Analytics software provider SAS Institute Inc. today is amping up development features in its Viya artificial intelligence and analytics platform and previewing some of the work it is doing with generative AI.

It also said users of its Customer Intelligence 360 customer relationship management application can now integrate with generative AI models of their choice for use in marketing planning, content creation and journey design. The new capabilities in Viya that are being previewed this week at the company’s Explore conference create lightweight environments for rapid AI model building, support multiple programming languages and run in a scalable cloud computing environment.

The SAS Viya Workbench is a cloud-native development platform that permits AI models to be quickly built and deployed, the company said. It supports the Python, R and SAS languages that are popular with data scientists along with Jupyter Notebook, Visual Studio Code and SAS Enterprise Guide automation. The workbench is currently available in private preview with general availability planned for early 2024.

The new SAS App Factory is a rapid application development environment for creating AI-driven applications. It automates the setup and integration of a cloud-native tech stack built on the React JavaScript library, TypeScript programming language and Postgres database that the company said helps speed models and AI-driven applications into production. It’s planned for general availability in 2024.

SAS also said Viya is now available as a private preview in Snowflake Inc.’s Data Cloud with Snowpark Container Services.

The first SAS product leveraging the App Factory, SAS Energy Forecasting Cloud, enables utility planners and managers to generate AI and analytics models that predict peak power demand and provide more accurate forecasts. The application also generates predictive insights into supply and demand to help utilities maintain stability while improving costs.

SAS said Cambridge University Hospitals is using App Factory to increase the success rate of kidney transplants by using computer vision and AI to score biopsies and set priorities for candidate kidneys so histopathologists can more quickly identify the most viable kidneys for transplant.

Bring your own AI

Customer Intelligence 360 customers now have the choice of adding generative AI models from commercial providers such as OpenAI LLC, open source or privately hosted models. They can also choose which capabilities to use and specify how the models are trained, configured and exposed to users via a custom integration framework.

Generative AI can be used to help create marketing plans, identify audience segments, and create content. It can also personalize communications around demographic, psychographic, behavioral and geographic variables as well as conduct testing at scale.

The company is using the conference to highlight other areas of AI it is actively developing. One is a patented extension to generative adversarial networks that creates statistically congruent tabular data that reflects the complexity of real-world environments. The capability is said to be useful in creating digital twins and improves predictive models while reducing costs.

SAS also said it’s developing large language models for vertical industries that preserve data privacy, a major concern of public models like ChatGPT. The company said it’s working with Microsoft Corp. to develop a generative AI integration that combines Microsoft Azure OpenAI with SAS’ enterprise task orchestration and analytics in a product to be previewed in the fourth quarter.

Image: SAS