Skip to content
Business ROI July 23, 2026 · 7 min read

Authorization Before Readiness: Why Inaccessible Onboarding Is a Governance Failure, and What It Predicts About AI Deployment

Heather M. Grizzle, M.A.
Heather M. Grizzle, M.A. Co-Founder & Principal Consultant
Audio Insights
Also listen on Spotify

Heather M. Grizzle, Novara Consulting Group LLC

Institutions routinely authorize participation before they have built the conditions that make participation possible. A hiring decision is finalized, a start date is set, and the employee arrives to find that the materials, systems, and accommodations required for the role were never prepared. The organization has authorized someone to participate in work it has not yet made participable. This is not an administrative lapse. It is a governance failure, and it is the same failure that occurs when an organization authorizes an artificial intelligence system for deployment before establishing the oversight, testing, and accountability structures that would make deployment defensible.

Onboarding is the useful case because the failure is visible on a short timeline and the costs are easy to trace. Consider a sequence that recurs across sectors. A department hires a disabled employee. Training materials have not been produced in accessible formats. What was scheduled as a structured two-week orientation extends past a month. Instruction is delivered inconsistently because the department is designing access provisions after arrival rather than before it. The employee spends the opening weeks of the role identifying missing materials, explaining why standard processes do not function, and constructing an access plan that the institution should have supplied. Productivity is delayed, and the delay is recorded against the individual rather than the process that produced it.

The significance of this pattern is not that it is inconvenient. It is that it constitutes structural exclusion within a formally inclusive act. The organization has hired the employee. It has satisfied the visible obligation. Yet the employee remains functionally excluded because the institution treated hiring as the endpoint of the accessibility question rather than the point at which the question becomes operative. Formal admission without operational readiness produces exclusion that no one has to intend and that no policy document will record.

The mechanism deserves precise naming, because how a failure is attributed determines who is expected to fix it. Deborah Stone’s work on causal stories describes how competing accounts of what caused a harm determine where responsibility is assigned and which remedies become thinkable. An onboarding delay attributed to individual circumstance generates accommodation requests, case-by-case exceptions, and a quiet performance record. The same delay attributed to institutional unreadiness generates procurement requirements, design standards, and sign-off procedures. The facts are identical. The governance consequences are not.

The Same Failure, Applied to Technology

Organizations now deploy automated and AI-enabled systems across the functions where this failure originates: recruitment screening, onboarding platforms, employee training delivery, performance monitoring, scheduling, and internal decision support. An AI-enabled process is not thereby an accessible or well-governed one. Where a system is procured and deployed without testing whether disabled employees can use it effectively, the technology does not correct the underlying pattern. It industrializes it, applying an untested access assumption uniformly and at speed.

The relevant question at the point of authorization is not whether a system functions according to its technical specification. A platform may perform exactly as designed and still fail as an institutional instrument if the people required to use it cannot access, navigate, or meaningfully engage with it. Vendor documentation attesting to conformance is a claim about the product. It is not evidence about the workforce. Public evidence in this market does not yet support treating the two as equivalent, and procurement instruments that conflate them transfer risk onto the institution while appearing to reduce it.

The legal architecture already anticipates this. The Americans with Disabilities Act frames accommodation as an interactive obligation of the employer rather than a request the employee must originate and justify. Section 508 of the Rehabilitation Act imposes accessibility requirements on information and communication technology procured by federal agencies, and its influence extends well beyond direct federal purchasers through contracting relationships and adopted state standards. In Europe, Directive (EU) 2019/882 has been enforceable since June 2025 and reaches organizations serving European consumers irrespective of where they are registered, with EN 301 549 operating as the technical benchmark. Employers with any European exposure are now operating under an accessibility regime with procurement consequences attached, including potential exclusion from public contracts.

What the legal architecture does not supply is a method. Statutory duties establish that accessibility is required. They do not tell a procurement officer what evidence to demand before signing, what a sufficient test looks like, or what happens when a system is found deficient after deployment. That gap is where governance instruments belong, and where most organizations currently have nothing at all.

Pre-Authorization, Not Post-Complaint

Governance that begins after a complaint is not governance. It is remediation with a paper trail. The corrective is to locate the accessibility determination at the authorization point, where the institution still has leverage over the vendor and has not yet made the operational commitments that make reversal expensive.

