AI news
April 8, 2024

Developing a Strategy for Integrating Generative AI

Think Big, Start Small, and Iterate

Kris Naleszkiewicz
Kris Naleszkiewicz

In this article, we discuss an approach for developing a strategy to integrate Generative AI and Large Language Models into core business functions, where the visions of the technologists/data scientists are balanced with the realities of executive decision-making. We all have seen news outlets discussing the power and potential risks of Generative AI coupled with the incredible speed that these technologies have evolved over the last 3, 6, 12, and 18 months. I expected nearly every organization to be well on its way to adoption and integration; however, the adoption of Generative AI in the core business processes seems to be lagging.

This article discusses an approach to building a strategy loosely based on the agile principles of development, where technologists/data scientists need to adopt a business mindset, and executives need to understand the broader implications of Generative AI and LLMs beyond technical solutions.

Think Big — Image generated in MidJourney by author.

By taking a ‘Thinking Big, Starting Small, and Iterating’ approach, organizations can develop an outcome-focused strategy with the flexibility to adapt to their needs, striking a balance between innovation and pragmatism.

Understanding the why

This section provides an overview of what executives should know about GenAI from a business standpoint. We will focus on notable and reputable studies done by McKinsey, BCG, Harvard, Standford, and others related to the impact and considerations associated with integrating LLMs and GenAI technologies.

Impact and Focused Areas: Incorporating Generative AI into business processes has tremendous potential across various sectors. McKinsey’s research highlights that as much as 75% of the annual value driven by Generative AI is expected in customer operations, marketing, and sales. Gartner’s predictions reflect a broader adoption, with over 80% of enterprises expected to have deployed GenAI applications by 2026, democratizing access to specialized tasks and enhancing user experiences across industries. The State of AI Report 2023 further reinforces the critical role of computing power in AI development, indicating a competitive edge for companies leveraging advanced technologies. These findings suggest that LLMs and Generative AI are not just technological advancements but essential drivers of strategic growth and innovation in the modern business landscape.

Jagged Technological Frontier: The concept of the Jagged Technological Frontier elegantly captures the uneven impact and adoption of Large Language Models (LLMs) and Generative AI across various industries and business functions. While some sectors, such as marketing and customer service, experience massive advancements and transformations due to these technologies, others, like manufacturing, may see more incremental changes. This disparity creates a landscape where strategic adoption and integration of these technologies become crucially sector-specific. Companies must navigate this jagged frontier by assessing where their industry stands regarding technological advancement and adoption and tailor their strategies accordingly. It’s not a one-size-fits-all scenario; businesses must identify their unique position on this frontier to maximize the benefits and minimize risks. The concept also highlights the necessity for ongoing risk assessment and adaptation, ensuring that investments in these technologies align with each sector's specific needs and potential. Understanding the Jagged Technological Frontier is essential to effectively leveraging the power of LLMs and GenAI strategically and competitively advantageously.

AI + Human Partnership: The AI + Human partnership in knowledge work represents a synergistic collaboration where the AI handles much of the data crunching and information processing and complements human creativity and strategic thinking. McKinsey’s insights suggest that AI can significantly enhance efficiency in knowledge work, particularly in tasks involving extensive data analysis and information processing. This shift from AI as a mere tool to a collaborator amplifies human capabilities and enables more informed and effective decision-making. This is not about replacing human intelligence but augmenting it, allowing organizations to leverage the best of both worlds. This collaboration is key to unlocking new levels of productivity and innovation, ensuring that businesses keep pace with technological advancements and use them to drive meaningful growth and competitive advantage.

Human + AI Partnership — Image generated in MidJourney by author.

Creativity and Innovation Enhancement: McKinsey’s study on the economic potential of Generative AI emphasizes its ability to expand the boundaries of what AI can achieve, particularly in driving creative processes and innovating business practices. This allows organizations to explore new, previously unattainable avenues in product design, marketing strategies, and customer experiences, fostering continuous innovation. Using AI in these creative domains accelerates the ideation process and introduces a level of depth and analysis that can refine and enhance the final outputs. GenAI acts as a catalyst for creative thinking, providing organizations with a powerful tool to reimagine and redefine their products and services, setting new industry standards.

Ethical Challenges and AI Misuse: The ethical considerations surrounding Generative AI are complex and related to bias, misuse, and fairness. The Stanford AI Index Report 2023 highlights how model scale affects bias and toxicity, revealing that larger models, while powerful, struggle with ingrained biases and toxic outputs. These challenges can be mitigated through thoughtful training data selection and rigorous mitigation methods. However, addressing these issues is not straightforward, as the relationship between AI fairness and bias indicates that efforts to create fairer models do not always correlate with reduced biases. Additionally, there has been a rise in AI misuse incidents, adding another layer of complexity and signaling a growing awareness of ethical AI usage and the potential for AI systems, like chatbots, to be exploited for unethical purposes. This has increased focus on ethical considerations in AI development and deployment, ensuring that systems are designed with safeguards against misuse and unintended consequences. These challenges underscore the need for a multifaceted approach to AI ethics that balances technological advancement with societal values and norms.

