Seven rules that hold whatever tool you pick, ranked by how much they matter. Each links to the best free source behind it.

• 1. Start with a business problem, not a tool
  Pick the task that costs you the most time or money, then look for AI that fixes it. Tools chosen first rarely stick.
  Source: The 4 Stages of AI Adoption, David Schonthal, Kellogg Insight · Mar 2026

• 2. Keep confidential data out of AI tools
  Do not paste customer or employee secrets into AI tools, and switch off training on your inputs. A leak cannot be undone.
  Source: AI tips, use cases and limitations, SCORE

• 3. Pilot one task and measure before and after
  Time the task before AI and after. People feel faster with AI even when they are not, so only the numbers count.
  Source: Measuring the impact of early-2025 AI on experienced developers, METR · Jul 2025

• 4. A person checks every output that reaches a customer
  AI makes confident mistakes. Anything sent to a customer, an employee or a regulator needs a human check first.
  Source: AI Risk Management Framework, NIST

• 5. Be honest about AI
  Do not claim AI powers you when it does not, and never let AI write reviews or testimonials. The FTC fines both.
  Source: FTC final rule banning fake reviews and testimonials · Aug 2024

• 6. Bring your team along
  Your staff find the best uses for AI in their own work. Give them permission, time and a few rules.
  Source: Making AI Work: Leadership, Lab, and Crowd, Ethan Mollick · May 2025

• 7. Read every AI number with its definition
  Reports say anywhere from 9% to 76% of businesses use AI, depending on what they count. Check what was asked before you compare.
  Source: AI use among businesses, US Census Bureau · May 2026

Curated by effectiveAI. Every link checked November 2026.