Analyst firm Omdia fingered five leaders in AI (artificial intelligence) and ML (machine learning) development platforms that offer a variety of capabilities for enterprises to use. They are C3.ai, Dataiku, IBM, Microsoft and SAS.
ML and its sub-branch of deep-learning (DL) have left the research laboratory and are being applied in real-world applications,” said Bradley Shimmin, an Omdia analyst in data management and analytics.
He said high-tech internet giants are already leading the field, while many other enterprises are hoping to reap the same benefits in digital transformation. AI is already used to run speech assistant applications like Siri but also to make buying recommendations for consumers, such as saying, “people who bought X also bought Y.” The predictions are based on previous purchases on Amazon or other platforms.
ML is also widely used in various forecasts, such as weather as deployed by IBM with The Weather Channel, for instance. With video surveillance, ML software can help decide which footage to keep and analyze, picking only the scenes where people enter a vacant alley, for instance. Some companies are developing the capability to spot when automated machinery on a production line is showing erratic movements, which could be an indication of a mechanical failure. Or, ML could analyze the electrical current to a device for problems.
In medicine, some hospitals are developing the ability to scan images of lungs for cancer. In a few cities, like Montreal, ML tools have been used to detect cracks on bridges and dying street trees.
To generalize, Shimmin said that AI and ML can automate largescale tasks that manual effort can’t cope with effectively or manage at all, which can result in new business opportunities while transforming existing business practices. The leading ML development providers are generally offering pre-built software as well as augmented and automated development tasks.
Omdia picked the five leading vendors that offer enterprise customers a platform for AI and ML development with products that typically manage the entire cycle of development to deployment.
The Omdia rating gave SAS the top position of the leaders in terms of enterprise execution, with Microsoft and IBM in second and third, followed by C3 and Dataiku in fourth and fifth place. Omdia’s assessment of their technology gave Microsoft the top spot, with nearly a tie with C3.ai, followed by SAS and with IBM and Dataiku tied for fourth. Omdia ranked H20.ai and Petuum out of the top five but described them as “challengers,” while Evolution.AI was described as a “follower.”
Shimmin said that IBM has a complete set of software for build-deploy-monitor-govern lifecycle for ML applications suitable for technical and business roles. He also said IBM has the best of open source ML software, such as TensorFlow and PyTorch, with updates from IBM. Users can rely on Watson OpenScale for monitoring and testing.
SAS, Shimmin said, has a long history as a statistician’s tool of choice, adding it has been “well-positioned to tackle the rose of ML…putting the company into a central role within the enterprise.”
For all its high technology ranking, Microsoft provides ML model creation that is “complex” that requires skills that are scarce in the market, he said. Microsoft’s Automated ML fulfills the automation task, while Azure Machine Learning allows enterprises to scale out their use of ML.
Omdia also found that Dataiku offers software that allows multiple teams with thousands of users to access the same platform. They can be technical coders and non-coders, data scientists and business analysts.
C3.ai’s platform relies on a modular approach known as MDA, which comes with an abstraction layer. Shimmin said C3 can extend and evolve custom solutions, which makes it easier for ML developers and data scientists to focus on core functions.