The Joxel Group pushes governance-first AI adoption for behavioral health

The Joxel Group is advising behavioral health providers, county human services agencies, and healthcare organizations to put AI governance, risk controls, and clinician oversight in place before deploying AI tools. The Wisconsin firm says that approach is especially important for records covered by 42 CFR Part 2 and for organizations using AI in documentation, analytics, and prior authorization.

Why it matters: - Behavioral health organizations face higher privacy stakes when AI touches patient data, especially records covered by 42 CFR Part 2. - A governance-first model can help providers adopt AI without weakening confidentiality, clinical oversight, or compliance controls. - The approach is aimed at organizations that already have access to AI tools but have not built formal policies around them.

What happened: - The Joxel Group, a healthcare and public-sector consulting firm founded in 2006, said it is advising behavioral health providers, county human services agencies, and healthcare organizations on AI adoption. - The firm is promoting a governance-first model that requires policy, risk controls, and clinician oversight before any AI tool touches a patient record. - The company is based in West Allis, Wisconsin, and it was founded by Sushil and Nan Pillai.

The details: - Behavioral health organizations are evaluating AI documentation assistants, predictive analytics, and automated prior-authorization tools. - The Joxel Group says many organizations have not yet adopted a formal AI governance policy. - Substance use disorder treatment records fall under 42 CFR Part 2, which imposes stricter confidentiality and consent rules than HIPAA alone. - The firm’s governance work begins with an inventory of AI-enabled tools across the organization, including EHR analytics, third-party integrations, and staff-piloted tools. - The Joxel Group classifies tools by risk. - Low-risk tools do not interact with protected health information. - Moderate-risk tools process protected health information without producing clinical decisions. - High-risk tools influence clinical documentation, treatment recommendations, or risk stratification. - The resulting policies require Business Associate Agreements with AI vendors. - The policies bar vendors from training models on organizational protected health information without authorization. - The policies define how Part 2 records are handled. - Treating clinicians must review and edit AI-generated documentation before it enters the medical record. - AI risk scores remain in a decision-support role. - Ongoing oversight includes accuracy audits, bias audits, logging of AI interactions with protected health information, incident response tied to existing HIPAA procedures, annual recertification of each AI tool, and role-specific staff training. - For organizations using Netsmart myAvatar or myEvolv, the firm applies more than 10 years of myAvatar experience and seven years of myEvolv experience to configure audit logging, access restrictions, and signing and locking workflows. - The goal is to make sure clinicians review AI-assisted content before it becomes part of the record. - The consulting engagement model is built around the firm’s “Guiding Change. Delivering Value.” framework. - Consultants interview clinicians, operational leaders, finance teams, and executives before evaluating any vendor or tool. - Senior advisors lead the work directly. - The firm designs solutions around the staff who will use them. - Each project includes knowledge transfer so internal teams can sustain the results. - Every engagement begins with agreed measures of success. - The Joxel Group’s AI consulting services include Fractional Chief AI Officer services, Fractional CIO services, AI Strategy and Governance, Rapid AI Custom Development, and an AI Readiness Assessment for behavioral health organizations. - The AI Readiness Assessment scores AI maturity, governance readiness, and implementation risk. - The assessment also flags critical control gaps that a strong average score cannot offset.

Between the lines: - The message is less about AI experimentation and more about operational control. - The firm is positioning governance as the prerequisite for safe AI use in heavily regulated care settings. - By emphasizing transfer of capability to client staff, the consulting model appears designed to reduce long-term dependence on outside advisors.

What's next: - The Joxel Group says it will continue offering AI governance and implementation support through fractional leadership, strategy, custom development, and readiness assessments. - Behavioral health organizations evaluating AI tools will likely need to decide whether their current policies, audit trails, and clinician review processes are ready for deployment before expanding use. - Clients using myAvatar or myEvolv may add or tighten logging, access controls, and signing workflows as part of implementation.

The bottom line: - The Joxel Group is betting that AI adoption in behavioral health will succeed only if governance comes first, not after deployment.