The new data essentials

Data readiness is the top challenge in applying generative AI

Generative AI is changing everything—including how we think about enterprise technology itself. Companies today require a digital core that’s built for both machines and humans, and that can unlock the value of AI at scale across the organization.

The good news is that you already own the most valuable asset you can in the era of generative AI: your data. But is your data ready for generative AI?

With the rapid advancement of generative AI changing the sheer amount and types of data companies need…the road to readiness can seem complex. But there is a clear path forward.

We’ve identified six key things you need to know about how data essentials are changing in the era of generative AI— and what it means for your business—so that you can pull ahead of the pack.

1. Your proprietary data is your competitive advantage

Generative AI foundation models become relevant— and therefore useful—when combined with your company’s own data. This combination unlocks specific insights into your customers, products, and operations, providing a competitive edge.

Tapping historical and real-time institutional knowledge can improve internal decision-making, reduce risks and identify new efficiencies, as well as open up attractive monetization opportunities.

Key things to consider

• Start with a value-led approach—thinking about data as a product—to developing and maintaining accurate and relevant data, to ensure the correct level of investment.
  
  
• Identify the unique data generated at each step of your business processes, and the data that’s needed to differentiate decision making. Think beyond your own first party data to consider 2nd party from partners and 3rd party collected externally.

2. Your unstructured data holds untapped potential

Unstructured data—encompassing formats like text, images, audio and video—is rich with contextual information.

Generative AI excels at processing this data type, transforming it into valuable business insights and applications. Like turning a how-to video into a list of product features, summarizing a voice call or spinning up marketing content.

When combined with structured data, it adds the context needed to enable more human-like communication: it contains signals for tone and personality, look and feel that drives much richer interactions.

For many years, Fortune has rigorously collected and analysed complex financial data on the largest companies in both the US and the world in order to create the iconic Fortune 500® and Fortune Global 500™lists. Over the years, Fortune has amassed a wealth of data that the company offers through a comprehensive analytics spreadsheet product. Accenture and Fortune transformed that business knowledge into an AI-driven platform that can give business leaders access to insights like never before.

Key things to consider

• To unlock the potential of unstructured data, it must be made more available: Extend data architectures, security and governance to make unstructured data more usable across your business.

3. Synthetic data is key to filling in data gaps

AI is hungry for data—and the more complex the task or output, the more data is required. Synthetic data addresses the scarcity of specialized datasets, enabling companies to explore multiple scenarios without the extensive costs associated with real data collection.

For example, a company might use synthetic product and customer data during market-testing to save time and resources. It can also be used for risk-management, designing “what-if” scenarios, and even to remove bias.

Synthetic data also addresses certain data risks. It can be used to train AI models without transgressing privacy if the data is sensitive. In cases where data is regulated, keeping copies of synthetic data rather than the original reduces risk in case of a breach.

Key things to consider

• Generative AI itself can be used to create synthetic data: using a larger LLM to generate the data needed to fine-tune a smaller LLM offers a cost-effective approach, without sacrificing accuracy.

4. Gen AI makes it easier to contextualize and find new relationships in data

So much of today’s data is locked in silos and functional domains, limiting potential and collaboration. Generative AI facilitates the use of cross-functional data, enabling the reinvention of end-to-end business processes that cut across functions and value chains.

Generative AI helps surface the right information at the right time to the right user. Think, how much better would life be if customer service could “see” the required updates based on exact specifications from product R&D. Or marketing could know right away that supply chain can keep up with their promotion.

Access to cross-functional data breaks down boundaries and opens up the organization to new ways of working.

Accenture and a large luxury automotive company teamed up to create a new platform called EKHO that uses generative AI to drive decisions across North America, accelerating productivity and experiences. When an employee asks a question through the platform’s simple interface, it selects the right knowledge base and continues to refine answers based on the user’s feedback. EKHO solves new issues by learning from and applying past scenarios and pulling any new information added to the knowledge bases in real time.

Key things to consider

• Every part of the business must make the shift to make data available, treating it as a product that's packaged to be safe, easy-to-use and able to provide trusted insights.
  
  
• Companies must invest in the architectures and operating model needed to create, use and manage these data products. For example, a semantic layer that captures the context needed to make data easier for humans to understand and for generative AI to work with.

5. Generative AI accelerates data risks

Most new opportunities come with new risk, and generative AI is no exception. It introduces new challenges particularly when it comes to data governance and security. There are a number of common blind spots that organizations must address to mitigate these new risks:

• New data types: Organizations typically use data process designed for structured data, but generative AI introduces new data types and more dynamic data flows, increasing vulnerability.
  
  
• Greater access: Generative AI makes data and AI tools more accessible but lacks safeguards against human error, emphasizing the need for training and a culture of collective responsibility to mitigate risks.
  
  
• Increased attacks: Generative AI introduces new attacks to data whether through creating deep fakes, data poisoning, or even making it easier to de-anonymize data.
  
  
• Maintaining data quality: Data quality in the context of generative AI is an ongoing requirement, not a one-time task. Continuous enforcement of data quality and lineage is essential to ensure scalability and model accuracy.

Accenture’s own Responsible AI Compliance program includes built-in measures to mitigate each of these potential data risks. Our four key program elements include: Establish AI governance; conduct and AI risk assessment; enable a systematic RAI testing program; and ongoing monitoring and compliance of AI.

Key things to consider

• Transparent and public commitment to robust governance and security measures will build trust and enhance brand value.
  
  
• A well-communicated strategy should be paired with training and programmatic tooling (e.g., including emerging capabilities privacy preserving technology).

6. Generative AI, applied to data, jumpstarts data readiness

It’s not just about what your data can do for generative AI, it’s also about what generative AI can do for your data. Applying generative AI to your current data processes data can enhance various aspects of the data supply chain, from capture and curation to consumption.

Generative AI can help summarize and classify business data requirements; automatically generate design documents, test cases and data; and generate runbooks and deployment scripts. It can be used to help users find, contextualize and use data.

It also provides opportunities to leap-frog legacy systems and slow ways of working. For example, generative AI supports the reverse-engineering of an existing system prior to migration and modernization.

Key things to consider

• Applying generative AI broadly across the data supply chain, requires investing in maintaining a knowledge base of data about data (metadata, descriptions, service tickets, etc.)
  
  
• In transforming the data lifecycle, processes like data governance and quality also need to be updated to keep pace.

Mining your data’s full potential

Many companies are sitting on a goldmine in potential generative AI value in the form of their proprietary data. It’s time to dig in. The journey to data-readiness can be accelerated by keeping these six things in mind.

If you’re ready to tap your data’s potential, the first step is to assess your current readiness and determine how fast and how far you want to go with your data strategy. Whether you need to refamiliarize yourself with the fundamentals—or you’re ready to dive into our 12-step data maturity journey—we’re here to light the way.