How Solid Data Strategies are Fueling Generative AI Innovation

If innovation is the ultimate goal in business and technology today, then consider generative AI (gen AI) the vehicle taking us there — and a strong data strategy, the fuel. Despite all its promise of productivity gains and new discoveries, gen AI alone can’t do it all. The technology needs a “very ready” data foundation to feed on, something the vast majority of businesses today (78%) do not possess, according to a new report by MIT Technology Review Insights, in partnership with Snowflake.

The report surveyed more than 275 global business leaders, from a broad range of industries, about their hopes for gen AI. From that survey, it found 4 out of 5 businesses aren’t ready to capitalize on the technology’s benefits because of poor data foundations. 

In my conversations with customers, I’ve seen organizations look at gen AI as a way to reshape how they operate and what they sell. But this isn’t possible without a strong data foundation, the core of gen AI’s capabilities. Business leaders must move on this quickly to combat the concerns with AI adoption, such as data security and cost, and establish the foundation they need to deliver on the technology’s promise.

Our joint research with MIT shows that as organizations feel increased urgency to deploy AI applications, they’re realizing that their data can help deliver insights from previously untapped sources of information. The survey broke down the priorities of AI customers, revealing where organizations are focusing their efforts and what they expect from AI solutions.

Organizations demand trust and accountability from AI

One of the most striking revelations from the survey was that 59% of the leaders prioritize data governance, security or privacy above all else. This speaks volumes about the level of trust and accountability that organizations expect when integrating AI into their operations. Following closely behind are quality concerns — particularly reducing the phenomenon of “hallucinations” in AI outputs — and cost management. These are all areas Snowflake has been working extensively to address.

We’ve seen organizations already investing heavily in their data infrastructure, recognizing that data is the lifeblood of AI. Our Snowflake platform provides robust support in this regard, offering access controls and secure data sharing that help organizations govern their data effectively.

Our customers can maximize their investments and drive productivity gains by positioning AI capabilities directly alongside their data. For instance, Siemens Energy built an AI chatbot based on a retrieval-augmented generation (RAG) architecture to quickly surface and summarize more than 700,000 pages of internal documents. The tool has helped accelerate research and development and transformed previously locked information into a searchable asset for their research organization. This not only boosts productivity but also fosters innovation across various sectors.

Interestingly, while organizations navigate some of the challenges, they also have ambitious aspirations for gen AI because of success stories like that of Siemens Energy. 

Specifically, nearly 3 out of every 4 respondents (72%) said they want to enhance their efficiency, while 47% focused on developing better products and services through AI. These findings highlight a dual focus: the foundational need for secure and trustworthy AI and a relentless drive for innovation and improvement.

Overcoming the challenges of AI adoption

Despite the optimistic outlook, many organizations still face hurdles in AI adoption, particularly around data governance and privacy. Survey respondents said these challenges were significant obstacles to scaling their AI initiatives. To navigate this landscape, organizations must establish a governance framework that fosters trust while ensuring compliance.

We’ve seen organizations increasingly implement rigorous access controls and data governance protocols to great effect. For example, consider an HR chatbot designed to answer sensitive inquiries, such as promotions within a company. Ensuring that the right information reaches the right user while maintaining data privacy is crucial. Building robust security measures into AI systems allows organizations to harness the power of AI without compromising on trust.

The imperative of data quality

Data quality is another critical area that organizations must prioritize. AI accuracy is non-negotiable. To address this, Snowflake is focused on providing platforms that help our customers build AI systems capable of distinguishing between when to respond confidently or when to seek clarification. This capability is vital in preventing misinformation and helping ensure that AI serves as a reliable tool for decision-making.

Looking to the future of AI

As we consider the future of AI, the opportunities are boundless. We’re currently seeing productivity gains across various domains, from customer service enhancements to insights gleaned from vast data sets. But there is an even more transformative future where AI systems will not only answer questions, but also take proactive actions based on insights derived from data. This shift toward more agentic systems marks an exciting chapter for AI, and we’re eager to lead this charge at Snowflake.

In recent months, we’ve been on a fast-paced journey, rolling out several exciting features like Snowflake Arctic and Snowflake Cortex AI. Snowflake’s goal is to make AI easy, efficient and trusted. By running AI near the data, our solutions prioritize quality and cost-effectiveness, all within a secure framework.

The future of AI holds immense promise, and with the right data strategies and governance in place, organizations can unlock unprecedented value. At Snowflake, we are committed to empowering our customers on this journey, ensuring they can harness the full potential of AI with confidence and security.