Policy and Regulation Considerations: Navigating policy and regulation in Generative AI is crucial to strategic integration. The evolving regulatory landscape underscores the need for compliance with diverse and sometimes complex legal frameworks. This is especially true given the global nature of AI development and deployment, which often crosses national and jurisdictional boundaries. As governments and regulatory bodies worldwide grapple with the implications of AI technologies, organizations must stay on top of these changes to ensure their AI strategies align with current and forthcoming regulations. This consideration of policy and regulation is not a static process but a dynamic aspect of AI strategy that requires continuous attention and adaptation.

While the integration of Generative AI represents a new era of enhanced creativity, innovation, and knowledge management, we should consider their adoption with a broad understanding of strengths and risks. These technologies come with considerations that, if overlooked, could negatively impact organizations. Acknowledging and addressing these factors at the outset ensures that the deployment GenAI aligns with business objectives and ethical, legal, and societal standards.

Developing your GenAI Strategy

Let’s examine how organizations can develop a strategy for integrating Generative AI into their core business functions using a ‘Think Big, Start Small, and Iterate’ approach. Our strategy needs to align GenAI integration with overarching business objectives with sufficient flexibility that can be tailored to each organization’s needs and context. It safeguards technological advancements, drives operational efficiencies and innovations, and contributes to broader strategic goals like market expansion, customer satisfaction, and competitive advantage.

Developing your strategy — Image generated in MidJourney by author.

Think Big

Research has shown that ‘thinking big’ is not only a prerequisite for strategic foresight but an imperative. Harvard Business School highlights the critical roles of talent and data in this new era, where AI, data analytics, and IoT are not just tools but foundational elements of business strategy. McKinsey’s study on disruptive technologies further emphasizes the need for leaders to be proactive in understanding and preparing for the impact of technological advances. BCG’s insights on deep tech innovation reinforce the importance of problem orientation and the convergence of various technologies in driving impactful innovation.

This approach to thinking big requires a shift in mindset — from viewing technology as a support function to recognizing it as a core strategic driver, capable of redefining industries and creating new paradigms of service and efficiency.

To truly ‘think big,’ organizations should envision the future with AI and other emerging technologies at the forefront of their strategy. This involves identifying opportunities where AI can solve current problems and uncover new avenues for growth and innovation. Executives should encourage a culture of creative exploration, where ideas leveraging AI are nurtured and valued. This includes imagining new business models, redefining customer experiences, and using AI to drive unprecedented operational efficiencies. ‘Thinking big’ means looking beyond the immediate horizon and imagining a future where AI is intricately woven into the fabric of the business, driving transformation and delivering tangible, long-term value.

Start Small

While embracing a ‘think big’ mindset is essential, implementing these ideas starts with pragmatic, manageable steps. This leads us to the ‘Start Small’ approach, where organizations begin their AI journey with focused, high-impact pilot projects, laying a strong foundation for future scalability and success.

This approach serves several purposes:

Building Momentum and Buy-In: Pilot projects, especially successful ones, create a ripple effect within the organization. They demonstrate the tangible benefits of the technology, fostering confidence and buy-in from leadership and stakeholders. This is particularly important when the vision is bold, but the path to achieving it isn’t clear to everyone involved. Early successes serve as proof showing that investment in new technologies can lead to tangible, measurable improvements.

Mitigating Risks: Starting small allows organizations to test the waters while limiting risk. If a pilot project doesn’t yield the expected results, the scale of failure is contained, and the lessons learned can be invaluable for future endeavors. This approach is vital for avoiding costly and high-profile failures that could set back the entire initiative.

Learning and Refinement: Pilots are excellent opportunities for learning and refinement. They provide real-world data and insights that can’t be gleaned from hypothetical scenarios or projections. This learning feeds back into the strategic process, helping to refine the technology application, improve implementation strategies, and better align with business objectives.

So, how do we identify pilots?

Criteria for Selecting High-Impact Pilots

Selecting high-impact pilots means focusing on measurable outcomes, which is crucial for organizations looking to integrate emerging technologies effectively.

Here’s a proposed framework to guide this process:

  1. Alignment with Strategic Objectives: The pilot should directly contribute to the organization’s strategic goals, such as improving customer experience, increasing operational efficiency, or driving innovation.
  2. Executive Sponsorship: Identify a respected and willing executive who can champion the pilot. Their support can be pivotal in gaining organizational buy-in and showcasing the pilot’s impact.
  3. Feasibility and Scalability: Assess the technical and resource feasibility of the pilot. It should be challenging yet achievable and have the potential for scaling up based on success.
  4. Measurable Outcomes: Define clear, quantifiable metrics for success. These could include metrics like improved efficiency, reduced costs, increased revenue, or enhanced customer satisfaction.
  5. Visibility and Influence: Choose visible pilots within the organization and can influence broader perceptions and attitudes towards technology adoption.
  6. Potential for Quick Wins: Prioritize pilots that can deliver quick, visible results. Early successes can generate momentum and build confidence in the technology.
  7. Risk Mitigation: Consider the potential risks associated with the pilot and plan for mitigating these risks. This includes evaluating the impact of potential failures and having contingency plans.
  8. Stakeholder Engagement: Involve relevant stakeholders in the pilot selection process. Their insights can help identify areas where technology can impact the most.
  9. Problem-Solving Orientation: The pilot should address a specific, well-defined problem or opportunity within the organization.
  10. Data Availability and Quality: Ensure sufficient quality data to support the pilot, particularly for AI-driven projects.

