Organizations may encounter challenges in managing enterprise data; however, they can devise strategies to optimize this strategic asset.
The beating heart of AI is data, and it has the potential to enable tremendous business value. However, managing data can be a challenging endeavor, even when it appears to be relatively simple.
Juan Tello, a principal at Deloitte Consulting LLP and chief data officer for Deloitte LLP, asserts that AI is a critical endeavor that affects all organizations, and data is the source of power for these algorithms. “In our capacity as leaders, it is imperative that we are prepared to engage in a candid and open dialogue regarding the challenges that data presents—including its perceived unpleasantness—and devise strategies to resolve them.”
Nitin Mittal, a principal at Deloitte Consulting LLP, concurs. “He asserts that leaders emphasize the significance of being an AI-fueled organization, and data serves as the catalyst for competitive advantage and differentiation through AI.”
Dr. Alex “Sandy” Pentland, the director of MIT’s Human Dynamics Laboratory and the MIT Media Lab Entrepreneurship Program, further comments, “There is an immense amount of excitement surrounding all things data, and AI can be frightening due to concerns about job security.” “However, there are measures that organizations can implement to enhance their chances of success with AI and data.”
Tello, Mittal, Pentland, and Lacy Blalock, a vice president at Deloitte Consulting LLP, discussed five unflattering truths about data during an episode of Deloitte’s LinkedIn Live series, “Winning With Data.” These truths are the most challenging data conundrums that organizations face in the pursuit of their AI-enabled aspirations, and they provided strategies for overcome them.
Ugly Truth #1: The accurate determination of data’s true value can be a difficult task
According to Pentland, it is crucial to acknowledge that data is no longer a mere “extra,” but rather a primary means of production that is comparable to money, labor, land, or facilities. “Organizations should carefully evaluate the allocation of their data and AI resources and adjust their company culture to prioritize data as a critical component of any solution.”
Revenue, growth, and positive consumer experiences—the list is endless—are all highly valued by organizations. However, Blalock emphasizes that all of this is contingent upon data.
According to Mittal, the process of identifying the data associated with a particular function, assigning it a meaning, and calculating its fundamental value was significantly simpler during the period in which data was digitized as part of enterprise system implementation. “Today, the velocity, variety, and volume of data we receive—whether from digital apps, internet of things sensors, the edge, or other inputs—make that much more difficult,” he asserts.
Mittal asserts that leaders can more effectively connect the underlying data to specific value elements in their organization if they have a better understanding of the data. “The initial step is to establish a clear understanding of the organization’s objectives, which may encompass the transformation of the experience, the acceleration of revenue growth, the enhancement of cost efficiency, or the promotion of innovation,” he asserts.
Tello asserts that leaders must comprehend the precise data necessary to achieve their desired result. “Data will never be incorporated into the fabric of how we measure and drive business value unless we link business benefits to the data that enables.”
Pentland observes that there is a trend among certain organizations to include data assets on their balance statements. “Eventually, the system will experience increased efficiencies and a more liquid market for data assets as a component of the overall economy as companies become more accustomed to sharing insights,” he notes. “Although it may appear to be a distant future, the implementation of such a system could result in more consistent data pricing.”
Ugly Truth #2: Systems Design Frequently Ignores Insights and Data
According to Mittal, it is equally crucial to consider data when designing systems, as they are typically designed in accordance with the aspirations of consumers, stakeholders, and the marketplace. “Frequently, we overengineer systems rather than establishing a foundational understanding of the data that is available, the data that can be utilized, and the information that the data provides,” he asserts.
Tello concurs. “Organizations are merely stabs in the dark when they engineer systems based on an off-the-shelf product without considering the data and insights that are available,” he asserts. “Organizations may occasionally implement a system without first identifying the issues they wish to resolve, resulting in a lack of buy-in, regardless of whether it is an ERP, CRM, or other major system.”
Ugly Truth #3: A scarcity of multidisciplinary skills can impede the progress of data and AI initiatives
Without the appropriate personnel in place across business and technology functions, organizations will face significant challenges, regardless of the amount of funding that leaders allocate to address a data issue.
“To succeed with data, it is necessary to possess a comprehensive understanding of technology and a granular, functional understanding of the business,” asserts Tello. Organizations should make a concerted effort to cultivate talent with a converged set of skills, as skills shortages are present in both sectors.
To succeed with data, it is necessary to possess a comprehensive comprehension of technology and a granular, functional understanding of the business.
— Juan Tello, Chief Data Officer, Deloitte LLP
Mittal asserts that the concept of data fluency should not be restricted to the domains of data scientists or data engineers. “Core data literacy should be an essential component of the skill set of all individuals.” “They should all comprehend the significance of data and its potential benefits, which subsequently enables them to regard it as a strategic asset,” he asserts.
According to Blalock, it is challenging to identify individuals who possess a strategic mindset and can overcome organizational barriers in order to make data an enterprise asset and advance data initiatives. “It necessitates a fundamental change in the way we develop our talent to guarantee that they are proficient in data, have a business impact mindset, and be able to transition into any industry,” she says.
Ugly Truth #4: Scaling and operationalizing data fluency and literacy can be challenging.
“Scaling and operationalizing data literacy should not be as complex as calculating the value of data.” Tello asserts that the required tasks are relatively simple. “However, it is crucial to avoid becoming overwhelmed by the organization’s global reach or size, as it is a significant amount of work.” He suggests that concentrating on the end user is an effective approach to overcoming this obstacle.
Mittal emphasizes that “data is the lifeblood of organizations and the catalyst for innovation.” However, the majority of individuals concur that data is essential; however, it can be overwhelming. “Organizations must prioritize outcomes in order to more effectively operationalize, scale, and leverage data as a strategic asset,” he asserts.
According to Pentland, governance is also a critical factor. “Data governance is the most prevalent data-related concern that I have encountered, as evidenced by the statements of former presidents and prime ministers.” However, data require no longer be classified as a “black box.” He asserts that it is significantly more auditable. “You would never establish an accounting system that you were unable to audit.” Nevertheless, we implement this approach with our artificial intelligence systems. Organizations should commence with an understanding of the functions they anticipate an AI system to perform, the methods by which they intend to audit it, the potential consequences of a system that deviates from its intended course, and the methods by which they can adapt the system as time progresses.
Ugly Truth #5: Data Silos Can Result from Data Privacy and Security
Blalock asserts that numerous data-related measures are intended to safeguard data; however, “data silos may arise when data sharing is restricted to those who require it, as is the case in numerous government and military environments.”
Mittal asserts that the liquidity of data is decreasing as a result of the proliferation of data regulations and data sovereignty regulations. “However, this does not necessitate that the insights derived from the data be. He suggests that in order to achieve a correct balance, we should alter the dialogue to focus on the sharing of insights rather than basketball stars data.
Tello asserts that organizations must begin sharing information more extensively, extending beyond their suppliers, customers, and consumers. “They should endeavor to strike a balance between safeguarding data at the appropriate levels and facilitating essential data sharing.” He claims that this will catalyze the emergence of a new era of AI.
Pentland asserts that this paradigm shift can enable organizations to investigate API-based innovations that facilitate the extraction of data from various divisions, thereby establishing the foundation for federated AI and machine learning. “The challenge is frequently influenced by the culture of the company.” He argues that organizations are frequently hesitant to share their data externally, despite the fact that they would benefit from the realization that they almost never need to share their data in its entirety—only their insights.
Pentland asserts that individuals are less inclined to perceive their data as being removed from them when they are requested to provide insights, which in turn reduces their perception of job insecurity. “In reality, this approach has the potential to enhance the value of their positions.”