Microsoft ramps up small language model effort
Microsoft is working with several industry sector specialist software providers to bring industry-specific AI models to its Azure AI platform
In a bid to make artificial intelligence (AI) more applicable, Microsoft has worked with a number of industry platforms to deliver small language models (SLM) to its Azure AI catalogue.
The industry-specific models, built on Microsoft’s Phi family of SLMs, represents an expansion of the company’s industry-tuned AI capabilities, which it said targets specific needs and challenges in sectors ranging from healthcare to finance and manufacturing.
In a blog post discussing the new models, Satish Thomas, corporate vice-president of business and industry solutions at Microsoft, said: “By integrating the Microsoft Cloud with our industry-specific capabilities and a robust ecosystem of partners, we provide a secure approach to advancing innovation across industries.
“This collaboration allows us to create extensive scenarios for customers globally, with embedded AI capabilities – from industry data solutions in Microsoft Fabric to AI agents in Microsoft Copilot Studio to AI models in Azure AI Studio – that enable industries to realise their full potential.”
He said the models will be available either through the Azure AI model catalogue, which offers a range of AI models to build custom AI applications in Azure AI Studio, or directly from Microsoft partners, adding that those available in the Azure AI model catalogue can also be used to configure agents in Copilot Studio, Microsoft’s platform for creating and deploying customised AI-powered agents.
These include ELY Crop Protection from healthcare and agriculture life sciences firm Bayer. According to Bayer, this specialised SLM is designed to enhance crop protection, sustainable use, application, compliance and knowledge in the agriculture sector.
Other big names include Siemens Digital Industries Software, which has developed a new copilot for its NX X software. Siemens said this uses an adapted AI model that enables users to ask natural language questions, access detailed technical insights, and streamline complex design tasks for faster and smarter product development. The copilot aims to provide CAD designers with AI-driven recommendations and best practices to optimise the design process in the NX X experience.
Read more stories on industry-specific AI
- Five industries where AI adoption is lagging: While AI captivates users worldwide, that doesn't directly translate into widespread adoption across all industries.
- Even AI laggards can see triple return on investment: A study from IDC shows that those businesses that have an aligned AI strategy can achieve accelerated benefits.
In the financial sector, regulatory compliance tech firm Safir, backed by Fidelity Investments’ innovation incubator, Fidelity Labs, is planning to introduce four new models in the Azure AI model catalogue.
Aimed at financial institutes, Safir said the retail marketing compliance model is focused on compliance in text, while the image detection model looks at compliance in images. There is also a risk interpretation model, which tries to explain why something was flagged, and a language suggestion model, which suggests alternative language that might be more compliant.
Rockwell Automation is offering the FT Optix Food & Beverage model, which it claims brings the benefits of industry-specific capabilities to frontline workers in manufacturing, supporting asset troubleshooting in the food and beverage domain. Rockwell Automation said the model provides recommendations, explanations and knowledge about specific manufacturing processes, machines and inputs to factory floor workers and engineers.
Other models include fine-tuned SLMs from in-car software developer Cerence, which run within a vehicle’s hardware and Factory Namespace Manager from SightManager, which analyses existing factory data, learns the patterns and rules behind the naming conventions and then automatically translates these data field names into a standardised corporate format.
Originally published at ECT News
