Generative AI is revolutionizing enterprises by enhancing customer-facing chatbots, preventing fraud, and accelerating time-consuming tasks such as developing code, preparing pitch book drafts, and summarizing regulatory reports. However, to capture the full value of Generative AI, enterprises need to develop strong capabilities across seven interconnected dimensions: strategic roadmap, talent, operating model, technology, data, risk and controls, and adoption and change management. To fully harness its potential, enterprises must develop a robust operating model for Generative AI.
The Role of an Operating Model
An operating model represents how a company functions, encompassing its structure (roles and responsibilities, governance, decision-making), processes (performance management, systems, technology), and people (skills, culture, informal networks). Enterprises that successfully leverage generative AI typically design a tailored operating model that accounts for the technology’s nuances and risks, rather than trying to fit generative AI into an existing model. Here, we delve into what an operating model is, why it is crucial, the archetypes that have emerged for generative AI, and the key decisions enterprises need to make when establishing a generative AI operating model.
Why an Operating Model is Important
A well-designed operating model is essential for effectively scaling generative AI and aligning it with the enterprise’s organizational structure and culture. It acts as a blueprint for how the enterprise puts strategy into action, ensuring that generative AI initiatives are implemented efficiently and effectively. Here are the key reasons why an operating model is crucial:
- Strategic Steering: This model helps in identifying clusters or domains of generative AI use cases that align with the enterprise’s strategic objectives. It sorts these use cases by priority, creating a roadmap that maximizes value while managing risk. Additionally, it monitors value creation to ensure efficient resource allocation.
- Standard Setting: It defines common standards for technology architecture choices, data practices, and risk frameworks. This standardization increases efficiency and allows the enterprise to leverage insights from completed projects for new ones.
- Execution: Such a model guides the design and testing of technical solutions for generative AI use cases. It ensures that use cases meeting performance and safety criteria are put into production and scaled if there is a business case, while tracking and delivering their impact.
Operating-Model Archetypes for Generative AI
Enterprises generally choose from three archetypes when designing their generative AI operating models:
- Centrally Led Model: This model centralizes oversight and management of generative AI initiatives. It is beneficial in the early stages of adopting generative AI, as it allows for consistent decision-making, efficient resource allocation, and cohesive risk management. Centralization is particularly effective in rapidly evolving environments, ensuring that the enterprise can stay on top of new developments in generative AI technology.
- Federated Model: As generative AI matures within the enterprise, a federated model can be adopted. This approach decentralizes certain aspects of the operating model, allowing individual business units to tailor generative AI applications to their specific needs while maintaining central oversight for standard setting and strategic steering. This balance can foster innovation and adaptability.
- Hybrid Model: This model combines elements of both centralized and federated approaches. It centralizes critical functions such as risk management, technology architecture, and partnerships, while allowing for decentralized strategic decision-making and execution. This model is flexible and can be tailored to fit the enterprise’s unique structure and culture.
Key Decisions in Establishing a Generative AI Operating Model
To establish an effective generative AI operating model, enterprises need to make several key decisions:
- Leadership and Governance: Decide which leaders will define the generative AI strategy and whether it will be developed enterprise-wide or at the business unit level. Establish governance structures that facilitate decision-making and accountability.
- Use Case Identification and Prioritization: Determine who will identify and prioritize generative AI use cases. Create a roadmap that aligns with strategic objectives and maximizes value.
- Talent Allocation: Define the skills needed for generative AI initiatives and develop a strategy for acquiring this talent through hiring, upskilling, or outsourcing. Establish roles for “translators” who bridge business needs and technical requirements.
- Funding Models: Set out how generative AI use cases will be funded. Decide on the level of centralization for funding, balancing the needs of individual business units with the benefits of a central funding pool.
- Technology and Data Standards: Define technology architecture choices, data practices, and risk frameworks. Ensure these standards are consistently applied across the enterprise to increase efficiency and leverage learnings from previous projects.
- Risk Management: Assign responsibility for defining risk guardrails and mitigation strategies. Adjust existing frameworks to account for generative AI-specific risks and determine if additional governance is required for particular use cases.
- Change Management: Develop and execute a change management plan to ensure the successful adoption of generative AI across the enterprise. This plan should address necessary evolutions in mindsets and behaviors.
Final Words
An effective operating model is essential for enterprises to realize the full potential of generative AI. By carefully considering and aligning the various components of the operating model—strategic steering, standard setting, and execution—enterprises can scale generative AI initiatives efficiently and effectively. As generative AI technology continues to evolve, enterprises must remain flexible, adapting their operating models to fit their unique needs and circumstances. This approach will ensure sustained value and a competitive edge in the rapidly changing landscape of generative AI.
