Waraaqaha, warbixinnada gaagaaban iyo warbixinnada waxaa lagu soo bandhigi karaa far-qoraal gaar ah oo iyaga u gaar ah. Si toos ah ayay sidaas ku sii socotaa ilaa uusan wax ka mid ah dejinnadaada kore ku dhawaaqin si kale.
Seven productized services for institutions evaluating AI accessibility technology, each tightly scoped, priced to real procurement thresholds, and designed so every engagement creates a natural next step.
Trust ratings for sign language AI, computed from evidence.
DatasetDeaf education, fifty-one jurisdictions, primary sources only.
Public recordWhat public institutions publish about accessibility governance.
AdvisorySeven productized engagements, from a first read to full governance.
TechnologyAccessibility-first apps and the @learn platform.
Each instrument has its own site and its own governance; every one is built from primary sources, states its gaps, and makes no claim beyond what its own rubric supports. The services below draw on all three.
Our published trust framework for sign language AI has grown into its own home. It evaluates five trust domains (linguistic, transparency, governance, accessibility, and deployment trust) through 11 criteria and 54 indicators, and its ratings are computed from evidence, not negotiated from marketing.
Because your existing diligence cannot answer the question that matters.
An accessibility conformance report certifies that an interface works with assistive technology. It cannot say whether a Deaf person understands the output, and no standard procurement artifact can.
Peer-reviewed literature documents that accuracy claims in this market frequently omit what was measured, against what reference, with which signers, under what conditions. SLAT grades the evidence and treats its absence as a finding, never a pass.
Hospital intake, emergency notification, classrooms, courtrooms: a failed output is a safety event and a civil rights exposure. Since 2018, the World Federation of the Deaf and WASLI have cautioned against signing avatars as replacements for qualified human interpreters in live, complex, or high-stakes information.
Findings are criterion-based, Deaf-reviewed, and documented, written for the people who must defend the purchase to a board, an auditor, or a court.
A 51-jurisdiction reference on deaf education in the United States, built state by state from primary sources: enrollment and placement, fiscal allocations, and the institutional frameworks that decide who serves a deaf child and who pays. Sourced claims only, every figure carries a citation, and the gaps in the public record are documented as findings. Extracts, multi-state comparisons, and custom briefs are available as scoped engagements; fees apply.
Because decisions about deaf education are being made without a national record to stand on.
Federal special education reporting aggregates hearing-related categories in ways that dissolve the distinctions that matter, states publish to inconsistent standards or not at all, and the national annual survey that once filled the gap has been discontinued. NDEDI builds the record, jurisdiction by jurisdiction.
Deaf education funding moves through at least six channels: formula aid, state school appropriations, federal flows, interagency billing, capital projects, and scholarships. In the states indexed so far, none publishes a consolidated view, and district formulas rarely disaggregate by disability, so what a state actually spends on a deaf child is in most cases never calculated by anyone.
Among the first states indexed: one operates its deaf schools directly inside the education agency, one under an independent governor-appointed board, one runs a statewide school out of a single local district, one abolished its separate deaf school, and one has no campus at all. No decision about deaf education transfers between states without checking the machinery first.
Every figure carries a citation and a retrieval date. Conflicts between sources are recorded, not resolved silently, and nothing is estimated, modeled, or inferred. Where a state publishes nothing, that absence is documented as a finding. Extracts, comparisons, and briefs can be cited and defended when challenged.
A public record of what public institutions publish about how they govern accessibility, built entity by entity from primary sources: who is accountable, how a person complains or reports a barrier, the standard claimed, effective communication and language access, how technology is bought, and what is said about testing and the plan. Thirty indicators in five domains, six codes, and a Level from 0 to 4 computed by script against a published rubric. Its fourth section, the Procurement Accessibility Record, sets what vendors publish about the platforms those institutions deploy beside what the institutions publish. The Record documents publication; it does not test conformance and it does not state whether any entity complies with any law.
Because obligations are not publication, and nobody was keeping the record.
Every profile opens with the ADA or accessibility coordinator, the complaint route, and the barrier-report route, quoted as published or recorded as not found with the pages checked. A member of the public who cannot find those three things cannot use anything else the institution publishes, so the Record puts them first and measures how findable they are.
The Title II web accessibility rule reaches state entities and large local governments in April 2027 and the rest in April 2028, with the Section 504 rule for HHS recipients on its own calendar. A record of what each institution had published before those dates, re-audited on a published schedule afterwards, is the only way the change will ever be visible.
The Procurement Accessibility Record parses every accessibility conformance report a vendor publishes, criterion by criterion, and places it beside the vendor's marketing claims and the public entities that deploy the product, with flags such as report age, standard drift, and claim divergence shown with their computation. No score, no badge, and no vendor can pay for a row.
Every entity and vendor receives the draft before release with ten business days to point to anything missed, may comment in its own words beside the finding, and may ask for a declined decision to be re-decided by a second person. Nothing is ranked, no payment alters a code, and every relationship with the publisher is disclosed on the profile.
Pricing works two ways here, on purpose. Fixed-scope diagnostics carry set fees, documented and consistent, and priced against public-sector purchasing thresholds; the Trust Snapshot fee is published below, and the full fee schedule is issued on request. Governance-grade engagements are scoped individually, because the governance surface of an AI system that reads, decides, and acts is not the governance surface of a six-policy review.
The full evaluation across all five trust domains: structured vendor documentation review, Deaf reviewer evaluation of output samples, deployment-context analysis specific to your setting, and recommended contract terms and acceptance criteria you can put in a solicitation.
A short preliminary read of a product’s public claims and evidence: what the vendor asserts, what supports it, and what to ask before proceeding. Priced below common direct-purchase thresholds so you can buy it on your own authority.
An organizational review rather than a product one: where AI is already in use across your institution, which tools bypassed procurement, and where accommodation and governance risk sits, with a prioritized remediation sequence.
A structured briefing for boards, commissions, cabinets, and legislative bodies: the taxonomy, the current legal obligations, the procurement risks, and the governance actions, delivered as a written document with presentation.
Recurring surveillance of designated vendors and the surrounding environment: claim changes, product updates, litigation, regulatory movement, and community response. Delivered as a monthly written update with an escalation notice when something material occurs.
Direct support through the purchasing process and the governance that follows it: specification language, evaluation criteria, solicitation review, contract terms, acceptance testing, and standing oversight structures for AI systems whose autonomy, data reach, and decision authority demand them.
Workshops, certificate curricula, and licensed toolkits for procurement teams, accessibility offices, and interpreter coordinators, delivered under license with attribution, so your institution can carry the method forward.
Start small. Each engagement qualifies the next, so you escalate only when the evidence warrants it.
We advise buyers. We do not publish evaluative findings about any organization that has paid us within the preceding twenty-four months, and vendor engagements are private, unrated, and contractually barred from marketing use. All published work discloses funding sources.
How we govern ourselves →SAM registration active, UEI on file, NAICS 541611, Minnesota SWIFT supplier registration, W-9 and HECVAT on request, and a sole source justification package prepared for you.
Where the market fails to build, we build. Auxentia ships accessibility-first apps for navigation, daily living, workplace accommodations, and play, and the @learn platform that packages scarce assistive-technology instruction as software. Available to individuals on the App Store and to agencies, schools, and employers through enterprise and seat licensing outside it.
Start a secure, passwordless consultation from your browser. Message our team directly; every reply is written by a person, supported by AI behind the scenes. Screen-reader friendly, keyboard-operable, with an option to listen.
We help you get the governance right before it ships.
Start a consultation