Meta, Microsoft, and Amazon Lead the Charge in Agentic AI Development with Llama 4, Avatars, and Cloud Innovations

Meta's chief product officer, Chris Cox, announced that the company is developing Llama 4, an open-source AI software set to empower AI agents with advanced reasoning abilities, allowing them to interact with web browsers and various tools to perform multistep tasks. This was revealed at Morgan Stanley's Technology, Media and Telecom Conference, where Cox emphasized the potential of AI agents in automating business tasks such as filing receipts. Clara Shih, Meta's head of business AI, highlighted that the company plans to leverage its relationships with 200 million small businesses to implement AI agents to automate workflows, provide customer service, and enhance customer interaction globally. Meta is also preparing to introduce a stand-alone Meta AI app in the second quarter and will host its inaugural LlamaCon AI conference on April 29.

Salesforce has launched Agentforce 2dx, the third iteration of its agentic AI platform, designed to streamline the creation and management of AI agents that automate various business operations, including sales, service, marketing, finance, HR, and operations. This developer-focused update introduces features like an API, DX Inspector, Interaction Explorer, and support for VS Code. The enhancements mainly focus on orchestrating and building AI agents to improve efficiency across departments.

Startup D-ID has partnered with Microsoft to integrate its agentic AI avatars into Microsoft's software, such as Microsoft Teams, enhancing customer interactions with engaging, human-like interfaces. Known for lifelike avatars offering real-time video streaming and accurate lip-syncing, D-ID aims to revolutionize digital interfaces by replacing traditional graphical interfaces with human-like interactions. Already serving major clients like Fidelity and PepsiCo, D-ID's avatars are used across marketing, sales, and customer service. Microsoft is integrating AI capabilities into its 365 suite under the "Copilot" brand, significantly boosting performance for Microsoft's sales and marketing teams. This collaboration highlights Microsoft's strategic focus on generative AI as both companies aim to capture a share of the projected trillion-dollar AI market by 2033, emphasizing responsible AI and security.

Swami Sivasubramanian, a veteran executive at Amazon, will lead a new initiative within AWS focused on developing agentic AI tools, which he claims will bring unprecedented efficiency by automating complex workflows with human-like reasoning. This strategic move aligns with Amazon's efforts to stay competitive in the rapidly evolving generative AI landscape, already leveraged by services like AWS's Q to modernize internal applications. AWS CEO Matt Garman identified agentic AI as a potential multibillion-dollar opportunity, and the initiative will involve several senior executives. This development follows similar moves by tech rivals, such as Microsoft's Azure AI Agent Service and Google's Agentspace, as Amazon maintains its leadership in cloud infrastructure with significant revenue gains and aims to enhance its Alexa assistant with AI models from partners like Anthropic.

Criminals are increasingly using AI to create voice clones of individuals, aiming to defraud people by impersonating their loved ones, bosses, real estate agents, and more, as highlighted by a McAfee survey showing that a quarter of adults in seven countries have suffered such scams, with 77% losing money. Scammers typically extract voice samples from social media, needing only a few seconds of audio to create convincing replicas using AI tools. Neal O'Farrell, founder of Think Security First!, warns that criminals may use cloned voices to engage in brief, urgent calls to avoid suspicion. Common scams include fake grandparent calls, dubious demands for money transfers from impersonated bosses, and fraudulent real estate and financial advisor calls. To combat these scams, experts advise recognizing telltale signs such as unfamiliar numbers, strange payment requests, and discrepancies in familiar voices, urging people to report incidents and consider identity theft protection services like Aura.