How to bridge gaps between generative AI ambitions and readiness

By Jim Freeman

Generative AI is a topic of interest everywhere today–from homes, and educational institutions to offices and executive boardrooms. Generative AI, of course, refers to algorithms that can be used to create new content, including audio, code, images, text, simulations and videos based on the data they have been trained on.

Its applications are being tested and adopted across various industries, from healthcare to finance to retail. However, as with any groundbreaking technology, the response to generative AI has understandably been mixed.

While there’s enthusiasm for its potential benefits like improving human efficiencies and creating personalized experiences, the C-suite is also exercising caution as they consider its implications on safety, ethics and trust. Before diving in, business leaders must understand the potential impacts of generative AI on productivity and the future workforce.

In my own conversations with CIOs, a theme is clearly emerging–there is a gap between their Gen AI ambitions and their readiness for it. Here are a few things to help a business leader build their strategic roadmap for responsible, scalable generative AI adoption.

A strong data foundation

The success of any generative AI application hinges on the quality of the data feeding it. The old adage ‘garbage in, garbage out’ is more relevant than ever with generative AI.

If your data is flawed, incomplete, or siloed, you’re likely to end up with misleading—or worse, harmful—AI-generated outputs. That’s why building a strong data foundation is critical.

For scaling AI responsibly, organizations need their data foundation to include a strong and agile data strategy. This means having the flexibility to adapt to new data sources and changing business needs without missing a beat. It’s also vital to ensure compliance with regulations like GDPR, especially as privacy and data protection are a constant concern for companies of all sizes and in all verticals worldwide. Maintaining customer trust is paramount, and the last thing any organization wants is to compromise on data security.

A rigorous data cleaning and validation process is another cornerstone of a strong data foundation. By ensuring high-quality inputs, you set the stage for your AI to deliver valuable and actionable insights.

Additionally, integrating data across the organization helps avoid silos, ensuring comprehensive data coverage. This is where master data management comes into play—it helps achieve a unified view of your data, making it easier to manage and deploy across various AI applications.

Identify the right use cases

Generative AI’s potential spans a wide range of functions—from customer service and marketing to software development. However, not every use case will justify the investment. It’s crucial to prioritize those that align with your business goals and offer clear, measurable benefits.

My colleagues and I have seen a busy international airport use generative AI to help staff centralize customer feedback from multiple platforms, do sentiment analysis on the feedback and generate personalized responses.

We’ve seen a bank use generative AI to help developers automate routine code generation, so they could focus on delivering more business value.

We’ve also seen marketers use generative AI to create personalized campaigns that truly resonate with individual customers, tailoring messages to their unique preferences and behaviors.

By adopting a collaborative approach across business functions, your organization can identify the most impactful generative AI use case for your business to help you receive a return on your investment.

Leverage LLMOps

As you begin to operationalise generative AI, it’s essential to incorporate Large Language Model Ops (LLMOps)–a framework of practices, techniques and tools used for managing large-language models effectively and securely. This helps ensure your AI deployments are reliable, scalable and cost-effective.

LLMOps is not meant to be a standalone framework—it needs to be part of your broader data and AI strategy to ensure seamless integration and optimal performance. Some best practices in LLMOps include implementing stringent data security measures to protect the integrity of your information. Regular updates and maintenance of AI models are also vital to ensure they remain relevant and effective over time.

Generative AI is an emerging technology, and integrating it into your operations requires a strategic, considered approach. This isn’t a race, but a journey.

With a solid foundation, a proactive strategy, cost optimisation and the right use cases, enabled by a trusted partner, this journey can be a successful one.

Jim Freeman is CTO for Kyndryl ANZ