C-suite point of view: Top 5 challenges to AI adoption

Generative AI (GenAI) has disrupted how virtually every organization operates. In fact, 81% of organizations around the globe plan to increase their AI spend next year, according to the Enterprise AI Maturity Index by ServiceNow and Oxford Economics. But are they fully prepared to tap into the opportunities?

I recently had the privilege of discussing this challenge in a ServiceNow Executive Circle webinar with Hartmut Mueller, chief transformation officer at ServiceNow, and Philip Carter, general manager and group vice president of tech buyer digital platform research at IDC.

Our discussion focused on the game-changing value of GenAI: better user experiences, faster ways to work, and smarter decision-making at the organizational level.

Philip shared insights from the IDC Worldwide CEO Survey, 2024, which examined the role of leadership in the “age of AI everywhere.” Drawing from the input of C-suite leaders at more than 350 companies across North America, Western Europe, and Asia Pacific, Philip laid out five key challenges to AI adoption. Let’s break them down.

1. Measuring return on AI investments

One of the foremost challenges CEOs face is measuring the return on investment for AI initiatives. Unlike with traditional projects, the benefits of AI often become clear only in the long term.

"CEOs are very much aware and focused on the opportunity associated with AI investments," Philip said, but they often struggle to quantify these benefits in the early stages of projects. We’ve seen this challenge ourselves at ServiceNow.

That’s why, rather than jumping into a new AI initiative, we first define the business problem that needs to be solved (and determine if it really needs an AI solution). The issue should link closely to a measurable impact goal. Across the organization, we’re defining specific and measurable goals for AI projects, beginning with pilots that can demonstrate value before scaling.

The majority of our GenAI use cases focus on enhancing productivity, primarily by automating repeatable tasks. To demonstrate the financial impact, we must convert the saved time into a specific dollar value using reasonable and defensible assumptions.

To support this goal, we’ve established an AI dashboard that displays the results across all use cases. We expect that more than half of these use cases will not deliver value initially but will pay off over time. In the meantime, our task is to continually refine and improve these capabilities.

2. AI governance

Another concern keeping many executives up at night is putting AI governance frameworks in place that are robust enough to help ensure ethical, transparent, and secure AI use.

Our experience at ServiceNow has underscored the need to inventory all the AI models we’re using across the organization, as well as the data and algorithms they use. AI is an enterprise asset, and we need to manage it as such.

For example, we developed a governance app to measure and manage the inventory of our AI models, including data, security, privacy, and performance considerations. This real-time application, built on the Now Platform, provides visibility into all the GenAI models deployed across the company.

3. AI skills and talent strategy

According to IDC’s research, 60% of CEOs report their organizations don’t have the skills across business functions to execute on their AI initiatives.

At ServiceNow, we believe low-code and no-code AI development will continue to grow rapidly. As a software as a service provider, we can easily deploy AI solutions because we have no need to move data for model development.

Our out-of-the-box solutions also accelerate time to value. The speed offers plenty of experimentation opportunities to refine the model quickly.

4. Cost concerns

Most organizations like the power of large language models (LLMs)—until they get the first bill. The costs associated with AI initiatives, including maintenance and upgrades, can be quite unpredictable.

Leaders are increasingly exploring options to optimize costs—such as cloud-based AI services and predictable pricing models—and ensuring their budgets are flexible enough to accommodate unexpected expenses.

With the rise of domain-specific models, we can produce faster, more accurate results since the LLMs are trained on focused domains.

5. Prioritizing the right use cases

Identifying and prioritizing the most effective AI use cases can be difficult, especially given the broad applications of AI across different business functions. Use cases often fall into three categories: productivity, experience, and growth.

Boosting productivity with AI involves breaking down every job into a series of tasks. Many of these tasks are ripe for automation. The question is: Which tasks are best suited for people, and which should be delegated to machines?

To identify tasks suitable for AI, consider these criteria: repetition, data dependency, and cognitive demands involving synthesis, prediction, and translation. AI can handle these tasks extremely well, especially where some margin of error is acceptable, as accuracy improves with scale.

In regard to experience, we’re seeing many use cases focused on AI-enhanced user interfaces. Think about how the practice of “prompt chaining” is changing how we engage with data and business logic.

Rather than navigating multiple screens to develop a financial report, for example, you can simply generate this analysis through a conversational interface and have it presented to you in real time.

Regarding growth, more organizations are incorporating AI into their business models to generate new revenue streams. In fact, IDC predicts that by 2026, organizations infusing AI into their business models could see double the revenue growth compared to their peers.

The bottom line is AI adoption presents a transformative opportunity for businesses. Leaders must navigate these five challenges to realize the full potential of AI. Staying proactive and adaptable will be essential to maintaining a competitive edge as the AI revolution continues to evolve.

Find out how ServiceNow helps organizations put AI to work.