Artificial intelligence, like humans, needs quality "nutrition" to function effectively. Wrong data types and poor data hygiene can dismantle a company's AI ambitions before a project even starts.

When CIOs and their executive peers feel urgent pressure to advance initiatives, they first need to assess the resilience and strength of their data strategy.

"You can use a general-purpose version of generative AI, or you can train your own large language model," Gartner Senior Director Analyst Roxane Edjlali told CIO Dive. "But in all cases, your only differentiating advantage is data. To maximize that differentiation, you must ensure data is in a state of readiness."

Technology leaders and market analysts suggest that CIOs and their enterprises should ask themselves a series of questions before launching AI projects to ensure the strategy yields expected outcomes:

  1. What do you hope to achieve? How does that goal align with the business strategy?
  2. How well do you understand the distribution of your data?
  3. To what extent does the enterprise want to share its data openly?
  4. In what way should data be accessed?
  5. What is the current organizational structure of the data?
  6. Where does the data originate?
  7. How are privacy and security ensured?
  8. How is data quality quantified and measured?
  9. Who will be responsible for data cleaning?
  10. How will success be continuously monitored?

For enterprises, bringing data strategy up to par comes at a high cost.

Weak data frameworks directly lead to poor performance and business outcomes. According to a SoftServe report, about two-thirds of decision-makers believe no one within their organization truly understands the mechanisms of data collection or access. Nearly three-fifths (3/5) of business leaders say inaccurate or inconsistent data is affecting critical decision-making.

Poor data management not only hinders corporate ambitions, but these flawed practices also erode trust and drive up costs, a Semarchy report reveals. Enterprises that push forward before resolving data ailments are only prolonging their cycle of pain.

Vendors are actively positioning themselves to help enterprises streamline data strategies. Tim Guleri, managing partner at early-stage venture firm Sierra Ventures, notes that established vendors are seeking partnerships with AI startups to attract enterprise customers.

Large vendors have distribution channels and an existing customer base, but often lag behind startups in innovation speed. Meanwhile, startups are typically at the forefront of innovation but need to put in more effort to earn the trust of enterprise-level clients. Alliances between the two are expected to become increasingly common.

"This is exactly the risk mitigation that large enterprises are seeking," Guleri said.

However, regardless of who assists, getting a data strategy on track is a formidable task.

"There are almost no shortcuts when it comes to maintaining data quality, solidifying the data foundation, and ensuring data connectivity," Shiyi Pickrell, senior vice president of data and AI at Expedia Group, told CIO Dive. The travel company, which operates brands like Vrbo, Orbitz, and Travelocity, has invested significant effort in integrating data lakes and data assets across different brands.

"We went through a very difficult phase to bring data together," Pickrell said. "We really did the hard, unglamorous work first—getting data connected, then improving data quality and usability—so that we can truly leverage the latest technologies like generative AI and large language models."