This commentary is provided by McKinsey & Company, a Fortune Global Forum Knowledge Partner. Rodney Semmel He is a senior partner in McKinsey & Company's New York office and a global leader for McKinsey Digital.He is a co-author of Rewired: A McKinsey guide to winning the competition in the age of digital and AI.
The natural language capabilities of generative AI are so easy to use that even CEOs who aren't usually early adopters give them a try. In late 2022, he said, less than a year after AI-based tools exploded onto the scene, a quarter of executives were already using them.
While widespread interest in generational AI is creating a massive wave of use cases and experimentation, there are also challenges. Such initiatives are relatively easy to start, but they can eat up resources without creating much value.
To emerge from this pilot purgatory, connecting AI to business outcomes must be a priority. Here are four ways CEOs can do just that.
Focus on something important. As Gen AI proliferates to a wide variety of pilots, it can look like a technology looking for problems. However, when Gen AI is directed at an area large enough to effect change, such as the customer journey or functional area, meaningful change will occur. For example, McKinsey has teamed up with financial services giant ING to develop a solution powered by his GEN-AI, which can deliver precise solutions to customers through language and data capabilities. This improved service while freeing up agents to deal with more complex issues.
Create a business-driven technology roadmap. There are so many unknowns with Gen AI that developing protocols and standards requires a central team comprised of all relevant capabilities, including risk, legal, compliance, finance, human resources, and strategy. That effort must start with the CEO and executives agreeing on what needs to be done. The CEO must work closely with the chief information officer or chief technology officer (CIO or CTO) to translate that commitment into a concrete roadmap that directs the company's path forward. Of course, domain transformation is not just about Gen AI applications. It also includes process digitization and other forms of AI. If your application is built around reusable modules, it can also be applied to a variety of future problems.
Build a talent bench. Building a talent bench is non-negotiable. Partnering with external providers, such as senior engineers who are already building generative AI products, can be an important part of your generative AI strategy. But you need to focus just as much, if not more, on your internal talent as well as your technical team. People on the business side also need to understand what genetic AI can and cannot do.
Companies can improve the skills of data engineers to learn multimodal processing and vector database management, for example, while data scientists can develop skills in rapid engineering and bias detection. And it's important to retain these professionals. A recent McKinsey survey of nearly 13,000 employees found that 51% of Gen AI creators and heavy users plan to leave their roles within the next three to six months. Compensation is always important, but talented people are more likely to stay if they have opportunities to develop their skills, engage in meaningful initiatives, and advance.
For example, McKinsey worked with DBS Bank in Singapore to successfully implement its digital transformation. The results showed that 80% of the talent was insourced and 20% was outsourced. This combination has enabled organizations to move faster and make decisions faster. The principle is clear. You can't outsource great things.
focus on what's important. Companies consume a lot of oxygen deciding which large-scale language models (LLMs) to use. But all his new generation LLMs can do amazing things. It's more important to have the right efforts in the right places, including context engineering, security, governance, and technology upgrades to ensure support for generational AI at scale. It may sound obvious, but many pilots find themselves in protected environments that don't reflect reality on the ground.
Improving the data required for a particular solution can have a significant impact on the quality of the output. The same goes for investing in an orchestration engine. Gen AI requires a lot of interaction and integration between models and applications. Application programming interface (API) gateways are a key element of this orchestration capability, as they mediate access and enforce compliance. A good API not only reduces risk but also gives your team confidence.
The performance gap between leaders and laggards in digital and AI technology is widening, with leaders faring better financially. If this trend spreads to artificial intelligence, latecomers could fall further behind.
Deriving true value from genetic AI is certainly possible, but it is harder than it looks. This is partly because it looks so easy. However, that is not the case.
The opinions expressed in Fortune.com commentary articles are solely those of the author and do not necessarily reflect the author's opinions or beliefs. luck.
