Claudeforce and What Comes Next for Agentic AI
InformationWeek, Tuesday, October 27th, 2026
Crescendo CEO Andy Lee argues Salesforce opening up to Claude signals agents running apps, and CIOs should rethink their stacks around that.
Andy Lee says framing Salesforce's work with Anthropic as AI killing SaaS is the wrong debate, since systems of record may become more valuable as the trusted data and logic layer beneath agents.
Giving agents tools does not teach them a business function, so enterprises must define skills, connect context, set guardrails and orchestrate agents.
Using customer experience as an example, he urges CIOs not to replace six applications with six agents but to design a unified agentic system with a continuous learning loop.
CIOs should decide which systems stay essential, which layers agents can absorb, and where rebuilding a function creates real operational change.