For well over a decade, Tom Davenport has been the preeminent thinker on the subjects of data and analytics. His bestseller, "Competing on Analytics: The New Science of Winning," transformed how business leaders consider the role of data and analytics. And his "Analytics at Work" provided the first maturity model for data and analytics. In 2010, when it was published, only 5% of organizations had reached the highest stage of data maturity.
Artificial Intelligence (AI today) is at a very similar point. In the upcoming book from Davenport and Nitin Mittal, "All-in On AI: How Smart Companies Win Big with Artificial Intelligence, (it releases on Jan. 24), they dig into how organizations use AI to transform their businesses. And not unlike Analytics at Work, they find only a small number of firms have reached the zenith of maturity. Davenport and Mittal say less than 1% of large organizations are AI-fueled.
Furthermore, Davenport and Mittal claim these organizations have become learning machines where their employees focus on AI acceleration. To achieve tangible value from AI investment, the authors say companies should rethink the way humans and machines interact within today’s working environments. Similar to the authors of "Designed for Digital," Davenport and Mittal claim AI should drive new product and service offerings plus transformative business models.
Citing research from Deloitte, Davenport and Mittal list the three most common uses of AI:
- Making business processes more efficient.
- Improving decision making.
- Enhancing existing products and services.
The Human Side of AI
Unlike other books on this topic, "All-in On AI" looks at the human side of the equation. Davenport and Mittal claim smart organizations actively manage potential people issues — but only by worrying about people and process can they make substantial progress with their AI agenda.
Importantly, the authors suggest real transformation starts by creating a culture that emphasizes data-driven decisions and actions. People must be enthusiastic about the potential for AI to transform their business while at the same time, their leaders need to make sure the right people are hired and used effectively. Because a critical dimension of success with AI is allowing people to learn and grow, the workforce at AI-mature organizations need to be data literate and data fluent.
Powering these organizations are data leaders throughout the organization — not just within IT. In fact, as "Future Ready" co-author Stephanie Woerner noted in a recent conversation, 50% of the leadership team and at least three members of the board need to be “digital and data savvy.”
A great example of a data leader driving this, according to the "All-in On AI" authors, is Vipin Gopal at Eli Lilly. By interviewing business leaders across Eli Lilly, Gopal gained a concrete understanding of areas of focus for AI, and it provided the corporate endorsement for the project and the support Gopal needed in order to deliver. I recently met Gopal at an event. He has one more superpower: he is extremely humble and gracious.
Related Article: Data Governance for Dummies
Strategy in an AI World
Without question, strategy is no longer something that happens in ivory towers or separated from technology. For this reason, the authors suggest there’s a demand for corporate leaders who ask how AI can improve their business. As well, they should ask what their organization can do with AI to create new offerings to help organizational growth — while also staying abreast of what the vanguard of AI-fueled companies are doing.
Davenport and Mittal suggest AI and strategy are connected in two ways. First is AI enabling business strategy by improving products and services: by augmenting business models, by transforming channels to customers, and by optimizing supply chains. Second is about developing a strategy for AI itself. Clearly, the former products and services is output of AI in enabling digital transformation.
The Role of Data in AI
Davenport and Mittal write, “If AI can fuel the company, data fuels AI” (page 20). Without question, organizations that are serious about AI are also serious about data. This means they are good at collecting, integrating, storing and making it broadly accessible. This is very similar to the views of the authors of "Future Ready," who say for firms that are future ready “data is a strategic asset that is shared and accessible to all in the firm that need it” (page 11).
Treating data as a strategic asset, write Davenport and Mittal, starts by modernizing the data infrastructure for AI. AI-fueled organizations have good data and are using it to transform their business with AI. Increasingly they have unique or proprietary data — this is similar to the argument in of "Beyond Digital" that digital winners have “privileged insights” unique to their firm.
To succeed, AI-fueled companies need to have what Marco Iansiti and Karim Lakhani call a "data pipeline." “This process gathers, inputs, cleans, integrates, processes, and safeguards in a systematic, sustainable, and scalable way” ("Competing in the Age of AI," page 58). Or, in analyst parlance, they have a data fabric. Davenport and Mittal claim that every organization serious about its data has to do the following tasks: “structuring or re-architecting it, putting it on a common platform, and addressing “pesky issues like data quality, duplicated data, and siloed data throughout the company. It’s fair to say that the single biggest obstacle for most organizations scaling AI systems is acquiring, cleaning, and integrating the right data” (page 83).
Related Article: How AI-Driven Data Enhances CX
Machine Learning Approaches
Machine Learning represents a foundational building block of AI-powered companies. According to Deloitte’s research, supervised learning is the most common machine learning approach. This approach trains models using a training data set. It then tests the model’s fidelity against the remainder of the data set. And it then uses the resulting model to make predictions or classifications upon additional data.
Davenport and Mittal want business leaders to know well-developed models are labor-intensive to develop and deploy. With this, Davenport and Mittal are clear it can be challenging to scale models across complex businesses and geographies. Their book provides great examples of the application of AI within industries including consumer, industrials, financial services, government and public sector, life sciences and healthcare, tech, media, and telecommunications.
Building AI Capabilities
Davenport and Mittal assert that no company adopts AI extensively and deeply at once. They candidly admit that success requires experimentation, developing capabilities of times, fits and starts, and mistakes and setbacks. With this they move to defining an AI maturity model. Similar to "Analytics at Work," the model has five components.
Interestingly, Davenport and Mittal suggest that because of the business risks, smart organizations work to be trustworthy and ethical. To address AI risks, governance and policy statements are required. To govern effectively, they suggest governance have the following components: be fair and impartial, be transparent and explainable, be responsible and accountable; be robust and reliable; respect privacy; and be safe and secure. In talking with data governance experts, this is without question the next wave of governance. As such, the authors are at the leading edge here.
So, your management and board have decided to transform and be AI-fueled, are you too late if you start now? Davenport and Mittal assure us that we’re only at the beginning of AI lead transformation. Clearly, no company was powered by AI a decade ago. There is still time to apply AI strategically and in large doses.
At the same time, the authors suggest that AI is critical to almost every business. As well, the organizations that apply it with vigor will dominate their industries for decades. To succeed, the authors practitioners take ten steps to succeeding.
Parting Words: Truly Understanding AI's Potential
"All-in On AI" extends the work of others on the topic of AI. But I think it goes further by providing managerial guidance for building a successful AI program in the eyes of business leaders and suggests the use cases and applications for AI by industry.
As such, it is a great primer for business leaders needing to understand the potential for AI — as well as data scientists who need to garner the most interesting use cases for their industry. Most important, and regardless of the reader’s role, is the concluding guidance on driving business success from AI.
Learn how you can join our contributor community.