SAS Viya Workbench is a lightweight development environment to execute code quickly and efficiently in a cloud-native, efficient and secure environment. SAS App Factory is a rapid application development productivity framework for creating fit AI-driven applications.
LAS VEGAS – On Tuesday, at their SAS Explore event here, AI and analytics vendor SAS started the show by introducing new capabilities within their Viya ecosystem, their core platform moving forward. The two new SaaS offerings are both considerably more than nice new features. SAS Viya Workbench is a lightweight development environment that spins up and executes code quickly and efficientry. in a cloud-native, efficient and secure way. SAS App Factory is a rapid application development environment for creating fit AI-driven applications, with several of the apps created so far in early stage access being highlighted at the event.
“Who is SAS today?” Bryan Harris, Executive Vice President and Chief Technology Officer at SAS, asked rhetorically in his keynote. “We deliver tangible, real business value with the fastest analytics platform in the market, and we are unifying our entire portfolio to have a single unified operating model with our software. I want our software to be unified software. We still have SAS 9 and a lot of customers are still on it, but we want to get them together.”
Harris discussed how Generative AI is being adapted into SAS’s core capabilities.
“We are excited about our work in this space,” he said. “It is very calculated with respect to our vision and approach. There are three things we believe provide Generative AI value – synthetic data generation, digital twin simulations, and large language models. We believe these three offer practical approaches compared with other Generative AI approaches in the market.”
Harris then detailed how all this fits together.
“Synthetic Data Generation can create more robust models and add additional things to see how models react,” he said. “It also serves as the foundation for digital twins. Digital twins take enterprise workflows and simulate that, putting synthetic data sets into an environment that can interact to ask ‘what if.’ It lets you ask questions to see if you will be disrupted.
“Large language models apply to industry models as both builders and consumers of AI,” Harris continued. “For builders, it gives the capability for conversations about creating AI for consumers. For industry-specific use cases and workflows, it combines Generative AI with traditional predictive AI, which is what our expansion of our partnership with Microsoft is all about, moving forward the ability to increase Generative AI workflows and providing the best of both worlds in the market.”
“With Viya as the foundation we are creating a powerful ecosystem to provide the power of AI and data management platform with no code, low code and yes code options,” said Udo Sglavo, VP of Advanced Analytics and Development at SAS. “Today, we are announcing two new SaaS offerings that will be generally available in early 2024, which create lightweight environments in the user’s preferred program environment.”
WorkBench executes code in a cloud-native efficient and secure way.
“We want to take computer science out of data science,” Sglavo said, “You don’t need to rewrite existing platform code – with one flip of a switch, regardless of language, you can use code to build model using tabular data. It gives access to data scientists and developers in their preferred language of choice.
“Workbench is about expanding our total addressable market,” Harris noted. “We were communicating too narrowly to the market here before. We want every type of developer and language to use Workbenches.”
Harris also emphasized that Workbench isn’t just an enterprise tool.
“We think Workbench addresses SMBs as well,” he said. “Viya on Azure is pay as you go and very cost effective.”
Workbench also provides flexibility in the development environment with plans to include three clients – Jupyter Notebook, Visual Studio Code and SAS Enterprise Guide. SAS Viya Workbench is currently available under private preview.
The other new offering, App Factory, is used to create specialized AI-driven applications. It lets developers build and modernize product quality cloud applications quickly and with minimal cost. It is also scheduled for availability at the end of Q1, 2024.
“App Factory is a productivity environment for developing applications, and the efficiency with which you can create assets and get them into production is really shocking, Harris said. “Many don’t yet understand what it can do for customers. It lets them focus on the problem and not the minutia of putting a tech stack together, allowing them to put billion dollar solutions together faster and more effectively. “It lets you import your specific data model into App Factory and easily deploy it into production. It also doesn’t get in your way with AI production. It gets you on your way.”
One of the reference customers for this, Cambridge University, is using it to treat kidney disease by being able to score every biopsy from a potential donor kidney, and quickly diagnose if a potential donor will be a good match for a transplant
“We allow them to accelerate their productivity while adapting to their workflows,” Sglavo said.
Nik Markovic Director R&D at SAS, explained how App Factory is customized to different developer needs with the no code, low code and yes code options
“With no and low code, they say you don’t need to be a developer to write software, and that is true to some extent,” he said. “But that will only get you 80% of the way there. You still need a real code base. But coding today is way easier. App Factory tries to build everything according to open standards, with no lock in or requirement to use any particular framework, and no hyperscaler lock. We want the developer experience to be so good they will use our apps.”
All this will help fulfill a key objective – to demonstrate that while Generative AI is still an early stage technology, it still has value today, and customers who invest in it are not wasting their money.
“We are at the peak of the hype cycle, which makes it critical that companies achieve tangible results and deliver that last mile of value,” Harris said. “Generative AI is just one way to meet customer needs.”
He made a comparison to the early days of the transition to the cloud
“We have one of the largest on-prem customer bases in the world,” he stated. We remade our software as cloud native, but it wasn’t cheaper, like everyone was promising. It was actually more expensive, and people just keep buying more. We analyzed and optimized our analytics over the last two years. Generative AI is another enabling technology for that. To do that, you have to have deep domain understanding.”