Transforming Business with AI: How to Grow $Billion Algorithms and Escape Pilot/MVP Gravity

I vividly remember when a business president called me because an AI model was not working. This was in 2016 just before Chinese New Year, when I was leading Digital & Technology for our Greater China business unit... and it was a great day in my AI journey which started in 2000 when I was a college student. In today's rapidly evolving digital landscape, Artificial Intelligence (AI) stands at the forefront of revolutionary changes across industries. The concept of $billion algorithms is pivotal to understanding the enormous value AI can bring to a large company. These algorithms are not just pieces of code; they are strategic assets that can dictate the direction and success of a business by unlocking new capabilities, optimizing operations, growing sales, creating superior customer experiences, and creating unprecedented competitive advantage. The role of proprietary data to train and operate these algorithms at scale is also more important than ever. To truly transform a business, however, it's essential for humans and AI to work together to scale these AI models across the company, ensuring that their impact is not isolated but widespread, contributing to a significant portion of how a company operates and where growth/profit comes from. Of course, you need to start small, and learn fast with pilots/MVPs but the very same pilots/MVPs create a gravity companies need to escape from for true success.

The Essence of $Billion Algorithms

If you want to understand how advanced a company is using AI/ML or GenAI ask them this simple question: How many million or billion-dollar algorithms are running in the company? Million and billion-dollar algorithms refer to the AI models that either directly generate or have the potential to generate $millions to $billions in revenue or cost savings for a company. They are typically easy to spot in large tech companies and often these companies are built on algorithms. Think of Google's page rank algorithm or Amazon's product recommendation engine. For consumer companies, it might be a model to predict best audiences for media targeting, or a demand vs supply optimization model to satisfy customer orders while minimizing inventory, or a personalization engine for direct to consumer selling, or a pricing optimization model for sales and marketing teams. These algorithms become the backbone of strategic business decisions, enabling companies to automate complex processes, enhance decision-making with predictive analytics, and personalize customer experiences at scale. In 2018, when I took the Global Digital Leadership at P&G, I asked this question: What could be a $billion algorithm? Over the next 2 years following that, we went on to building P&G's first $billion algorithm which was the proprietary media targeting engine predicting the right audience, right channel, right bid, and the right brand / content to display as well as the underlying Real-time Consumer Data, Media and AI/ML ops Platform to support it. In Whirlpool, as we established our Data & AI strategy, we did a similar exercise on Value vs Complexity to build few core AI applications. The distinction between $million and $billion algorithms lies not just in their immediate financial impact but in their scalability and the breadth of their application across the business.

Scaling (Gen)AI to $Billion Algorithms for Sustainable Impact

When I was preparing for a Data & AI strategy presentation for the board, a senior executive stated that we should select and pilot a set of models in different parts of the value chain. While this is a good approach, I rarely see companies which do not already have these pilots in place. In fact years ago during a business trip, a data scientist told me "the demand forecasting algorithm she created has been out performing existing legacy model for more than a year but it was still only being used for just one customer/brand". Typically what differentiates leaders from laggards is not whether they have pilots in place but rather how well they codify and scale these models across the company. For AI to truly transform a business in a sustainable way, it must be scaled across. This scaling ensures that AI's benefits are integrated into the core of the company's operations and how a particular business function is run... switching from old ways to new ways for the company and the organization. Many AI models / algorithms cannot escape the Pilot / MVP gravity in the funnel below. They stay in this stage for a long time wasting company resources and causing an even bigger opportunity cost. To get to $Million and $Billion algorithms you need to elevate the strategic importance and the scaled value, making these AI models the new way to run the business. This approach not only accelerates the pace of transformation but also maximizes the return on investment in AI technologies.

5 Building Blocks on How to Grow $Billion Algorithms for Successful AI Transformations

#1. Change Agents with Conviction, Mastery, and Perseverance

Transformations often start with leaders who operate at the intersection of "what is need" + "what is possible" + "what is breakthrough". They are well connected with companies strategic direction, business model, and operations. They have good visibility on what are the big problems or high value ideas. They then bring the art of what is possible, which requires technology and digital leaders with technical mastery. Then they need to have conviction for what is breakthrough, envision and inspire the organization, and own it to drive it home with perseverance, enabling and empowering teams through a bumpy road.

#2. Executive Alignment and Multi-functional Partnerships

Successful AI transformations begin with aligning AI initiatives with the company's strategic goals and securing unwavering support from the top leadership. This alignment ensures that AI projects are not just technological endeavors but are integral to the company's broader objectives. Leadership commitment is crucial for securing the necessary resources, fostering a culture of innovation, and navigating the organizational changes that accompany AI integration. Often this requires underlying multi-functional partnerships. AI transformation is inherently multi-disciplinary, requiring collaboration between data scientists, IT professionals, business analysts, and operational teams. Establishing cross-functional teams promotes the integration of AI into various business processes and encourages the sharing of insights and best practices. This collaboration is essential for identifying high-impact use cases and ensuring that AI solutions are designed with a deep understanding of business needs. Do you have an idea to step change how marketing is done? You better have a senior marketing executive partnering with you. Will you entrust your demand and supply forecasting to AI? You better start with a supply chain executive who has the reason to believe.

#3. Quality Data, AI/ML Models, and Cloud Infrastructure

At the heart of every $Billion algorithm is data. As AI/ML models are democratized and get commoditized by big tech and open source community, companies need quality proprietary data to train, optimize, and run $Billion algorithms at scale. Building a robust data infrastructure and a data management system, including people and process, are fundamental to developing effective AI algorithms. This includes implementing systems for data collection, storage, and processing, as well as protocols for maintaining data accuracy, privacy, and security. Without a solid data foundation, even the most sophisticated AI models cannot deliver their full potential. You also need to transform to become a data company creating a virtuous cycle to grow your company with data and AI/ML and have your business model generate more data in exchange feeding exponential growth. Data itself is a commercial asset and beyond supporting your own business operations, you can commercialize it for value, which becomes an even more important asset in GenAI era.

#4. Talent Development, Acquisition and Strategic External Partnerships

The success of AI transformation heavily relies on having the right talent. This involves both developing the AI and data literacy of existing staff and attracting new talent with specialized skills in AI, machine learning, and data analytics. Investing in continuous learning and development ensures that the organization can keep pace with technological advancements and remain competitive. This upskilling needs to happen at all levels across all functions. Also strategic partnerships with cloud players such as Google, AWS or Microsoft as well as Systems Integrators or select SAAS companies is key.

#5. Ethical AI and Governance

As AI becomes more integrated into business operations, addressing ethical considerations and establishing robust governance frameworks is paramount. This includes ensuring fairness, transparency, and accountability in AI systems, as well as compliance with regulatory requirements. You need to have the right trust and safety layer on top of your $billion algorithms. Ethical AI practices not only mitigate risks but also build trust among customers and stakeholders.

When the president of the business calls you on AI/ML ops, it means AI has become core to your business model and mission critical to the company success. Transforming a business with AI is a comprehensive endeavor that goes beyond deploying isolated models. It requires a strategic approach to scaling AI across the organization, building & leveraging million and billion-dollar algorithms for maximum impact, and codification to everyday operations. You need to take learnings from pilots/MVPs quickly and persevere or pivot, and watch out for staying in that stage too long with zombie AI/ML applications. By focusing on these 5 building blocks, companies can navigate the complexities of AI transformation and emerge as leaders in the new digital AI era.