SAS aims to solve 'last mile' issues with analytics, put more models in production
SAS is launching SAS ModelOps in a bet that enterprises will increasingly have to manage, optimize and watch spending on their analytics, machine learning and artificial intelligence efforts.
According to the company, SAS ModelOps is designed to address the "last mile" challenges with enterprise models. IDC estimates that only 35% of analytics models make it to production. SAS has made a series of moves to raise its profile in the AI market while touting its data science experience.
The SAS ModelOps effort aims to package SAS Model Manager as well as advisory services using open source and SAS analytics models. SAS is also launching ModelOps Health Check Assessment, a service to optimize deployments.
SAS' Model Manager is a suite that enables collaboration on models, scoring, test and production tools as well as code testing and a bevy of other items.
It is increasingly clear that enterprises will need to layer management, deployment, monitoring and governance to handle what is likely to become model sprawl. Tech giants such as IBM are also eyeing the model and data science management market.
The AI, machine learning, and data science conundrum: Who will manage the algorithms? | Executive's guide to prescriptive analytics
SAS is aiming to take its position in the data science market to pitch companies looking to make their data, AI and analytics investments pay off. SAS cited Commerzbank, Connect Financial Services and Telenor as model management customers.
According to SAS, managing models will mean adopting DevOps best practices as well processes knowhow.