Why AI Transformation at Enterprises is Crawling (and Why it Should Sprint)

The age of AI is here, super charged by the interest in Generative AI, and employees are ready to embrace it. Unlike its predecessors, Gen AI does not just analyze existing data; it creates novel content, insights, and can even contribute to the design of new products, fueling a new wave of innovation and growth. Microsoft's 2024 Work Trend Index?reveals that 79% of business leaders recognize the need to integrate AI to remain competitive. Yet, a significant 59% are grappling with how to measure the return on investment (ROI) of AI initiatives. This uncertainty has created a strategic challenge, with 60% of leaders expressing concerns about the lack of a clear vision for AI implementation within their organizations.

The result? A frustrating paradox. While the workforce, across all age groups, is eager to harness AI's potential to streamline tasks, boost creativity, and drive innovation, many companies are stuck in neutral, hesitant to shift gears towards a full-fledged AI transformation. This is especially true in highly regulated sectors like financial services and healthcare. The technology's inherent risks, such as the potential for biased algorithms, the 'black box' problem of explainability, and the possibility of unintended consequences from autonomous decision-making, are compounded by concerns around data privacy, confidentiality, IP, and copyright.

In this blog post, I will challenge the traditional "crawl, walk, run" approach to change and advocate for a bolder strategy – a "run, walk, crawl" approach that allows enterprises to seize the opportunities AI presents while building the necessary foundation for long-term success.

Rethinking the Approach: From Crawl, Walk, Run to Run, Walk, Crawl

The "crawl, walk, run" approach, while a common change management strategy, can be too slow and cautious in the face of transformative technologies like AI. The AI landscape is evolving at breakneck speed—what was once a frontier model is soon surpassed. As Sam Altman, the CEO of OpenAI, recently shared during a Q&A session at Stanford University, even groundbreaking models like ChatGPT are "mildly embarrassing at best" compared to what's on the horizon.

By the time a company has "crawled" and "walked" its way to the "run" phase, the AI tools and techniques it initially adopted may already be outdated. This diminishes potential ROI and creates a frustrating cycle of playing catch-up.

Instead, I propose a "run, walk, crawl" strategy that allows enterprises to rapidly adopt AI, demonstrate tangible value, and position themselves for long-term innovation and growth:

"Run" – Empower employees with existing AI tools:

Don't wait for perfect solutions. Start by utilizing low-friction entry points for your employees to experiment with AI in a secure and controlled environment.? Look to readily available AI-powered "assistants" (like M365 Copilot) and other SaaS solutions.

Foster a culture of experimentation by encouraging employees to explore how these tools can streamline their workflows, automate tasks, and enhance decision-making, building internal momentum for AI adoption. By taking a ‘human in the loop’ approach—where humans oversee and validate AI outputs—and focusing on internal use cases, you can responsibly experiment with AI while mitigating risks to customer data and avoiding premature public releases. Establish clear guardrails for AI usage, and collaborate closely with risk, compliance, and legal teams to build robust frameworks that evolve as you learn, ensuring a responsible and successful AI transformation.

As Wharton Professor Ethan Mollick suggests in his excellent guide to navigating this new age of AI "Co-Intelligence: Living and Working with AI," inviting AI into your daily work (within legal and ethical boundaries) is crucial to understanding its potential benefits and risks.

"Walk" – Customize and integrate for tangible ROI:

Once your workforce has experienced the potential of AI, it is time to take a more strategic approach. Leverage foundational large language models (LLMs) and low-code platforms to customize AI solutions that align with your specific workflows, datasets, and business objectives.

●????? Focus on tangible ROI: Identify use cases where AI can deliver measurable returns, such as automating customer service, optimizing marketing, or improving onboarding. These early successes validate AI's value and secure further investment.

●????? Start with low-code platforms (such as Microsoft’s Copilot Studio): Utilize low-code or no-code platforms to accelerate the development and deployment of AI solutions without requiring deep Data & AI expertise.

As you experiment with customized AI solutions, you will uncover "big rock" opportunities – novel business models, process improvements, or product innovations that can fundamentally transform your business - using your unique data, insights, and IP - setting the stage for the crawl phase.