When the baseline data is unavailable, we can select a pilot using a combination of qualitative and quantitative measures. Qualitative measures may include employee or customer feedback, observed inefficiencies, or areas where competitors excel. Quantitative Measures involve concrete data points like process times, error rates, or sales figures.

Things to keep in mind when selecting a pilot:

Flexibility in Assessment and Selection: Focusing on areas championed by influential stakeholders can be an effective alternative when detailed assessments are not feasible. Their advocacy and the pilot’s success in their domain can create a compelling narrative for broader technology adoption across the organization.

Emphasizing the Ripple Effect of Success: Success in a pilot project, especially backed by a respected executive, can create a ripple effect, encouraging wider acceptance and enthusiasm for technology integration. The success story told by an internal champion can be far more impactful than any external advocacy, underscoring the real-world benefits and potential of the technology.


After the initial success of a pilot project integrating emerging technologies, the journey toward broader integration or implementation begins. This phase is not just about expansion; it’s a continual process of refinement, learning, and alignment with the organization's strategic goals.

The first step involves an assessment of the pilot’s impact. This means going beyond just looking at whether the objectives were met. For instance, if a pilot project in customer service reduced response times, measuring the time saved and the impact on customer satisfaction and employee workload is crucial. This dual approach helps understand the full spectrum of the pilot’s effectiveness.

Success in the pilot often leads to consideration of scalability. However, this isn’t a simple replication of the pilot on a larger scale. Each organization must tailor its expansion plans depending on its size, industry, and specific circumstances. For instance, a successful pilot in a small department might lead to a gradual rollout in similar departments, carefully monitoring the results at each step.

Jagged Technological Frontier — Image generated in MidJourney by author.

Adopting an agile methodology is critical during this phase. Feedback loops become essential. Regular input from stakeholders, including leadership, helps quickly identify areas that need adjustments. This agile response is vital, allowing the organization to adapt to changing business needs and stakeholder expectations.

Gaining leadership buy-in is another critical aspect of this phase. Communicating the successes of the pilot in a language that resonates with the leadership — in terms of ROI, efficiency gains, or customer satisfaction improvements — can secure the necessary support for broader implementation. It’s about aligning the expanded use of technology with the organization’s broader goals and demonstrating how it contributes to achieving these objectives.

Refining the strategy based on the feedback received is an ongoing process. It might involve adjusting the technology application, revising implementation strategies, or redefining some strategic objectives. The priority should always be to focus on the most important areas to stakeholders and leadership. For example, if a pilot shows qualitative improvements but leadership seeks more quantitative data, subsequent initiatives might focus on generating and analyzing more data-centric outcomes.

Developing a comprehensive roadmap becomes essential as the organization prepares for long-term technology integration. This roadmap should outline the technological aspects and consider the resources needed for a successful integration — budget, personnel, and any additional technology requirements. Risk management is an integral part of this phase. Identifying potential risks associated with scaling the technology and having a clear mitigation strategy is crucial. Moreover, contingency planning is necessary to prepare for unforeseen challenges during expansion.

Finally, the continuous measurement of the value generated by the technology integration is crucial. This should be a combination of both quantitative and qualitative measures. Regular updates to leadership and stakeholders on the progress and value generated ensure transparency and help maintain the momentum gained from the pilot’s success.

The iterative phase is about building upon initial successes, constantly refining the strategy based on feedback, and preparing for a long-term, sustainable integration of emerging technologies. It emphasizes the need for a flexible and adaptive approach, recognizing that no one-size-fits-all strategy exists in this dynamic technological landscape.


Throughout this article, we’ve discussed the need to balance the innovation that technology inspires and the realities of executive decision-making. I encourage data scientists and technologists to view their work through the lens of business impact and for executives to appreciate Generative AI's capabilities and considerations. By doing so, organizations will foster a culture of innovation and be poised to harness the ripple effects of their early successes, paving the way for widespread adoption and deep integration of these transformative technologies.

In summary, integrating Generative AI into core business processes is not merely an IT upgrade — it’s a strategic imperative that demands vision, leadership, and a relentless pursuit of excellence.

Let this be a call to action: strategizing with foresight, innovating with intent, and executing with agility. The rewards are not just incremental gains but the redefinition of what’s possible in their domains.