Lessons from the cloud

Prior to the development of public clouds, organizations ran their computing services via private data centers. IT departments spent decades securing their digital fortresses, where everything was tightly managed, safeguarded, and controlled.

About 15 years ago, organizations started to outsource on-demand computing services and infrastructure to third-party vendors. These vendors offered services over the public internet, meaning they were available to any individual or organization that wanted to purchase them. The advantages were—and are—numerous. It’s easier and faster to deploy public clouds than on-premises infrastructures, and public clouds are far more scalable than their private counterparts. With internet access, every employee in the company can access the same source of truth from anywhere in the world.

But when public clouds emerged, people panicked. There was a tremendous amount of fear and doubt over whether this was the right path to take. Suddenly, employees were throwing private data into the cloud vendor’s infrastructure, sometimes without oversight or coordination across their companies. It felt a bit like the Wild West, with executives either going all-in or catastrophizing about the future.

The fear and excitement over AI mirrors the rhetoric of the public cloud era. On the one hand, executives are thrilled about the possibility of making work faster and smarter. On the other hand, they’re incredibly anxious about the future. 

When executives talk about AI, they’re naming many of the same anxieties I was hearing 15 years ago. This is good news for executives. Because many of the questions are the same, the answers are, too.

1. What if there’s a data breach?

One common mistake from the public cloud era was waiting to create a security strategy. Companies were blindsided by data breaches because they thought they could put off security planning until the last minute.

In the AI era, companies must think about security, compliance, and governance from the get-go. Gartner recommends that organizations apply AI trust, risk, and security management (dubbed “AI TRiSM”) frameworks before and while adopting the technology, not after. These frameworks should encompass AI “governance, trustworthiness, fairness, reliability, robustness, efficacy, and privacy,” according to Gartner. Organizations that work to operationalize AI TRiSM will deploy models that work better, drive business goals, and increase customer trust, while organizations that don’t will face negative outcomes such as breaches, privacy failures, and reputational loss. 

2. What if the cost spirals out of control?

When businesses moved toward public cloud adoption, expenses quickly got out of hand. Businesses rarely developed a unified enterprise architecture for managing cloud storage. As a result, individual departments didn’t reliably communicate who was doing what. Confused, businesses spent money on multiple vendors that were basically offering the same services, and various teams developed redundant workflows and technologies to interface with these vendors.

To avoid making the same mistakes with AI, it’s important to build a centralized enterprise architecture: a platform through which everyone can see what processes are in place, what is being automated, what data is being generated and collected, and what roadblocks employees are encountering. Creating a centralized architecture reduces the amount of duplicate work that teams end up doing, and it makes it easier to communicate about strategies and goals.

3. Do we have the organizational culture to make it work? 

At the beginning of the public cloud era, executives didn’t make any changes to organizational culture prior to adopting the new technology. That’s understandable. It had never been done before. But the status quo was not up to the challenge. Public cloud technology is fundamentally about sharing resources and tools. Employees were not ready to think about projects that way.

What executives should have done—and what they should do with AI—is foster a culture that emphasizes three things: communication, collaboration, and agility. To make AI work without breaking the bank, employees must talk openly about failures, work together to iterate, and pivot when things go wrong. Above all, they should feel comfortable sharing ideas and resources across teams.

While AI is an innovative technology set to radically change the way businesses operate, putting it to work quickly and with minimal missteps does not mean having to reinvent the wheel. Past experiences with tech offer a guide for executives looking to the future.