Google expands Gemini capabilities
Google's Gemini model has been integrated across Google Workspace, with the company claiming it saves users an average of 105 minutes per week. The Gemini Pro model is now also available on the Vertex AI platform for enterprise use. A new feature for the Gemini 3 Deep Think model can now turn sketches into 3D-printable models, signaling advances in multimodal AI.
- The Gemini 3 Deep Think model's new sketch-to-3D feature analyzes a drawing's geometry to generate a printable 3D model, a tool aimed at accelerating prototyping for engineers and designers. This advanced reasoning capability is now available to Google AI Ultra subscribers through the Gemini app and to select researchers and enterprises via an API. - For enterprise use, Gemini can be accessed through Vertex AI, a platform designed for building and scaling machine learning applications with enterprise-grade security and governance. This is distinct from the Gemini API offered via Google AI Studio, which is geared more towards prototyping and experimentation. Vertex AI allows for the integration of Gemini models with other Google Cloud services like BigQuery and supports building custom AI agents. - In the insurance sector, AI platforms are being used to automate up to 70% of underwriting tasks by analyzing vast datasets to assess risk, detect fraud, and personalize pricing. For example, Hong Kong-based MediConCen uses Gemini on Vertex AI to process medical insurance claims, achieving 98% data extraction accuracy and reducing processing time from days to hours. - MLOps best practices for enterprise-scale AI include versioning everything (code, data, models), automating CI/CD pipelines for testing and deployment, and continuous monitoring of models in production for data drift. Tools like MLflow and cloud-native platforms such as Vertex AI are recommended for managing the machine learning lifecycle. - In the fashion and retail industries, AI is being used for hyper-personalization, with recommendation engines analyzing browsing history and purchase behavior to tailor suggestions. Companies like Nike and Stitch Fix use these technologies to customize products and curate clothing selections. AI is also optimizing inventory management by forecasting demand, which can reduce overproduction. - Google's broader AI strategy in 2025 involved embedding a portfolio of Gemini models (including Flash and Nano for efficiency) across its products, from Search to Android, rather than focusing on a single model. This ecosystem approach aims to make AI a fundamental layer of the user experience.
