Microsoft’s expansion of AI agents within the Copilot Studio ecosystem was a central focus of the company’s Ignite conference. Since the launch of Copilot Studio, more than 100,000 enterprise organizations have created or edited AI agents using the platform. Copilot Studio is getting new features to increase productivity, including multimodal capabilities that take agents beyond text and Retrieval Augmented Generation (RAG) enhancements to enable agents with real-time knowledge from multiple third-party sources, such as Salesforce, ServiceNow, and Zendesk. Integration with Azure is expanded as 1,800 large language models in the Azure catalog are made available. Continue reading Microsoft Pushes Copilot Studio Agents, Adds Azure Models

The theme of Zoomtopia 2024 was “an AI-first work platform for human connection,” with the release of the custom Zoom AI Companion 2.0 for Zoom Workplace, an AI assistant that costs $12 per month that starting next year will have the ability to create custom AI avatars. Initially available for short video presentations to share with internal teams, the eventual goal is to have a “digital twin” that can participate in meetings and calls. The avatars will “save time and production costs” with segments that work with Zoom Clips for internal circulation in brief messages to colleagues. Continue reading Zoom Updates Its AI Assistant and Previews Custom Avatars

Facebook disclosed plans to buy Kustomer, a customer relationship management startup, in a deal valued by sources at close to $1 billion. Crunchbase stated the New York-based Kustomer has raised about $170 million in venture capital funding. Kustomer would help Facebook provide customer support on its main platform and its services such as WhatsApp, Instagram and Facebook Messenger. The deal, subject to regulatory scrutiny, comes as Facebook is already being investigated for its acquisition of startups as anticompetitive behavior. Continue reading Facebook Plans to Buy Customer Service Startup Kustomer

Zendesk is introducing Automatic Answers, a service that answers a business’s emails from customers, without human intervention. But it’s far more sophisticated than today’s well-known auto-responders, having been built on a machine-learning platform that in-house data scientists and engineers have been developing for some time. The platform is also used for another Zendesk service, Satisfaction Prediction, which helps determine the customer experience. Automatic Answers will be rolled out later in the year. Continue reading Zendesk Customer Service Product Built on Machine Learning