Google doubles down on agents
- Google framed AI agents as the centrepiece of its enterprise push at Cloud Next, pitching workflow automation over standalone models. - It announced a $750 million fund to help consulting firms deploy agentic AI and highlighted partnerships with Salesforce and Merck. - The moves signal Google is selling implementation capacity and integrations, not just models, as it showcased startups and cloud deals at the event ( )
Google used its Cloud Next conference in Las Vegas on April 22 to tell business customers it wants to sell AI agents that do work, not just models that answer prompts. (reuters.com) An AI agent is software that can take actions across apps and data systems, like updating records or moving a task through a workflow. Reuters reported Google made those agents the centerpiece of its enterprise pitch as Alphabet looks for clearer ways to turn artificial intelligence into revenue. (reuters.com) Google Cloud also announced a $750 million fund on April 22 for consulting firms, systems integrators, software partners and resellers building agentic AI on its platform. Google said the fund will support development, adoption and training across its 120,000-member partner ecosystem. (googlecloudpresscorner.com) Bloomberg reported the fund is aimed at firms including McKinsey, Accenture and Deloitte, with Google DeepMind offering some of them early access to Gemini models. That points to a sales strategy built around outside implementers that already run large corporate technology projects. (bloomberg.com) Google paired that funding push with integrations meant to make agents useful inside the software companies already buy. On April 22, Salesforce and Google Cloud said their expanded partnership will let AI agents execute end-to-end workflows across both platforms with shared context from customer records, documents and collaboration tools. (salesforce.com) The companies said the tie-up includes Google Workspace inside Salesforce’s Agentforce, Salesforce inside Google Workspace, and “zero-copy” links so data can be used across systems without being duplicated. The pitch is less about a standalone chatbot and more about connecting customer relationship software, email, files and meetings into one chain of actions. (salesforce.com) Merck announced the same day that it will expand work with Google Cloud in a multi-year deal valued at up to $1 billion. Merck said it plans to use Gemini Enterprise and related cloud tools across research and development, manufacturing, commercial operations and support functions. (merck.com) Reuters separately reported that Merck will invest as much as $1 billion with Google over several years to fund AI infrastructure and software. That gives Google a showcase customer in a regulated industry where companies usually move slowly on new enterprise systems. (reuters.com) The backdrop is a crowded market where Microsoft, Amazon, OpenAI, Anthropic and Salesforce are all pushing workplace AI. Reuters said Google used Cloud Next to argue it can win by combining its models with cloud infrastructure, security tools and partner-led deployment. (reuters.com) TechCrunch reported the same week that Google also deepened ties with Thinking Machines Lab in a new multibillion-dollar deal, another sign the company is spending heavily to secure both talent and distribution in AI. Cloud Next showed the customer-facing side of that push: Google is trying to be the company that wires agents into the systems businesses already run. (techcrunch.com)