Survey of over 9,000 customers worldwide reveals AI usage trends: Chatbots increasing in LLM apps

Python remains the preferred AI language among developers: Unstructured data processing volume increased by 123%

Snowflake, a global data cloud company, has announced that the proportion of chatbots among Large Language Model (LLM) applications has increased by 46% since last May.

https://seenthis.kr/newspage/2448

Snowflake has published its "Data Trend 2024" report based on a study of over 9,000 customers to analyze patterns and trends in data and AI adoption. This report, which examines how global enterprise companies are leveraging AI technology and data in their businesses, analyzes that the proportion of text-based LLM applications is decreasing (82% in 2023, 54% in 2024) while interactive chatbots are increasing.

Additionally, a survey conducted among the Streamlit developer community revealed that approximately 65% of respondents are working on LLM projects for work purposes. In reality, enterprise customers are widely using generative AI-based technologies to enhance work productivity, efficiency, and analytical capabilities.

Jennifer Belissent, Snowflake's Senior Data Strategist, remarked, "Interactive applications are programmed in a way that mimics real human interaction, enabling people to easily interact with LLMs as if they were having a conversation. Particularly, if the governance and security of the underlying data for LLM applications are guaranteed, interactive applications are expected to expand in their usage, satisfying the expectations of both businesses and users."

Construction of over 33,000 LLM applications in 9 months

Over 20,000 developers worldwide belonging to the Snowflake Streamlit community have built over 33,143 LLM applications in the past 9 months. Their preferred programming language was Python. Python usage on Snowpark, Snowflake's application building platform, surged 5.71 times last year. This is a higher figure compared to Scala and Java, which saw increases of 3.87 times and 1.31 times, respectively. Python accelerates prototyping and testing speeds, boosting overall task speed as well as the overall learning speed during the initial stages of cutting-edge AI projects.

Recently, there's been an increase in cases of developing LLM apps directly on data management platforms. The usage of Snowflake Native Apps, which enables direct development of applications on the Snowflake platform, has increased more than threefold between July of last year and January of this year, spanning a period of seven months. Applications developed on a single data platform eliminate the need to export data copies to external vendors, thereby speeding up app development and deployment, and reducing operational and maintenance costs.

Growing importance of data governance within enterprises

Alongside the adoption of AI, analysis and processing of unstructured data within organizations have also increased. This process can lead to the discovery of untapped new data sources, and data governance has become crucial for protecting sensitive personal data. According to Snowflake's report, businesses have processed 1.23 times more unstructured data over the past year. Global market research firm IDC estimates that up to 90% of global data consists of unstructured video, images, and documents. Unlike structured data, where language model training is already commonplace, processing unstructured data represents a new opportunity for businesses to enhance their competitive edge.

Jennifer Belissent, Snowflake's Senior Data Strategist, emphasized, "Data governance isn't about controlling data; it's ultimately about leveraging the value of data." She explained, "Snowflake categorizes governance into three areas: data collection, data security, and data utilization. Customers can tag and classify data to apply access and usage policies suitable for their businesses." Consequently, the number of Snowflake customers who have adopted data governance capabilities has increased from 70% to 100%, and the number of queries operating within governance has also surged by 142%.

Furthermore, Belissent added, "Each piece of data reflects how businesses respond to the challenges they face. When these individual data points are viewed as a whole, they enable organizations to formulate comprehensive strategies that leverage the latest AI technology opportunities. The core strategy in the age of generative AI is not about fundamentally changing data, but rather about putting that strategy into action immediately. To this end, businesses need to open and share data sources scattered across their vast data ecosystems to dismantle data silos."

The full "Snowflake Data Trend 2024 Report" can be found here.

Website: https://www.snowflake.com/?lang=ko

Contact
Snowflake PR Agency
KPR
Su-bin Kim
02-3426-2281