## The Challenges of Implementing AI at Scale
Implementing new technologies at scale has always been a challenge for large organizations. Whether it’s mobile, Big Data, or the cloud, many companies have struggled to fully embrace and implement these technologies. Now, it’s AI that is forcing companies and their employees to change, whether they like it or not.
One of the biggest challenges is technical debt. Organizations often have a tech stack that is designed for a prior era, and changing it to take full advantage of new technologies can be risky. Managers are often hesitant to fully embrace such changes, as they involve tremendous risk along with enormous potential.
Another challenge is institutional inertia. Changing how people do things is difficult, as people tend to stick to what they know. To illustrate this, consider the story of a small town register of deeds implementing a computer system. The workers at the front desk were resistant to change because they associated their rubber stamp with their identity and sense of power. Eventually, the system architect allowed them to keep their stamp, even though it was no longer necessary. This compromise helped them buy into the change.
Change management is perhaps the biggest challenge of all. Implementing new technology isn’t just about shopping, buying, and implementing it. The hardest part is getting people to use it. Companies often have to find ways to accommodate employees’ preferences or they risk sabotaging the implementation of the solution.
## The Radical Adjustment of AI
AI represents a whole new way of working. While previous technological shifts like the PC and the internet lowered the cost of information transmission, AI is different. It’s lowering the cost of expertise. This presents a new set of challenges and requires organizations to rethink the role of computing in their operations.
Organizations need to consider issues like answer accuracy, data leakage, and the data used to train AI models. Companies like Box are developing frameworks and paradigms to help organizations navigate these challenges, but with multiple vendors offering similar solutions, finding the right one can be difficult.
## The Promise and Perils of AI
One of the challenges organizations face is determining whether AI is truly delivering on its promise of increased productivity. It’s difficult to make a direct connection between AI capabilities and increased productivity. This can make it hard to sell AI internally to skeptical workers who are concerned about their own futures.
However, there will also be employees demanding these new tools, creating tension within organizations. Some argue that companies have to adopt AI, even if they don’t see immediate benefits, in order to avoid losing market share and becoming irrelevant. AI brings transformative potential and companies need to embrace it.
## Overcoming the People Problem
While CEOs can easily see the power of AI without needing a technical explanation, vendors still need to show value to organizations. However, the bigger challenge lies in getting past the people problem. There are three truisms to keep in mind when implementing AI: machines won’t replace humans, but humans with machines will replace humans without machines; AI will fail if change is not driven from the top down and incentives are not created for employees to adopt it; and change cannot be forced, it must be defined and communicated effectively.
Implementing AI at scale won’t be easy. Each organization has different levels of maturity and technological readiness. However, by addressing the people problem and embracing the potential of AI, companies can navigate this challenge and thrive in the age of AI.
