Set priorities
In a successful technology transformation, actions are closely tied to solving business problems. That’s why CIOs need to work with business unit leaders to set priorities and make those choices work. The principle is to identify use cases where Gen AI will drive your strategy forward. That requires abandoning failed pilots and focusing on promising ones. This analysis should also consider cost-benefit. Because Gen AI is still new, costs can balloon and it can be difficult to scale. One rule to remember is that for every $1 spent building a Gen AI application, you need roughly $3 in change management, including training people and actively tracking performance.
It's also important to resist the temptation to just disconnect your technologists, which leads to multiple, sometimes overlapping, platforms that cost both money and time. A better approach is to architect your infrastructure and applications in a way that gives you the flexibility to switch providers and models with relative ease.
Don't treat next-generation AI as a technical program
Gen AI is a team sport, and the CIO is the head coach. To have true impact, Gen AI must leave the IT function and become embedded in the business. That means integrating the technology with product, risk, legal, and other departments. One of the key focuses of this cross-functional team is to develop and adopt protocols and standards that support scale. There are different ways to develop such a team, and the CIO has a big say in its composition and mandate. Some companies have launched centers of excellence, while others have chosen to have separate strategy and delivery units. The key is for the team to work well together, understand what they are trying to accomplish, and check in regularly along the way. The CIO needs to ensure the team acts as a value builder, not just a manager of the work.
The principle to keep in mind is not to create different pieces, but to make them all work together. Each use case needs to have access to multiple models, vector databases, prompt libraries, and applications. This means that companies need to manage different sources, such as applications and databases that are in the cloud, on-site, at a vendor, or a combination, while being resilient and consistent with existing protocols, including access rights.
