Governance that fits a 50-person business
You do not need a committee. You need a policy people read, an inventory that is true, a risk register with owners, and an evidence file. The quarterly cycle keeps them current. Everything maps to the frameworks below, so when someone asks, you can show it.

The quarterly governance cycle
The four documents
Acceptable-use policy
What may be entered, shared, automated and published with AI, by data type and role. Two pages, signed annually.
AI inventory
Every AI app, copilot, agent and integration: owner, plan, data types, decision, review date.
Risk register
Each finding rated by likelihood and impact, with the control, the owner and the due date. Mapped to NIST AI RMF functions.
Evidence file
Vendor contracts, DLP reports, consent reviews, training records and the audit report, ready for insurers and clients.
Framework and regulation map
| Framework or rule | What it asks for | Who it reaches | Audit output that answers it |
|---|---|---|---|
| NIST AI Risk Management Framework | Govern, map, measure and manage AI risk; Generative AI Profile adds confabulation, privacy, security | Voluntary, widely referenced by insurers and clients | Risk register mapped to the four functions |
| ISO/IEC 42001 | An AI management system: policy, roles, risk assessment, lifecycle controls | Organizations seeking certification or asked by enterprise clients | Policy, inventory, risk register, review cycle |
| HIPAA Security Rule | Risk analysis covering AI tools that touch ePHI; BAAs with AI vendors | Covered entities and business associates | ePHI data-flow map, vendor BAA list, DLP evidence |
| PCI DSS | No cardholder data in unapproved systems; scope control | Any business handling card data | DLP rules and scope diagram |
| California Civil Rights Council ADS rules (effective Oct 1, 2025) | No discriminatory automated-decision systems in employment; keep records four years | California employers with 5 or more employees | Inventory flag for employment-decision AI, record-keeping note |
| California CPPA ADMT regulations (effective Jan 1, 2026) | Risk assessments, notices and opt-outs for automated decisions about people | Larger businesses under the CCPA | Inventory flag and referral to counsel |
| Cyber insurance applications | Questions about AI use, data handling and controls are appearing on renewals | Every insured business | Evidence file and executive report |

Why this matters in 2026
A page from the sample report
Every finding in the real report has this shape: domain, finding, severity, evidence, fix. See the full sample report.
Pairs with HIPAA Security Readiness and vCISO
Healthcare practices get the AI findings folded into the HIPAA risk analysis. Businesses without a security lead get the quarterly cycle run by the vCISO service.
HIPAA security readiness
Guided by Ali Hassani, CISO
Certified CISO (C|CISO) and CISSP with 25+ years in IT, cybersecurity, compliance and infrastructure, and dozens of Microsoft and Cisco certifications. Every AI Security Audit is read-only, evidence-based, and written so an owner and an administrator can both act on it.
Questions business leaders ask
Do the California rules apply to a 30-person company?
The Civil Rights Council rules apply to employers with five or more employees that use automated-decision systems in employment. The CPPA rules have revenue and data thresholds. The audit flags any AI used in hiring or employment decisions and recommends you confirm with employment counsel; we are not a law firm.
Is NIST AI RMF mandatory?
No. It is voluntary, but it is the framework insurers and enterprise clients most often reference, and it maps cleanly to the audit's domains.
How long does governance take to set up?
The audit delivers the four documents in first-draft form. Most businesses adopt them within 30 days and run the first review at 90.
Implementation partner
OC Security Audit stays independent and read-only. If you want hands-on administration, IT Perfection can implement the fixes: DLP rules, Conditional Access, app blocking, agent identities, and ongoing Microsoft 365 support.
Need to show an insurer, client or regulator what your AI controls are?
The AI Security Audit produces the policy, inventory, risk register and evidence file, mapped to the frameworks that apply to you.
Read-only audit. No changes to your systems. Irvine, Orange County and Southern California on-site; remote nationwide.