AI Adoption: Diverse Strategies for Enterprise Success

The journey from an artificial intelligence prototype to a full-scale production system is often fraught with difficulties for many organizations. While a universal blueprint for this transition remains elusive, the successful experiences of companies like Ernst & Young (EY) and Lumen offer valuable insights, showcasing markedly different yet effective strategies. These companies highlight that successful AI deployment is less about a single approach and more about tailored innovation.

EY, a global leader in tax and advisory services, operates within a heavily regulated financial landscape. This environment naturally steers them towards a cautious and highly responsible methodology for AI integration, prioritizing risk management alongside technological advancement. Their strategy is built on establishing clear frameworks and ensuring compliance. In contrast, Lumen, a telecommunications firm, has taken a more aggressive stance, fostering an organization-wide AI culture by immediately equipping all employees with AI tools. Both demonstrate unique pathways to leveraging AI effectively within their respective operational contexts.

Navigating AI Deployment with Prudence: The EY Approach

Ernst & Young, a prominent global tax and advisory organization, is actively developing and deploying portable frameworks designed to guide its clients through the complexities of AI adoption. Internally, the firm manages an impressive 30 million documented processes and supports 41,000 agents in production, leveraging this extensive internal knowledge base to inform client strategies. This deep well of experience allows EY to provide practical, evidence-based advice for successful AI integration.

The accelerating pace of technological evolution means that even more foundational AI technologies are emerging, poised to transform enterprise IT operations further. Joe Depa, EY’s global chief innovation officer, notes this rapid advancement. He points out the shift from generative AI to agentic and then physical AI, with quantum computing also on the horizon. This constant innovation creates a dynamic landscape that businesses must navigate strategically.

Organizations today find themselves in a challenging position, simultaneously implementing new AI processes while still grappling with legacy infrastructure. Depa highlights that many companies have not yet fully modernized from previous technology cycles, adding another layer of complexity to current AI initiatives. This necessitates a careful balancing act between innovation and maintaining existing critical systems.

A cornerstone of successful AI implementation, according to Depa, is establishing a robust data foundation. Without this critical element, AI prototypes are highly likely to fail before they can even be fully developed or deployed. A survey conducted by EY in late 2024 underscored this point, revealing that a significant 83% of organizations lacked the necessary data infrastructure to effectively capitalize on AI capabilities. This statistic emphasizes the foundational role of data in any successful AI endeavor.

Depa stresses that whether discussing generative, physical, or quantum AI, the underlying data set is often both a critical resource and a potential barrier. He strongly advocates for governance and responsible AI frameworks as essential components that enable scaled deployment. These frameworks provide the necessary structure and guidelines to ensure that AI applications are developed and utilized ethically and effectively, minimizing risks and maximizing value.

Implementing responsible AI frameworks into workflows, processes, and employee training has tangible benefits for clients, Depa explains. Such frameworks significantly reduce compliance risks, which is particularly vital in regulated industries. Beyond risk mitigation, these clients also observe greater growth and derive more substantial value from their AI investments, indicating a direct correlation between responsible practices and positive business outcomes.

The presence of responsible AI guardrails empowers teams to experiment with greater freedom and confidence. Depa clarifies that these guardrails create a safe sandbox environment, allowing teams to explore new AI applications without undue concern about potential negative repercussions. This fosters innovation and encourages a proactive approach to problem-solving using AI.

When clients encounter difficulties with AI rollouts, Depa frequently inquires about their training strategies. He notes that he has “never heard any client say they’ve over-invested in training,” suggesting a common underestimation of its importance. He further advises that successful deployment often requires moving beyond traditional training methods, advocating for more integrated and contextual learning experiences.

Effective employee training, Depa argues, must occur at the point of application for AI solutions. This hands-on, immediate learning approach ensures that employees truly grasp how to utilize AI tools in their daily tasks. This method is far more effective than theoretical instruction, fostering deeper understanding and practical proficiency.

Depa illustrates this point with the example of robotic surgery. This technology offers the potential to perform procedures with precision equal to or surpassing human capabilities, addressing physician shortages and enhancing health outcomes. However, the advanced technology’s potential remains untapped if hospitals and doctors are unwilling to adopt it. This highlights that the primary hurdle is often not technological, but rather a challenge in change management and human processes.

