To lead with AI, focus on people and governance

Here's how senior leaders can optimize AI for not just their organization but also for their workers and the world.

The possibilities of artificial intelligence are huge and rapidly evolving. The technology’s ability to process vast amounts of information will lead to breakthroughs that seemed unimaginable just a few years ago. No area will be unaffected: AI is positioned to help organizations accelerate, augment and automate business activities around hiring, customer-facing processes, back-office operations and beyond.

By connecting a wider range of parameters and data sets, AI could help us push past yesterday’s tradeoffs. If implemented and managed correctly, profit and sustainability will no longer be in conflict as AI can churn through the variables and dependencies to arrive at a balanced solution (although reducing the carbon footprint of power-hungry AI needs to become a high priority). Workers, freed from rote tasks, could focus on complex problem-solving that celebrates their humanity. Supply chains could be both resilient and cost-effective.

While AI’s potential is exciting, leaders must approach this powerful technology with a clear-eyed understanding of how to align it with business missions and objectives. Purposeful, centrally organized implementation and governance of AI—that ensures we anchor humans in the process—will lead not only to improved performance but also to greater trust on the part of workers, consumers and society as a whole.  

Start the AI strategy with the human factor

When creating an AI strategy, the first agenda item is to account for its impact on humans. Our latest research New Work, New World forecasts that as generative AI takes hold over the next 10 years, it could impact 90% of jobs. In a disrupted environment like that, employers need to take an aggressive approach to upskilling and reskilling to retain the trust of their workers.

At Cognizant, we believe AI should be leveraged as an assistant within a “human in the loop” model. Our philosophy is to use AI to automate tasks and enable humans to work smarter, freeing up capacity to focus on innovating. We believe in the right mix of human and machine intelligence to deliver the right outcomes. 

Upskilling will be a moving target. Continuously training people to successfully work with AI will be an ever-changing process, requiring cross-enterprise buy-in and participation, beyond just technology teams.

Incorporate purposeful governance

Recent business history teaches us that in large organizations, multiple duplicative technology initiatives tend to spring up in various business units, regions and even departments (as many users of Slack can attest). This phenomenon, dubbed “shadow IT,” can lead to poor interoperability, silos and insufficient data security.

As we make recommendations on product, policy and process across the whole business, it is critical that models are informed by consistent objectives and a collective understanding of their operating context. We want to avoid a proliferation of AI point solutions focused on localized objectives and partial or inconsistent views of domain data. Left unchecked, this approach could lead to an environment where AI and people undermine each other and produce weak or even dangerous outputs.

Enterprise-level governance and assurance will be essential to align parallel AI initiatives with a common enterprise purpose. Accountability for AI outputs and transparency of AI-supported work needs to be established across the definition, development and operation of all models.

Differentiate with data

Data is the lifeblood of business—and the engine powering AI. To derive the best possible outcomes, leaders must ensure that both the quantity and quality of data in their organization is watertight. The data sets used to train AI models must be rigorously controlled, and responses and performance should be continuously monitored.

The challenge is: the data that enterprise AI use cases depend on to produce high-quality outputs is largely domain specific. Moreover, generative AI requires sophisticated knowledge management that enables machines to understand products, customers and operations. When businesses make investments in centralizing and augmenting their data assets, they will see greater returns that can be channeled into exploring more cases.

Businesses need to also consider the security, privacy and trust issues around data and knowledge management. Data should be sourced responsibly and used in ways that are transparent and acceptable to consumers. Our view is that platforms will play a key role in safeguarding the data needed for AI, while also making it available to innovate and adapt at pace.

Anticipate a changing AI regulatory landscape

Organizations with purpose, central visibility into AI activity and comprehensive quality assurance systems adapted to AI needs—these are the businesses that will have their house in order long before regulators come knocking. AI regulation is coming, and leaders can help shape it. Strong AI governance will be more than compliance—it will be a competitive differentiator.

Today, AI regulations look like a patchwork quilt. But this will change; while various entities will approach the technology in different ways, I believe a broad consensus will emerge. We can look to data privacy regulations as a potential model. The European Union, China and many states in the US have all implemented regulations. While they differ in important ways, it’s possible to draw a through-line that connects them.

Whatever happens on the regulatory front, internal governance is the key—if a company’s governance program is solid, regulatory compliance should be straightforward. For organizations lacking strong AI governance, regulation will prove helpful in driving better maturity and control. But it’s wiser to get out in front of the issue.

Leading with AI

To ensure AI implementations maintain uninterrupted momentum, leaders must champion a “human in the loop” model as they take ownership of AI implementation and governance. In doing so, they will better serve not just their organization but also workers and the world.

To learn more, follow us at the World Economic Forum Annual Meeting at Davos.

Visit the Data & AI section of our website or contact us.