Scaling AI Projects: Why Many Corporations Stumble After the Pilot Phase

As artificial intelligence (AI) becomes increasingly integral to business operations, corporations are facing a sobering reality: many AI test projects fail to deliver expected results after being rolled out to a wider audience. The consequences of these failures can be far-reaching, leading to finger pointing, recriminations, and embarrassment. But the problem often lies not in the technology itself, but in the planning, processes, and expectations that companies have established around their AI projects.

Background & Context

AI has the potential to revolutionize the way businesses operate, from automating mundane tasks to providing insights that inform strategic decisions. However, the journey to achieving these benefits is often fraught with challenges. Many companies are finding that their AI projects, which initially showed promise in pilot phases, struggle to scale and deliver meaningful business outcomes. This is a critical issue, as the failure of AI projects can not only waste resources but also undermine confidence in the technology.

The consequences of AI project failures can be significant, extending beyond the immediate impact on the project itself. They can erode trust in the technology, making it harder for companies to invest in future initiatives. Moreover, the failure of AI projects can also lead to a lack of transparency, as companies may be hesitant to share information about their experiences with AI. This can make it difficult for others to learn from their successes and failures, hindering the broader adoption of AI in business.

Key Details

According to business leaders who spoke at a recent roundtable discussion, the problem lies in the planning and expectations that companies have established around their AI projects. Amgen Chief Technology Officer Sean Bruich emphasized the importance of having a wide number of ideas, but a very tight governance on which pilots are actually greenlit. "It's so easy with a pilot to let a thousand flowers bloom," he said, "but the key to making pilots scale successfully is actually having a tight governance on which pilots are actually greenlit."

Salesforce Chief Customer and Commercial Officer Lashonda Anderson-Williams highlighted the need to understand the intended outcome of the project. "Too many companies are focused on the successful implementation of AI features—the technological bells and whistles—instead of the business outcome," she said. This mentality is a recipe for disappointment, as the AI features may work great, but the new technology isn't driving meaningful business results.

Anderson-Williams also noted the importance of understanding the workflow when it comes to agentic AI. "A detailed understanding of the workflow—which individuals, groups, or touch points are necessary to complete a task—is critical," she said. However, many companies are finding that documentation of the workflow either doesn't exist or is poorly documented. "When you put AI on top of that, the expectation is you're going to see some magic, and there's no magic there," she said.

Access to data is another common stumbling block that AI projects encounter in the transition from the pilot phase to full deployment. With data often scattered in different silos throughout an organization, and with all that data governed by different access privileges and by varying privacy and security considerations, things can get complicated fast. It's essential to map out the contours of the AI project and all the potential data that will be required ahead of time, as Thomson Reuters Chief Data Officer Caitlin Halferty stressed. "The earlier we can uncover that in discovery, the better we'll be set up for success," she said.

What Experts Say

The experts agree that the key to scaling AI projects successfully lies in careful planning and governance. By having a clear understanding of the intended outcome and the workflow, companies can avoid the pitfalls of AI project failures. Moreover, the experts emphasize the importance of access to data and the need to map out the contours of the AI project ahead of time. By doing so, companies can set themselves up for success and avoid the disappointment and embarrassment that can come with AI project failures.

Key Takeaways

• The key to making pilots scale successfully is having a wide number of ideas, but a very tight governance on which pilots are actually greenlit.
• Companies need to focus on the business outcome, not just the successful implementation of AI features.
• Understanding the workflow is critical when it comes to agentic AI.
• Access to data is a common stumbling block that AI projects encounter in the transition from the pilot phase to full deployment.

What This Means For You

The failure of AI projects can have significant consequences for businesses, from wasting resources to undermining trust in the technology. However, by understanding the key challenges and taking steps to address them, companies can avoid the pitfalls of AI project failures. This means being clear about the intended outcome, understanding the workflow, and mapping out the contours of the AI project ahead of time. By doing so, companies can set themselves up for success and achieve the benefits of AI that they are seeking.

As you consider investing in AI projects, remember that the key to success lies not in the technology itself, but in the planning, processes, and expectations that you establish around your AI projects. By being clear about your goals, understanding the workflow, and mapping out the contours of your AI project, you can avoid the pitfalls of AI project failures and achieve the benefits of AI that you are seeking.