Delivering Enterprise AI Value Requires More Than Technology - Forbes

Despite widespread recognition among leaders of the potential value of enterprise AI as a driver of business transformation, too many organizations continue to struggle to translate ambition into measurable business value.

It would be easy to conclude that the issue must be the technology, that perhaps the value of AI in the enterprise has been over-promised and it’s still far too hard to implement.

But that doesn’t seem to be the case. While the technology certainly isn’t trivial to deploy, like a lot of change within the enterprise, many of the underlying obstacles are fundamentally human and organizational.

The Human Side Of Enterprise AI

For over a decade, Jaclyn Rice Nelson, the CEO and co-founder of Tribe, has seen the challenges of implementing enterprise-scale AI initiatives. She’s had a front seat view of the rise of AI.

Formerly a VP with Alphabet’s growth equity fund, Google Capital (now renamed CapitalG), she worked to place Google engineers inside the startups they invested in to accelerate market readiness. Google’s pioneering work on AI meant that she experienced some of the earliest pre-generative AI capabilities and saw what it was delivering for the firm’s portfolio.

As early as 2015, the future became clear for Nelson: virtually every business would become an AI-enabled organization.

Over the next few years she made an observation that would change the course of her life. Many companies, despite their enthusiasm, were simply not in a position to easily engage with emerging technologies from firms like Google and others in order to integrate it and extract business value.

In particular, with the rapid emergence of AI, Nelson saw with clarity that organizations would need a heavy dose of specialized technical and business skills, in addition to necessary technology prerequisites, such as quality data, strong data governance, and modern data infrastructure, to make any headway.

With this insight, Nelson left Google and started Tribe to address this opportunity.

Enterprise AI Requires More Than Good Models

While the vast majority of businesses have adopted the use of generative AI tools such as chatbots and benefited from many of the AI features now available in enterprise applications, the same can’t be said for success in building and integrating complex AI solutions into existing workflows. In those instances, the failure rate continues to be painfully high.

The reasons why complex, integrated AI efforts so often don’t succeed range from poor data quality and readiness to weak problem definition and project governance.

It’s also a result of too many leaders believing that today’s AI is just plug-and-play; it just works out of the box. Successfully integrating AI into existing systems and processes requires significant prerequisites, including the infrastructure, integration layers, security controls, and AI operational capabilities needed to support AI at scale.

Business leaders want AI to solve problems and create new capabilities, but the gap between accessing AI capabilities and operationalizing them securely, reliably, and at scale is far greater than many appreciate.

The Enterprise AI Opportunity Is Business Reinvention

Nelson validates the report findings on why so many AI initiatives never make it into production, but she also thinks that there’s room for CEOs and other leaders to reimagine the role AI can play in their business.

What she observes are well-intentioned leaders who view AI as a new way to solve a specific problem or build a point solution. She recognizes that both of these types of requests are reasonable and if appropriate scoping is performed, AI may be the right approach.

However, what AI brings to the table, unlike so many emerging technologies that have come before, is an opportunity to reinvent the business. Rather than using AI to simply automate or reduce costs in how the business currently works, Nelson wants more leaders to understand AI's potential to fundamentally redesign how their businesses operate. In the worst case, insufficient ambition could result in losing ground to a competitor who seizes the possibilities.

Leaders can be the disruptors or they can be disrupted. It’s a choice and Nelson fears many are not moving fast enough.

How Leaders Can Improve AI Success

If CEOs want AI to transform their business, they’ll need to ensure that they’ve defined the big picture and what it is that they want from it. Specifically, leaders should establish clear business objectives and define how success will be measured. In addition, considering how hard and complex the effort will be, rigorous diligence is required to justify the investment.

Nelson emphasizes this point. Poorly defined or insufficiently ambitious requirements are a limiting way to begin the AI journey. Accurate scoping at the start must be a priority as it will lead to understanding the gaps that exist in the business, how they must be addressed, and the real costs of the effort. It will also help to answer whether the business aspirations are technically feasible.

As part of the technical feasibility assessment, integration with existing systems and workflows must be analyzed and designed. This is an area where big AI efforts often stumble if not considered early. The worst time to discover an integration issue is when the time comes to integrate.

Assuming organizational readiness is established and the requirements are clear, Nelson recommends a fast, iterative build cycle. Foundational data and AI infrastructure should be deployed first followed by regular, small feature sprints that demonstrate measurable business value, build organizational confidence, and steadily advance the broader transformation.

Yet even organizations with strong technical foundations can stumble if they underestimate the human side of transformation.

Transformation Starts With People

Perhaps the most difficult part of a large, complex AI project isn’t the technology, governance, data, or skills, but rather the role that humans play in supporting and enabling the implementation of an organizational redesign.

While certainly not unique to AI, implementing new technologies and processes often meets resistance. The far-reaching implications of AI amplify this. People fear change and uncertainty, and it can manifest in many different ways including introducing all manner of blockers from slow approvals to persistent naysaying.

These blockers are certainly not insurmountable, but they must be anticipated and mitigated to the extent possible. Consider rigorous change management, stakeholder engagement, strong communications, and redesigned incentives.

The disappointing results so far for many enterprise AI initiatives don’t suggest a problem with the technology. Rather, they highlight how difficult organizational transformation has always been. Leaders should recognize that enterprise AI success depends less on the technology itself than on preparing their organizations for the scale of business transformation and human change it demands.