Evidence before adoption. Trust before deployment.
Sign language AI is being promoted faster than its evidence base has matured. Novara Consulting Group provides structured, criterion-based review of sign language AI systems, before adoption, procurement, pilot use, or public deployment, using the SLAT Index (Sign Language Access Trust), so procurement decisions rest on evidence rather than marketing.
Procurement is outrunning the evidence
Organizations are being asked to evaluate sign language AI without clear standards, consistent evidence disclosures, or Deaf-led governance. There is no established public benchmark, no independent certification regime, and no shared vocabulary for distinguishing a validated system from a well-marketed one.
When a sign language AI system fails, the failure lands on Deaf and hard-of-hearing people first, and on the deploying institution second, as access complaints, civil-rights exposure, procurement scrutiny, and public accountability problems. The central question is not simply “Does the technology work?”
Is this system appropriate, evidence-supported, accountable, and safe for the context in which it will be used?
That is a governance question, and it requires a governance answer. SLAT reviews exist to provide one.
Five ways to bring us in
Five review types under the SLAT Index, each a structured governance, procurement, and risk assessment grounded in public evidence and documented against defined criteria.
SLAT Vendor Snapshot
A concise, public-evidence review of a sign language AI vendor, product, or platform, public claims, stated use cases, governance disclosures, evidence gaps, deployment risks, and procurement concerns. A fast, credible orientation before deeper engagement.
SLAT Procurement Risk Review
A deeper review for adoption, pilot use, RFP language, vendor selection, or contract evaluation. Maps vendor evidence to your intended use case and produces procurement-usable findings: risk issues, vendor questions, contract safeguards, and evaluation criteria.
SLAT Governance Review
A policy and accountability review of the structures around a system: oversight, human-review pathways, Deaf and Disabled governance, risk escalation, transparency, and deployment safeguards. For building internal AI accessibility policy or evaluating a vendor.
SLAT Evidence & Claims Review
A focused review of vendor claims and the evidence behind them, validation statements, accuracy claims, dataset disclosures, testing methods, published studies, distinguishing what has been demonstrated, what has been asserted, and what remains publicly unverifiable.
SLAT Deployment Context Review
A context-specific review for high-impact settings, healthcare, emergency response, education, courts, employment, transportation, public benefits. The same system carries different risk in different contexts; this review asks whether the proposed context is appropriate for the current evidence.
The ten SLAT domains
Linguistic Integrity
Does the system handle sign language as a full language, grammar carried by facial expression, movement, and spatial structure, or treat signing as word-for-word translation?
Deaf & Disabled Governance
Are Deaf and disabled people in decision-making roles, not just testing or advisory ones? Who holds authority over how the system is built and deployed?
Transparency & Disclosure
Does the vendor disclose how the system works, what it was trained on, where it has been tested, and what its known limitations are?
Evidence & Validation
Are performance claims backed by independent, publicly available evidence, or only by internal statements and marketing?
Human Oversight
Is there a qualified human in the loop? What happens when the system produces an error, and who catches it?
Procurement Readiness
Can the vendor answer the questions a serious public buyer must ask, liability, service levels, error handling, accessibility compliance?
Accessibility Impact
Does the system expand access in practice, or risk replacing higher-quality access with a cheaper, less reliable substitute?
Deployment Context
Is the system appropriate for the specific setting? A tool acceptable for casual information may be unacceptable in a hospital or courtroom.
Data Governance & Privacy
How is signer data collected, stored, and used? Video of a person signing is biometric and identifiable, and deserves corresponding protection.
Institutional Accountability
When something goes wrong, who is responsible, through what mechanism, and with what remedy for the people affected?
For everyone on either side of the decision
- Public agencies evaluating sign language AI in RFPs, pilots, or accessibility programs
- Accessibility offices asked to sign off on technology they were given no evidence to assess
- Procurement teams who need defensible evaluation criteria and vendor questions
- Hospitals and healthcare systems where communication failure carries clinical and legal consequences
- Schools, universities, and education agencies weighing AI tools against accommodation obligations
- Courts and legal-access programs where accuracy and accountability are non-negotiable
- Transportation agencies considering AI signing avatars for announcements and alerts
- Employers evaluating workplace accessibility technology
- Disability service providers advising clients or funding decisions
- Deaf-led organizations seeking structured, documented grounds for their positions
- AI vendors seeking credible independent review before deployment or public claims
If you are being asked to approve, purchase, fund, pilot, or publicly endorse a sign language AI system, a SLAT review gives you a documented basis for the decision.
