Five key steps for C-suite execs navigating AI deployment challenges

Demystifying AI at scale

Even though generative AI (GenAI) has quickly become a global business imperative, many C-suite executives are still unsure how to deploy the transformative technology within their organizations. In the 2023 CEO Outlook Survey, 62% of executives said they knew their organizations needed to act swiftly to avoid giving competitors a strategic advantage. Yet, in the same survey, 61% said uncertainty around AI makes it challenging to develop and deploy an artificial intelligence (AI) strategy widely across an organization.

Two additional surveys point to further uncertainty surrounding GenAI. In the 2024 EY CIO Sentiment Survey, while only half of chief information officers (CIOs) said they believe GenAI will enhance the value of their organizations, 83% anticipate that their budgets for GenAI will increase over the next 12 months. In addition, an EY survey of employee attitudes on AI use found that while 76% of employees said they encountered a positive experience with AI at work, 75% were concerned that AI adoption would make certain jobs obsolete.

Moving forward with GenAI at scale

How can organizations overcome this hesitation and demystify GenAI for members of the C-suite? And, what’s the best way to move forward with an enterprise-wide strategy that enables them to reap a wide range of benefits: from improving research and development efforts to enhancing their supply chains?

One way that organizations are getting started is by defining a pipeline of new GenAI-powered use cases, such as a call center chatbot assistant, that serve as value-focused opportunities for deploying AI at scale. Attempting to scale a new technology prototype such as GenAI is not for the faint of heart, however. To be effective, GenAI requires access to AI-ready data, which is often difficult at the enterprise level. In addition, to deploy GenAI responsibly, organizations need to take measures to maintain confidentiality and prevent data leakage.

Organizations can start by identifying the assets necessary to deliver these high-priority use cases and create specific outputs, such as SEO-aware product listings. GenAI can synthesize data to trigger predictions that will activate text, image, video or a dashboard that can deliver product ideas, advertising copy or supply chain models via a standard query. This can’t happen without reliable data, however, which means that organizations need to extract data from smaller, more discrete databases and then train them to support specific use cases.

Another way organizations are gaining value, and driving cost effectiveness into their AI strategy, is by developing an enterprise-wide control tower or center of excellence. This function will enable visibility into AI use cases across the business, helping to reduce duplication of efforts, leverage reusable assets and focus on realizing maximum value from AI.

Organizations should plan to look ahead and carefully determine how they will deploy GenAI at scale. This step demands commitment and dedication from all levels of the organization and should be done in a phased approach. Here are five key steps organizations should consider:

1. Collaborate with partners and vendors: This will expose organizations to different views on GenAI and help companies address potential skill gaps, security and legal challenges and minimize development costs. This will also require them to take steps to protect and secure data if it is shared outside the organization. Developing interorganizational ecosystems of knowledge, training and security related to AI within the company and external relationships with companies that lead in this space could help to demystify anxiety around GenAI and empower employees to use it as a competitive advantage.
2. Prepare to deploy AI across systems and people. Deploying GenAI typically demands a strategic, coordinated approach that leverages significant infrastructure and employee investments. Offering training to upskill existing employees, restructuring business processes and establishing novel roles for people within the new workstreams will help smooth adoption.
3. Acknowledge security and ethical implications: Moving forward with GenAI also raises the stakes for potential ethical and regulatory issues that could result from its misuse. Organizations need to align GenAI deployment to legal frameworks and ethical considerations, particularly with respect to data privacy, security and brand standards. As an adjunct to this, organizations also need to take steps to prevent algorithmic bias and safeguard fairness and IP in their AI systems. Employees share these concerns. In our recent AI anxiety survey, 81% of employees said they want leading practices on responsible AI to be widely shared in their companies.
4. Embed risk management. While GenAI can yield great organizational rewards, it also can generate unintended consequences. As showcased in our Balance risk and reward with responsible AI article, organizations need to take steps to deploy AI responsibly and embed risk management into the development lifecycle. This means recognizing, mitigating and prioritizing AI risks, particularly to address cybersecurity and data privacy. Automated model risk management can help minimize risks to the organization while also enabling scaled deployments.
5. Identify change agents. Organizations also need to find the right people who can help with early adoption of GenAI and further secure buy-in within the workforce, Employees should be encouraged to develop and propose their own use cases for the pipeline on how they would use AI to help them achieve success in their daily tasks and for team objectives.

In the meantime, organizations also need to develop a strategic plan for deploying GenAI at scale that reflects their broader technology processes as well as goals. This should also include complementary workstreams to address human capital needs. AI at scale has the potential to drive substantial change across business operations, and organizations need to be ready for the potential impact on the enterprise, particularly employees.

Finally, when an organization is deploying AI across their companies, it’s important to remember that transformation does not happen overnight. AI models on their own do not transform the business. It is the bringing together of operations, technology, change management and data science teams to build and ensure adoption of AI over time that will unlock the most significant value. It will take time to achieve the full transformative impact from AI, and organizations need to give their businesses the time needed to redesign processes and ways of working that align with the new technology to foster effective company-wide success.