in AI outputs compared to one in ten desk workers.

Building Trust and Moving Beyond Pilots

Cardoso acknowledges that skepticism often accompanies the adoption of new technologies, including AI. However, he believes that as organizations demonstrate the value of AI, trust will gradually build. He draws a parallel to the evolution of self-driving cars, where initial skepticism has been replaced by increasing confidence as safety and sophistication have improved over time. To foster trust in AI, organizations must ensure that they use reliable data sources and maintain transparency regarding data transformations.

Keeping humans involved during the early stages of AI implementation is vital for building this trust. Cardoso emphasizes the importance of combining technical skills with business acumen, which enables professionals to review frameworks and guide decision-making effectively. As AI technologies evolve, so too must the frameworks that govern them, ensuring they adapt to new knowledge and changing expectations.

Another common challenge is the transition from pilot projects to full-scale production systems. Cardoso notes that projects with the potential for real value are more likely to receive the necessary organizational support. He advocates for adopting a platform approach from the outset, which provides a standardized method for accessing data sources and developing AI applications. This approach not only facilitates scalability but also ensures that data and workloads are reusable, allowing for consistent security, governance, and compliance.

Cloudera offers trial access to its software, enabling potential customers to evaluate both its capabilities and their specific use cases. The platform is designed to be flexible, capable of deployment on-premise, in the cloud, or in hybrid environments, allowing organizations to optimize costs while maintaining governance standards. Cardoso highlights the importance of bringing AI workloads to the data rather than moving data to AI, which can be costly and inefficient.

In conclusion, Cardoso asserts that a platform approach ensures AI operates where the data resides, safeguarding intellectual property and sensitive information while promoting sovereignty requirements. He stresses that AI should enhance human skills, inspire innovation, and create value rather than replace teams. Focusing on high-value use cases is essential; while low-hanging fruit can be beneficial for learning, addressing real business challenges is crucial for securing investment and gaining board-level support for scaling AI initiatives.