A data strategy is essential for every use case, but the rapid evolution of Artificial Intelligence (AI) has significantly bolstered the importance of a well-defined strategy.
All AI capabilities are data-driven, so you might assume one AI-ready data strategy will work for every AI use case. But traditional AI and gen AI have different data requirements. To get the most value from gen AI, you’ll need a data strategy that can help you manage your unstructured data.
That starts with making sense of your data landscape: your data assets, data infrastructure, and current data usage in your business processes. You’ll also need to imbue a culture of data literacy within your organization and empower people through data democratization and a foundational understanding of AI. It’s not the easiest task, but it’s an important and achievable one. The following framework will help you design the right data strategy to achieve your organization’s business goals and create success with AI.
Meet with senior leadership to gain a clear understanding of your organization’s top goals and priorities. These conversations will give you the chance to ask key questions and plot the best direction for your data strategy.
Understanding the quality of your enterprise data and how it flows, or doesn’t, between different areas of the business will allow you to unlock undiscovered business value.
Continue to check in with stakeholders as your data strategy takes shape. Keep their priorities and pain points top of mind.
Identify the most compelling use cases
Aligning the right data with your business objectives “starts and ends with the question, what business problem are you trying to tackle?” says Tony Giordano, who leads data strategy, consulting and transformation engagements for IBM.
As you search for a compelling use case, keep clear, attainable outcomes in mind. Leading CDOs understand what drives their business, and the criticality of connecting data and analytics to priority outcomes.¹
Protect your investments
Take your data strategy to the next level by leveraging your existing infrastructure, technology, and skills. Familiarize yourself with your organization’s technology ecosystem to determine where and how your data can help achieve business outcomes. When you truly understand your data, you can pinpoint outdated data architecture that doesn’t align with your business strategy, take better advantage of funded initiatives, and identify areas for improvement.
Identify barriers and gaps
Once you have your end goals and leadership buy-in, you can identify the barriers to building a true data-first experience. Silos often underlie the challenges with data integration, data management, and workflows. In fact, 81% of IT leaders say that data silos are hindering their digital transformation efforts.²
Data access should not be an obstacle.
Users should have access to the data that yields great outcomes. They shouldn’t have to think about where that data resides or whether it’s governed and compliant. They should be able to use the data they need with confidence.
Design thinking for data strategy
A design-thinking approach helps surface organizational pain points, which brings strategic value across multiple use cases, lines of business, and individual teams. This process helps generate achievable resolutions in a continuous cycle of observation, reflection, and creation, approaching problems and solutions as an ongoing conversation.
Inventory talent and skills
You can’t engineer data changes on your own. Make sure your organization provides ongoing training to keep up with the rapid pace of AI evolution and the IT industry as a whole. An IBM IBV survey found that 85% of leading CDOs are expanding training, 77% are reskilling their internal staff, and 70% are acquiring new talent to increase data literacy across their organizations.³
Prioritize governance
In the gen AI era, you must deliver end-to-end governance. Staying on top of critical, regulated data elements is essential to running your systems without duplication errors, unreliable searches, or privacy breaches. Consider who currently owns, manages, and defines your data policies, and whether that governance affects security, privacy, or compliance. Ensure the appropriate parties have the decision rights, accountability framework, and external resources to manage data effectively.
Define your data’s target state
“Many data environments are outdated and rarely have the flexibility to evolve in today’s digital environment,” says Giordano. A modern data architecture needs to be managed, governed, and secured to ensure consistent data quality. It requires the flexibility to evolve alongside your digital channels.
Measure progress toward your goals
While data leaders are often expected to drive transformational change, their success is measured against tactical, short-term business objectives. According to a CDO survey conducted by AWS, 74% of CDOs say their success is measured in terms of their business-focused achievements or equally divided between business and technology goals, while only 3% say their success is defined exclusively by their technical feats.⁴
Focus on your data objectives. Leverage insights from your data users as you consider the best ways to accelerate business value using AI.
Outline a data governance policy
A robust governance framework is based on quality, privacy, and security. A metadata and governance layer for all data, analytics, and AI initiatives increases visibility and collaboration across your organization, regardless of where your data resides. Your data governance policy will shape behavior around data quality, data privacy, data security, and data management, while also showing you where AI is streamlining your regulatory efforts.
Identify your data advocates
Find the people in your organization who are passionate about the impact data can have on their work. These are your success partners. Get them involved in regular meetings and standardization efforts.
You can find success partners within your data teams—enlist data engineers, data architects, or data scientists who are building AI models. Line-of-business (LOB) leaders whose teams rely on data analysis are also great candidates. They likely have experience using new technologies to help improve business processes and optimize the value of their data.
Set your sprint cycles
For a data and AI strategy to take hold, organizations often need to re-engineer their culture around new concepts and environments.
Start by setting goals that are quickly achievable, valuable, and viable. Assemble your cross-functional team against these objectives. Schedule short sprint cycles with actionable milestones that will help prove results. Finally, make sure your C-suite, technology teams, and business users all have the same finish line in their sights.
Collect small wins
Small, repeatable use cases can help you quickly prove the value of your data and AI investments. There’s no need to tackle your hardest problems at the onset. Use cases that are impactful, but simple, give you the opportunity to gather important insights about your technology and stack up early victories. Invest in pilot programs during the initial stages of AI adoption to gain the experience you need to enhance larger deliverables down the line.
Create a central data catalog
A central catalog stores and shares insights, allowing for simplified data consumption. In the catalog, data is augmented in original and curated forms with purpose-fit storage. Data-access tools look beyond individual apps or processes to assess how your data is being consumed and what knowledge is emerging. This level of detail enables users to make real-time decisions that consider data for every part of the organization.
Empower data consumers to adopt
Use your new data management framework to encourage enterprise-wide adoption. In this way, you’ll be influencing how your business communicates, improving key workflows, optimizing security, and unlocking new business models, market opportunities, and operational efficiencies.
Show and tell
Your use cases will be a significant source of empowerment. As a recent article from Harvard Business Review points out, CDOs and AI leaders see greater success when they “make data everyone’s business.”⁵ So, take advantage of the fact that use cases can span data science, operational analytics, digital transformation, business intelligence, new Generative AI initiatives, and more, allowing multiple teams to leverage data to make a difference in the enterprise.
Hire (and reskill) talent
Closing the skills gap means looking beyond traditional hiring and training strategies. As companies scramble to meet their talent needs, many are adjusting their education and experience requirements just to fill roles. When training and hiring are not enough, consider how your organization can leverage AI and automation to address skill gaps.
Build strong partnerships across the organization
On the most basic level, your job as a data leader is to help your organization make the wisest decisions about data collection, management, and usage. As you build and strengthen partnerships at every level, be open to feedback and collaboration.
Something fascinating happens as you build a data-first enterprise. The more your vision threads into the organization’s DNA, the more you can “let go” by simply supporting a culture in which people are motivated to learn and take on new roles.
Your organization is rallied behind you. As you augment existing technologies and introduce new solutions to simplify data access, remember, you’re doing more than just creating efficiencies and driving new insights—you’re building a culture that’s passionate about using data to its fullest potential.
¹ Turning data into value, IBM Institute for Business Value, April 2023.
² 85% of IT Leaders See AI Boosting Productivity..., Salesforce, January 2024.
³ 2023 Chief Data Officer Study, IBM Institute for Business Value, March 2023.
⁴ CDO Agenda 2024, AWS, Thomas H. Davenport, Randy Bean, & Richard Wang, October 2023.
⁵ Why Chief Data and AI Officers are…, Randy Bean & Allison Sagraves, June 2023