Your employees already use AI. Public chatbots for client emails. Free tools for summarizing documents. A meeting-note bot someone installed without asking. You never approved any of it, and you have no written rules about it.

This is not a rare edge case. The SBE Council's Small Business Check Up (October 2026, 516 small employers) found 58% of small employers already use AI agents. MainNerve's 2026 Small Business Cybersecurity Survey found 87% AI adoption — but only 46% have any formal AI guidelines. Anagram's 2026 research found 58% of employees admitted pasting sensitive data — client records, internal documents — into large language models. Nationwide (September 2026) found 60% of owners say staff use public AI chatbots, yet only 36% have written policies — and 31% were targeted by generative-AI scams in the past year.

The gap between "we use AI" and "we have rules for AI" is where leaks, violations, and scams happen.

THE KIT — an employer-side governance system in 11-language Excel workbooks (EN, ZH, FR, ES, AR, RU, DE, JA, PT, HI, IT). No software, no subscription, no IT department required:

• Employee AI Acceptable-Use Policy template — 9 sections, fill-in-the-blank, plain language
• RED / AMBER / GREEN data classification — tier table plus 24-item classifier with formula counts. RED never goes into any AI tool.
• Approved-AI-tools register + 10-question shadow-AI discovery sweep
• 12-check AI-output verification checklist with formula verdict — the gate before anything goes client-facing
• One-page employee training + sign-off log with coverage formula
• Incident-response sheet — data-leak playbook and gen-AI scam playbook
• 30-day rollout planner with progress formula

Two rules run the whole kit: (1) no AI output goes client-facing or drives a decision until it passes the verification checklist; (2) no employee touches company or customer data with an AI tool until they have signed the training log.

WHAT'S INSIDE (25 files): 11 workbooks × 9 sheets each, QUICK-START guide, fictional WORKED-EXAMPLE (a 12-person agency finds 7 shadow AI tools and one data leak in week 1), FAQ (12), LIMITS (11 honest limits), INVENTORY, PRICING-NOTE, QA-REPORT, FREE SAMPLE workbook (English, pre-filled with the example), 2 real screenshots.

HONEST LIMITS: Does NOT give legal advice — review the policy against local employment/privacy law. Does NOT prevent incidents by itself — policy and process, not cybersecurity software. Does NOT substitute for cyber insurance or an IT security review. Templates require customization. Industry figures are the cited sources' figures, not ours.

$29 one-time, Commercial license. One company / one legal entity.