Chiri blog
Ideas, research, and practical guides.
1–12 of 88 articles
What to Put in an AI Acceptable Use Policy for Employees
What an AI acceptable use policy for employees should include, which data never goes into public AI tools, and how to get people to follow it.
AI Agent Examples by Business Function
Real AI agent examples for sales, support, finance, HR, IT and security, drawn from 81 agent products in Chiri Atlas, with survey data on where agents are used.
What AI Governance Consulting Should Deliver
What AI governance consulting covers, what a useful engagement should hand over, which frameworks it maps to, and what to ask a partner.
The AI Governance Maturity Model
Five levels of AI governance maturity, a ten-question self-check, and what it takes to move up one level without a governance office.
AI in HR: What the Tools Actually Do
How HR teams use AI, what 49 HR tools in Chiri Atlas focus on, and the rules that apply when AI helps decide who gets hired.
AI Observability: What to Monitor for AI Agents
What AI observability means for agents in production: the traces, costs, quality scores and safety events to record, and who should watch them.
Chief AI Officer: Does Your Company Need One?
What a chief AI officer does, how common the role has become, and how a 50 to 500 person company can cover the job without a new executive hire.
Human in the Loop AI: What an Approval Must Show and Log
When a person approves an AI agent's action, what the approval screen must show, what must be enforced outside the model, and what the log should record.
Prompt Injection in the Data AI Agents Read
How prompt injection reaches AI agents through tickets, emails and logs, which controls limit the damage, and why no single filter is enough.
Single Agent or Multi-Agent System?
A multi-agent system splits work across AI agents with separate roles. When it beats a single agent, what it costs, and the checks to run before building one.
Palantir Alternative: How to Evaluate Enterprise AI Platforms
Evaluate a Palantir alternative through workflow fit, governance, deployment, portability, ownership costs, and a measurable proof of concept.
What Are Agentic Workflows?
An agentic workflow lets an AI agent decide some steps of a process while fixed rules run the rest. How it differs from automation and how to design one.