Anthropic repackages Claude as workflow products with enterprise connectors

Anthropic is pushing Claude further from a standalone chatbot toward packaged workflow software, with a new small-business product, an expanding connectors stack and signs that its security-focused Mythos research may be moving into developer tools. The clearest official step came on May 13, when the company introduced Claude for Small Business as a bundle of connectors and ready-to-run workflows inside existing software. Anthropic’s own platform pages also show a broader product lineup around Claude Code, Claude Security, Claude Cowork and connectors, alongside managed-agent tooling built to reach external systems. That combination matters because Anthropic is no longer presenting Claude only as a family of base models. It is also selling packaged ways to use those models inside finance, operations, sales and security work. Reports on May 23 and May 24 said users briefly saw “Mythos 1” and “claude-mythos-1-preview” references inside Claude Code and Claude Security, though those sightings came from third-party reporting rather than an Anthropic launch post. (anthropic.com) ### What did Anthropic actually launch? Anthropic on May 13 said Claude for Small Business is “a package of connectors and ready-to-run workflows” designed to put Claude inside the tools small businesses already use. The company said the package works with Intuit QuickBooks, PayPal, HubSpot, Canva, Docusign, Google Workspace and Microsoft 365. The company said the product includes 15 ready-to-run workflows across finance, operations, sales, marketing, HR and customer service, plus 15 “skills” built around recurring tasks. (anthropic.com) Anthropic said those tasks include planning payroll, closing the month, chasing invoices and running sales or marketing work, with users approving actions before anything is sent, posted or paid. ### Why do connectors matter more than another model announcement? (anthropic.com) Anthropic’s developer documentation says its Managed Agents product can connect to external tools and data sources through Model Context Protocol, or MCP, using remote servers and permissions controls. The docs say agents can be configured with up to 20 MCP servers and that tool access can require user approval. (anthropic.com) Anthropic’s platform pages also list “Connectors,” “Marketplace,” “Plugins” and several packaged surfaces including Claude for Slack, Claude for Excel, Claude for PowerPoint and Claude for Word. Taken together, those pages show Anthropic building distribution through software entry points and integrations, not only through direct model access. That is an inference from Anthropic’s product navigation and documentation. (platform.claude.com) ### Where does Mythos fit into this? Anthropic’s April 7 security post introduced Claude Mythos Preview as a general-purpose model that the company said was unusually capable at cybersecurity tasks. Anthropic said Mythos Preview had identified and exploited zero-day vulnerabilities in major operating systems and web browsers during testing, and said more than 99% of the vulnerabilities it found had not yet been patched. (platform.claude.com) TestingCatalog reported on May 23 that “Mythos 1” and a preview-labeled variant appeared to be preparing for Claude Code and Claude Security, citing temporary UI visibility and source-code strings. India Today and Times Now reported similar sightings on May 24, saying the traces later disappeared and Anthropic was still publicly tying broader release to stronger safeguards. Those reports have not been matched by a formal Anthropic product announcement. (red.anthropic.com) ### Is Anthropic now selling workflows instead of just models? Anthropic’s public materials suggest it is selling both. The company still maintains model and API documentation, but its May 13 launch and current platform pages emphasize product surfaces, skills, connectors and workflow-specific tools. Daniela Amodei, Anthropic’s co-founder and president, said in the launch post that Claude for Small Business is meant to run “inside the tools owners already rely on” and handle tasks such as payroll planning, invoice follow-up and marketing setup. (testingcatalog.com) That language places the product in everyday software workflows rather than in benchmark comparisons alone. ### What should readers watch next? (platform.claude.com) Anthropic’s next visible milestones are likely to show up on its product pages, developer docs and news posts, where the company has already published the May 13 small-business launch and the April 7 Mythos security write-up. Separate third-party reports on May 23-24 indicate any broader Mythos move would likely surface first in Claude Code, Claude Security or enterprise-facing dashboards. (anthropic.com)