Cultivating an AI-First Mentality: The Lumen Approach

Lumen, a company actively expanding its network infrastructure to meet the growing demands of AI, has made AI adoption a core strategic commitment at the highest levels of its organization. The company implements what Sean Alexander, Senior Vice President of Connected Ecosystems, describes as a “tops-down, bottoms-up” approach. This dual strategy ensures that AI initiatives are championed by leadership while simultaneously being driven by grassroots innovation.

Alexander firmly believes that “culture eats strategy for breakfast,” a principle he finds especially pertinent in the realm of AI. This philosophy underscores the importance of embedding AI into the very fabric of the company’s operational ethos. A strong, pervasive AI culture is seen as more critical than any specific strategic blueprint.

Lumen’s CEO, Kate Johnson, sets a powerful example by integrating AI tools into her daily routines, demonstrating a visible commitment to the technology. This top-down interest in AI permeates throughout the company, evident in the onboarding process where new employees are immediately provided with AI tools from their first day. This ensures early exposure and integration.

Alexander confirms that they activate Copilot Studio and Copilot Enterprise for every employee. For new hires, this immediate access to AI tools significantly shortens the time it takes for them to become fully productive, reducing the traditional six-month onboarding period to approximately four months. This rapid integration maximizes their potential contribution to the company.

Alexander is also spearheading a “Copilot Studio in a day” initiative, a program designed to quickly build employee confidence and proficiency in AI. Teams dedicate half a day to training, followed by hands-on “hacking” sessions. This rapid, practical experience encourages experimentation and familiarization with AI tools, fostering a “maker” culture.

Lumen has established a robust governance model specifically focused on the responsible adoption of AI. This framework encourages employees to take agency in solving problems using AI, while also ensuring that all initiatives begin with clear, measurable metrics. This strategic approach ensures that AI efforts are purposeful and aligned with business objectives, fostering innovation within defined boundaries.

Alexander shares an example where a sales leader leverages AI to enhance team performance. This leader records weekly one-on-one conversations with direct reports and then feeds these transcripts into a large language model (LLM) he developed using Copilot Studio. This allows him to pinpoint specific friction points, identify new opportunities, and detect any strategic planning drift, providing actionable insights for improvement.

The company has successfully deployed several AI solutions in production. One notable example is a “migration buddy agent” designed to assist customers in transitioning from legacy products to Lumen’s strategic portfolios. This agent automates various tasks, including customer lookups, product validation, offer validation, compliance checks, and contract reviews, significantly streamlining the migration process.

While a human agent still reviews the outputs, the AI-driven system dramatically reduces the time involved in these processes, leading to increased responsiveness and improved customer satisfaction, Alexander explains. This hybrid approach combines the efficiency of AI with the critical oversight of human expertise, ensuring both speed and accuracy.

Before deployment, Lumen forms dedicated teams comprising both business and technical leaders to develop and test these AI agents. This collaborative approach ensures that solutions are robust and meet business needs. Testing is followed by a carefully managed rollout process, typically involving groups of about a dozen customers.

Alexander highlights the importance of thorough testing, including A/B testing and controlled rollouts, to ensure that every AI solution consistently meets high-quality standards. This iterative process allows for fine-tuning and validation before broader implementation, guaranteeing reliability and effectiveness in real-world scenarios.

For customer service, Lumen prioritizes time-to-resolution for network outages as its most critical metric. One of Alexander’s colleagues transformed an initial “hack” into an LLM-based feature named “Ask Greg.” This system effectively resolves network issues by reasoning through problems and suggesting resolution steps, integrating data from dozens of systems, including health monitoring, telemetry, and geospatial information.

With approximately four million customer service requests annually, this pilot program, which originated as a simple hack, is now projected to save Lumen around $10 million in annual costs, Alexander estimates. This impressive saving underscores the significant financial impact that well-implemented AI solutions can have on operational efficiency and profitability.

Lumen also capitalizes on a knowledge graph built upon its Microsoft 365 data. By organizing its SharePoint data according to department and security levels, Copilot can augment conversations with near real-time understanding of Lumen’s products, services, and operational intricacies. This rich, interconnected data ecosystem enables more informed and efficient decision-making across the organization.

Alexander affirms that these continuous AI integrations are actively transforming the company daily. Lumen’s commitment to embedding AI into its core operations and culture positions it for ongoing innovation and efficiency gains, demonstrating a proactive approach to enterprise-wide digital evolution.