From individual coding tools to team development platforms, a shift in how agents earn trust
AITessl Blog author, an AWS Senior Principal PM who builds Amazon's Kiro agentic coding tool, argues that organizations trust AI coding agents in proportion to how cheaply they can verify agent output, not to model capability. He says agent costs scale with how much autonomous work the agent does, unlike the flat per-seat pricing of autocomplete tools, and that roughly 80% of developers use about 20% of the budget.
Source details
Tessl Blog · Massimo Re Ferre' · tessl.ioPublished · added here
Original title: From individual coding tools to development platforms