Stibo Systems provides customers with the ability to derive maximal value from AI by ensuring both high-quality mastered data input and high-quality AI-generated output through a variety of AI-powered features and functionalities. Stibo Systems' AI offerings include features that support user-level AI-assisted content generation, translation, and image analysis, among others, and also enables customers to connect STEP to Large Language Models (LLMs) of their choosing. These features can help customers automate routine data-management tasks, improve data quality, and compress time-to-market, among other goals.
Stibo Systems' AI functionality is accessible to users via the ProductGen AI solution.
ProductGen AI (recommended)
The ProductGen AI solution brings together multiple AI-capabilities into an accessible and easy-to-use offering. This solution is designed to address many common product data management use cases. With ProductGen AI, users can:
-
Generate product descriptions and marketing copy tailored to specific audiences.
-
Analyze product imagery to extract text from images, generate image descriptions for SEO, and ensure images accurately represent the products they visualize.
-
Translate attribute values into multiple target languages, accelerating expansion into new markets.
-
Validate data consistency across related attributes and flag likely outliers before publication.
Because ProductGen AI is designed for rapid enablement with out-of-the-box functionality, customers can quickly derive value without a lengthy professional-services engagement.
Integrations
For customers whose requirements extend beyond ProductGen AI's curated experience, STEP supports direct connections to Azure AI services. These are the same technical components ProductGen AI itself relies on, exposed to advanced users who need to assemble customized behavior. Two integration targets are currently supported:
-
Azure OpenAI Integration — Connect STEP to Azure OpenAI for text generation, translation, and content creation. For more information on this option, refer to the Azure OpenAI Integration topic.
-
Azure Vision Integration — Connect STEP to Azure Computer Vision for image analysis and metadata extraction. For more information on this option, refer to the Azure Vision Integration topic.
These integrations are intended for advanced implementers who need custom behavior beyond ProductGen AI, and for customers who must use their own Azure subscriptions to satisfy contractual or regional requirements. For configuration detail, refer to the Azure OpenAI Integration and Azure Vision Integration topics.
Machine Learning in Matching
Beyond generative AI, STEP also applies machine learning to improve the quality and efficiency of data matching. Two capabilities are available:
-
Machine Learning Matcher — A pre-trained model that matches individual party data elements, providing a strong baseline for entity (person), organization, and address matching without necessitating a model be trained from the ground up. For additional information on this feature, refer to the Matcher: Machine Learning Matcher topic.
-
Machine Learning Match Recommendations — A learning feature that observes user decisions during clerical review and presents merge or reject recommendations that reflect your organization's own matching conventions. For more information on this feature, refer to the Machine Learning Match Recommendations topic.
Together, these capabilities let teams start with sensible defaults and refine behavior based on human-made decisions, which is often more effective than requiring every matching rule be specified in advance.
Getting Help
Choosing the right AI approach depends on factors such as data volume, governance requirements, and existing Azure investments. For guidance on which approach fits your scenario, contact your Customer Success Manager or Account Manager.