Last week I wrote about how SaaS may not be disappearing. The interface may simply be changing.
I think Salesforce + Anthropic announcement on August 26th about “Claudeforce” is a good example of what I meant.
Salesforce has been building Agentforce, where AI agents sit on top of Salesforce’s CRM and other platform capabilities and customers use those agents to get work done.
Now, with Claude, Salesforce is taking another approach. You can come to Claude and have Claude interact with Salesforce. So practically what we are seeing is that, we now have two models running in parallel.
Agentforce: User -> Agentforce -> Salesforce
Claude + Salesforce: User -> Claude -> Salesforce
The important part is that Salesforce still provides the underlying capabilities: Data, business logic, workflows, permissions and actions. What changes is where the user interacts with those capabilities. For decades, we opened the application, learned the UI and then worked through it. That may start changing.
And this is not just Salesforce. Microsoft, ServiceNow and SAP are moving in similar directions. So I don’t think the question is whether AI will kill SaaS. The more interesting question is:
“Who owns the interface to SaaS?”
The value is now very much shifting towards the core capabilities that enterprise software brings to the table. The UX, which has traditionally been a huge part of the customer experience and differentiation, is now being completely reimagined. We are building for a world where the same underlying functionality can be consumed by both a human and an AI agent.
That, to me, is one of the more interesting shifts happening with Agentic AI.