At minimum, an organization should be able to answer six questions before an AI-supported employment system is approved for use. What accessibility requirements were written into the solicitation rather than raised afterward? What testing was conducted with affected users rather than with proxies or automated checkers alone? What communication access provisions are documented, and who is responsible for maintaining them as the system is updated? What human support procedure operates alongside the system, and is it resourced or merely named? What is the defined failure procedure when the technology does not work for a given employee, and does it place the burden of invoking it on that employee? And who specifically signs the authorization, on what evidentiary basis, with what standing to refuse?

None of these questions is technically demanding. Their absence from most procurement files is the point. Lester Salamon’s account of governance describes public action as a matter of tool selection, where the instrument chosen carries distributive consequences independent of the goal it was chosen to serve. Procurement is such a tool. Selecting an automated onboarding platform is not a neutral operational decision. It allocates who bears the cost of the system’s limitations, and in the absence of pre-authorization requirements, that cost falls consistently on the employees least positioned to absorb it.

Readiness as the Standard

The reframing this article proposes is narrow but consequential. Accessible onboarding is not a service courtesy extended to disabled employees. It is an observable indicator of whether an institution authorizes commitments it is prepared to honor. Organizations that prepare access provisions before an employee’s first day have demonstrated that they anticipate foreseeable requirements, account for users unlike the designers, and accept responsibility for equitable implementation. Organizations that respond only after a disabled employee encounters a barrier have not met an unexpected problem. They have disclosed that readiness was never part of the authorization criteria.

The same standard transfers directly to technology governance. Accessibility, oversight, and accountability must be established before authorization, not retrofitted after delay, harm, or exclusion has occurred. An institution that cannot answer the six questions above about a system it has already deployed has not made a governance decision about that system. It has made a purchasing decision and called it governance.

Inclusion cannot begin after arrival. Neither can oversight.

*Novara Consulting Group LLC advises public and institutional buyers on governance and procurement of AI systems affecting accessibility and workforce participation.*

References

Americans with Disabilities Act of 1990, 42 U.S.C. § 12101 et seq.

Bauer, T. N. (2010). *Onboarding new employees: Maximizing success*. SHRM Foundation Effective Practice Guidelines Series.

Dickson, R. K., & Isisah, O. S. (2024). An exploration of effective onboarding on employee engagement and retention in work organizations. *Researchjournali’s Journal of Human Resource, 9*(1), 1–21.

Directive (EU) 2019/882 of the European Parliament and of the Council of 17 April 2019 on the accessibility requirements for products and services (European Accessibility Act). *Official Journal of the European Union*, L 151.

ETSI. (2021). *EN 301 549: Accessibility requirements for ICT products and services* (V3.2.1).

National Institute of Standards and Technology. (2023). *Artificial intelligence risk management framework (AI RMF 1.0)* (NIST AI 100-1). U.S. Department of Commerce.

Rehabilitation Act of 1973, Section 508, 29 U.S.C. § 794d.

Salamon, L. M. (Ed.). (2002). *The tools of government: A guide to the new governance*. Oxford University Press.

Stone, D. (2012). *Policy paradox: The art of political decision making* (3rd ed.). W. W. Norton.

U.S. Equal Employment Opportunity Commission. (2022). *The Americans with Disabilities Act and the use of software, algorithms, and artificial intelligence to assess job applicants and employees* (EEOC-NVTA-2022-2).

World Wide Web Consortium. (2023). *Web Content Accessibility Guidelines (WCAG) 2.2*. W3C Recommendation.

Cite this
Grizzle, H. M. (2026, July 23). Authorization Before Readiness: Why Inaccessible Onboarding Is a Governance Failure, and What It Predicts About AI Deployment. Novara Consulting Group. https://www.novaracg.com/2026/07/23/authorization-before-readiness-why-inaccessible-onboarding-is-a-governance-failure-and-what-it-predicts-about-ai-deployment/
Heather M. Grizzle, M.A.
Heather M. Grizzle, M.A. Co-Founder & Principal Consultant

Content Partner and Client Success Lead, ensuring technical insights are communicated clearly, consistently, and in ways clients can confidently act upon.

Work with us

Turning this analysis into practice?

Novara advises agencies and organizations on AI governance, responsible-AI frameworks, and accessible procurement — including SLAT-based assessments of language-access systems.

Request a consultation →

Subscribe to Novara Consulting Group

Analysis on sign language AI, procurement, and Deaf-led governance — delivered to your inbox.

Consult