
The agent era doesn’t destroy all competitive advantages equally. Two moats die. Three survive and compound.
Moats That Die
Per-Seat Pricing
The revenue engine breaks when AI agents replace humans as the primary software consumer. 10 agents doing the work of 100 humans = 90% fewer seats.
Casualties: Salesforce -26%, ServiceNow -54%, Atlassian -35%, Workday, Paycom
UI / Trained-User Moat
The trained-user-interface moat collapses when the interface becomes natural language. Nobody needs to learn your UI when they can just tell an agent what to do.
Casualties: Intuit -34%, Zoom, DocuSign, HubSpot -51%, Dropbox, UiPath
Moats That Survive
Data Gravity
Data gravity is the ultimate moat. AI agents need clean, structured, accessible data to function. The companies that hold the data become more valuable, not less.
Winners: Palantir +70% YoY, Oracle, SAP, Snowflake, Wolters Kluwer
Regulatory Requirement
100% accuracy required. Governments will not accept “approximately compliant.” ERP, payroll, credit scoring, and defense systems demand deterministic outputs.
Winners: SAP, Oracle, ADP, FICO, Leidos, Veeva Systems
Network Effects
Self-reinforcing lock-in. Every new node makes the network stickier. Every agent that connects through the protocol makes the protocol more essential.
Winners: CrowdStrike, Palo Alto, Okta, Cloudflare, Datadog
The Bottom Line
Two moats die. Three moats compound. The per-seat model and the UI moat built a $2 trillion market. Both are now liabilities. The companies that survive — and expand — are those built on data gravity, regulatory requirements, and network effects. These moats don’t just survive the agent era. They get stronger.
Read the full analysis on The Business Engineer









