Enterprise AI Coding Costs

Enterprise AI Agent Deployment and Outcomes

Research how enterprises extend AI agents from pilots into knowledge-worker daily work: actual deployment patterns, measurement scope, governance frameworks, and the evidence required before scaling autonomy.

Lead research

Industry mechanism

Enterprise AI Agent Deployment: Which of BBVA’s Seven Metrics Actually Measure Agent Outcomes

A real enterprise IT question asks what knowledge workers actually do with more agentic AI. BBVA’s October 2026 presentation and deployment announcement, read alongside its March audit-assistant case, supply concrete workflows and seven different indicators. Their populations, units and measurement status differ. They offer useful pilot ideas, but do not isolate the additional benefit of autonomous planning over an existing internal chatbot.

Read the research

Research in this collection

Standalone research

What do “24/7 agents” do? Three configurations and one public run

A developer asks what always-on agents actually do. Two repository-report configurations and an event-triggered cost tracker make the work concrete: collect context, request a report and constrain how it is published. A public Team Status run shows execution stages and artifacts. These records help distinguish scheduled availability, model activity, permitted actions and useful output; none can substitute for the others.

Read the full analysis