New York, NY – July 16, 2021 – Ultimus has demonstrated the ability to generate software applications from simple, plain English sentences using machine learning-based text generation algorithms.
Machine learning is well established in uses cases such as image and speech recognition, sentiment analysis, classification, intelligent document processing, and medical diagnosis.
By leveraging the massive open-source investment in machine learning-based text generation, Ultimus has demonstrated the ability to generate software “code” from simple, plain English prompts.
The Ultimus Digital Process Automation Suite’s low-code development environment, Composed Process Solutions, is based on the concept of generative development – a domain specific language to specify software solution and solution family requirements, a domain model that pre-implements application commonalities, and a generator that merges the two when an application is called.
The Ultimus domain specific language (DSL) approach gives experienced developers virtually unlimited expressive power, as well as massive reuse capabilities. With visual extensions, Ultimus has made its concept accessible to “business technologists”.
Now, with the ability to generate the Ultimus DSL from simple typed or spoken sentences, software development is realistically within the grasp of the Citizen Developer. And the platform’s built-in governance, best practice, consistency, and security make leveraging citizen developers a practical and safe way to augment professional IT developer resources.
Example of simple form controls generated from simple sentence prompts using machine learning
Leigh Michl, Ultimus Chairman & CEO, said “Our AI/ML team got us off to a great start. They’ve proven that our text-based paradigm is ideal for applying AI in software development. That’s going to make implementations even easier and faster. From here, we will be expanding the training set to broaden and strengthen our capabilities. There’s still a long way to go, but the work ahead is straightforward.”