What you receive
Deliverables are scoped to the engagement and the decision you need to make. Depending on the review type, they may include:
- Vendor Snapshot: a concise public-evidence review of a specific vendor or product
- Procurement risk memo: documented risk issues tied to your intended use case
- Governance review memo: analysis of oversight, accountability, and safeguard structures
- Evidence-gap analysis: what vendor claims are supported, unsupported, or publicly unverifiable
- RFP & procurement language: criteria and requirements you can put into solicitation documents
- Vendor question set: structured questions your team can put directly to a vendor
- Public accountability summary: a plain-language summary for boards, councils, or community stakeholders
- Risk rating summary: findings mapped against the SLAT Index criteria
- Recommended safeguards: conditions, oversight measures, and review points to put in place before a system goes live
Every deliverable is written to be usable inside a real procurement or governance process, not as abstract commentary.
High-risk contexts require stronger evidence
Sign language AI should be held to stronger evidence and oversight in high-impact settings where a communication failure can cause serious harm:
- Emergency services: a missed or garbled message can be life-threatening
- Medical care: informed consent and clinical accuracy depend on communication
- Mental health: nuance, trust, and confidentiality are essential
- Legal & court access: errors affect rights, liberty, and due process
- Public benefits: miscommunication can cost people income, housing, or coverage
- Child welfare: the stakes involve family integrity and child safety
- Education accommodations: access failures compound over a student’s trajectory
- Employment decisions: communication quality affects hiring, evaluation, and discipline
- Transportation & safety alerts: time-critical information must be understood the first time
In these settings, the burden of proof sits with the system, not with the people relying on it, and review belongs before deployment, not after an incident.
A Deaf-led, accountability-first practice
Novara Consulting Group is a Deaf-led practice focused on AI accessibility governance, sign language AI procurement risk, and public-sector technology evaluation. It was founded by Heather M. Grizzle, a Deaf policy analyst who developed the SLAT Index and writes from inside the community sign language AI is marketed to serve.
Deaf-led
Grounded in lived expertise with sign language access, not secondhand assumptions about it.
Governance-focused
We evaluate accountability structures, evidence, and oversight, not demos and marketing.
Procurement-aware
Findings are written for the people who must defend a purchase, draft an RFP, or answer to a board.
Accountability-centered
We are not invested in any vendor’s success or failure, only in whether a system is appropriate, evidence-supported, and safe for its context.
We do not sell enthusiasm, and we do not sell opposition. We provide documented, criterion-based review that organizations can rely on when the decision matters.
The SLAT Index methodology is published and citable: SLAT Index Evaluation Standard (v2) · DOI 10.5281/zenodo.21537271
The SLAT review process
Request
Contact us with a short description of your situation and the decision you are trying to make.
Submit materials
Provide vendor or product information, procurement documents, RFP drafts, or links to public claims.
We review
We assess public claims, available documentation, the proposed use case, and the risk context against the SLAT Index criteria.
We document
We record findings, evidence gaps, risk issues, and recommended vendor questions and safeguards.
You receive it
A written review, briefing, or decision-support memo scoped to your engagement, ready to use.
Evaluate before you deploy
The cost of a structured review is small. The cost of an access failure in a hospital, courtroom, classroom, or emergency is not.
Request a SLAT Review Request a Vendor Snapshot →Request a SLAT review
Tell us the vendor or product, the intended use case, the deployment context, and the decision you are trying to make. We respond within one business day.
Scope of SLAT reviews. SLAT reviews are governance, procurement, and risk assessments. They are not legal, medical, technical, clinical, or linguistic advice, and not a formal certification. The SLAT Index is a preliminary evaluation standard developed by Novara Consulting Group and published as the SLAT Index Evaluation Standard (v2) (DOI 10.5281/zenodo.21537271); it is not a government-approved standard, and a SLAT review does not certify, endorse, or approve any technology. Findings rest on evidence available at the time of review, and public evidence may change. Organizations remain responsible for their own procurement, legal, and compliance determinations.