"Crawl" – Build or fine-tune your own models (when necessary):

While pre-trained models and customization can address many use cases, some scenarios might require building or fine-tuning your own AI models or the user experience. This is the most resource-intensive phase, but it offers the highest potential for differentiation and sustainable competitive advantage. This phase is about pushing the boundaries of what is possible with AI, combining it with your unique assets and creating proprietary solutions that set you apart from the competition.

A recent study that evaluated the effectiveness of LLMs in performing various financial text analysis concluded that across a number of tasks, foundational models (e.g., GPT4) outperformed purpose build models specific to financial domain (e.g., BloombergGPT). On tasks that needed a deeper understanding of financial language structure, fine-tuned models (e.g., FinBERT) performed better. In a rapidly evolving area, these findings are likely to change as new foundational and purpose build models are developed.

The lessons learned, the organizational muscle built during the earlier phase, as well as closer collaboration with the right partners will give you the confidence to go big and will set you up for longer-term success.

Why This Approach Works

The "run, walk, crawl" methodology addresses key challenges that hinder AI adoption:

●????? Overcoming the ROI Hurdle: The "run" phase focuses on readily available tools with low cost and immediate benefits, providing tangible evidence of AI's value and making it easier to justify larger investments.

●????? Building Momentum and Urgency: Demonstrating the potential of AI early on creates excitement and drives adoption across the organization.

●????? Agility and Adaptability: This approach allows enterprises to adapt to the rapidly changing AI landscape, experiment, learn, and pivot quickly if needed.

The Risk of Hesitation – Why Act Now

In the rapidly evolving landscape of AI, hesitation can be costly. While a cautious approach might seem prudent, it can inadvertently expose your organization to risks and missed opportunities.

Shadow AI and its Perils:

Microsoft’s 2024 Work Trend Index reveals a growing phenomenon: employees are increasingly bringing their own AI tools into the workplace. This "shadow AI" trend poses significant challenges for enterprises.

●????? Data Security and Privacy Concerns: Unsanctioned AI tools often lack robust security measures, putting sensitive company data at risk of unauthorized access, breaches, or even accidental leaks.

●????? Intellectual Property Risks: Employees might inadvertently share confidential information or trade secrets with external AI platforms, jeopardizing your company's competitive advantage.

●????? Compliance and Legal Issues: The use of unapproved AI tools could lead to violations of industry regulations or data protection laws, resulting in hefty fines and reputational damage.

●????? Diffusion of Innovation:? Without adopting a coordinated, intentional approach, while individuals will learn or make mistakes on their own, the organization will progress more slowly.

The Perils of Perfectionism:

While due diligence is essential, excessive planning and analysis can be just as detrimental as reckless experimentation. Spending too much time trying to find the "perfect" use case, build an airtight business case, or establish a comprehensive AI governance framework can lead to missed opportunities.

●????? Falling Behind Competitors: While you deliberate, your competitors might be rapidly adopting AI and gaining a significant edge in efficiency, innovation, and customer service.

●????? Missing the Learning Curve: The field of AI is constantly evolving. By delaying adoption, you miss valuable learning experiences that could inform your future AI strategy.

●????? Stifling Innovation: A culture of over-analysis can stifle creativity and discourage experimentation, preventing your organization from fully realizing the transformative potential of AI.

Learning Amidst the Hype:

Even if you believe that the current AI hype cycle will eventually subside, it is crucial to remember that the real work often begins when the initial hype fades. The most successful AI implementations emerge from a process of continuous learning, adaptation, and refinement.

●????? Embrace the Learning Curve: Do not be afraid to experiment, make mistakes, and learn from them. Every failure is an opportunity to gain valuable insights that can guide your future AI efforts.

●????? Focus on Long-Term Value: Instead of chasing short-term gains, focus on building a sustainable AI ecosystem that can deliver long-term value to your organization.

●????? Invest in Your People: Equip your workforce with the skills and knowledge they need to leverage AI effectively. This will ensure that your organization is well-positioned to capitalize on AI advancements in the years to come.

Call to Action:

The time for hesitation is over. Enterprises need to embrace AI as a strategic imperative and take bold steps towards adoption. By empowering employees with secure AI tools, customizing solutions for tangible ROI, and continuously learning and adapting, you can position your organization for success in the AI-driven future.

Don't let fear or indecision hold you back. The "run, walk, crawl" approach offers a clear and actionable pathway to harnessing the transformative power of AI. Start running today, and you will be well on your way to achieving AI-powered success.?????