Hiring guide · Accessibility

An accessibility audit is only half the engagement

Choose around who will change the product and verify the repair. These 20 partners have different roles in the path from an accessibility finding to a working customer journey.

A teal replacement section completes a navy pathway beside an unlettered report and checking frame.
In this article

An audit can identify the inaccessible form while the same form continues to ship. The missing purchase is often the work between finding a barrier and accepting the repaired product. These 20 options cover direct remediation, audit-and-advisory services and managed accessibility programs.

How we chose and ordered these partners

We recommend starting with the responsibility between a finding and an accepted repair. Deque leads because its service explicitly includes source-code remediation. Level Access follows for coordinating fixes across an existing development organization; TPGi, Allyant and TetraLogical provide evaluation, technical guidance and verification suited to teams that can implement. Other managed, specialist and technology-assisted routes follow with their boundaries stated. This is a buying-model order, not a conformance or quality score: if external coding is essential, compare that commitment before commissioning an audit.

Purchasing need Options to compare Proposal requirement
Direct or managed repair Deque, Equalize Digital, UsableNet, AudioEye State how each fix reaches the product and who maintains it
Program and workflow support Level Access, TPGi, Intopia, Accessible Web Connect findings to internal owners and release checks
Specialist assessment and guidance Allyant, TetraLogical, AbilityNet, Hassell Inclusion, DigitalA11Y, BarrierBreak, Accessibility.Works Include technical support and the required retest
Evaluation, documentation and team practice TestPros, Knowbility, Ablr, WebKey, Converge Accessibility Define product coverage and actual repair responsibility

1. Deque: audits with an explicit source-code remediation option

Deque is the more direct starting point when your team wants an outside specialist to work in the source code. Its services span accessibility audits, assistive technology acceptance testing, training, design and developer support, and remediation. Its DevServices description specifically says developers can address issues directly in a client’s code and update or add test cases. That is useful when your existing engineers understand the product but lack time or specialist knowledge to repair complex interactions.

There are several possible purchases inside that range. An audit can define the defects; advisory support can help your developers resolve them; a source-code engagement can supply implementation capacity. Those should become separate lines in the statement of work. The attraction is the potential continuity between the original finding and the repaired component. A product team still needs to agree which repository, release process and regression coverage the outside developers will use.

2. Level Access: coordinating repair across an existing development organization

Level Access is especially relevant when the larger problem is coordinating accessibility work across teams and keeping it visible after the first audit. Its manual evaluation offering puts findings into a shared platform, with integrations to turn them into work in Jira, Azure DevOps and Asana. Its build-and-fix offering combines support, developer tooling, CI/CD feedback and automated remediation for common issues.

That center of gravity suits an organization with its own engineering capacity and a fragmented repair process. Findings can enter the same workflow as other product defects, with progress tracked across the program. An automated fix and a source-code repair are different deliverables, however. If you need an external team to submit code changes, make that a named workstream in the proposal. The platform’s collaboration and remediation language alone does not assign responsibility for implementing every underlying application change.

3. TPGi: evaluation and remediation support within a continuing program

TPGi’s accessibility services include testing, audits, remediation support and conformance documentation, alongside program guidance and training. It is a useful candidate when a team can make code changes but needs specialist evaluation and a continuing source of technical advice. The service range also supports an organization building a sustained accessibility practice across products.

Ask for an example of a finding and the support that follows it. A developer should be able to reproduce the barrier, understand the expected behavior and know where to take a difficult implementation question. Agree whether retesting is included, how it is scheduled and which version it covers. TPGi’s public description supports an evaluation-and-support purchase; direct source-code implementation needs an explicit commitment in the proposal if your own engineers lack repair capacity.

4. Allyant: audit findings followed by verification of repairs

Allyant’s web-audit service explicitly describes retesting after a client starts fixing issues. That matters when the team wants a connection between the original finding and evidence that the barrier has been resolved. It is an evaluation and remediation-support option, with the application team still needing to establish who will make each code change.

Commission the retest at the same time as the audit. Name the journeys, important states and supported assistive technologies, then ask how the provider will track findings across successive builds. If the product includes documents, keep that remediation scope separate from the web application. Ask how the service will coordinate application findings with any document work your release also requires. The result should be an accepted repair against a clear boundary, not only a report showing fewer open tickets.

5. TetraLogical: assessment guidance matched to engineering experience

TetraLogical’s assessment offering covers representative layouts, content, functionality, design systems and component libraries, with assistive-technology testing. It distinguishes reporting packages according to how much implementation guidance a team needs. That is a useful purchasing distinction: engineers experienced in accessibility may need concise findings, while a team learning the work needs more explanation.

Use scoping to choose the level of support, not just the number of pages. Ask how findings about a shared component will be traced across the product and what consultancy or helpdesk support is available while fixes are made. The published model suits specialist assessment and guidance. If you need an external engineer to submit repairs, assign that separately rather than assume the assessment package supplies it. Retest the actual repaired behavior before accepting the work.

6. Intopia: assessments connected to ongoing accessibility operations

Intopia’s service range includes accessibility assessments, conformance reports, research and design, training and AccessibilityOps. It is relevant when the immediate audit sits inside a larger need to make accessibility part of everyday product delivery. The team may need to understand a barrier and change how similar components are designed and tested next time.

Ask which service will address the current release and which will establish a longer operating practice. For the first repair, require reproducible findings, named implementation owners and a retest on the agreed build. For ongoing support, define the decisions specialists will review and how they join the delivery process. Intopia’s range supports that combined conversation, but it should become a specific assignment rather than a broad promise to improve accessibility maturity without completing the product changes.

7. AbilityNet: specialist audits alongside user research and team development

AbilityNet’s digital accessibility services combine audits with user testing, training and strategy. It is worth comparing when a team needs technical evaluation and a better understanding of how people encounter barriers during real tasks. Those two kinds of evidence can inform the same repair while answering different questions.

Define the standards-based assessment and participant research separately. An audit should identify issues within the agreed scope; user sessions can show the practical effect on a journey. Ask how the provider will explain findings to engineers and help prioritize changes. The published service model is consultancy and evaluation, so establish implementation responsibility within your team or a separate development engagement. Include follow-up verification and training only where they support the repaired product and the people who will maintain it.

8. Hassell Inclusion: collaborative examination of complete user journeys

Hassell Inclusion describes live audits and related services that examine user journeys with the client team, alongside training and wider accessibility support. That is useful when a conventional report has failed to create shared understanding among product, design and engineering. A collaborative session can make a barrier and the decision needed to repair it visible to everyone responsible.

Choose one consequential task and include success, error and recovery states in the scope. Ask what documentation remains after the session and how unresolved questions become assigned work. The collaborative format should still lead to reproducible defects and a verification plan. Compare Hassell Inclusion when understanding and coordination are central to the repair, while separately specifying code implementation and retesting. A productive workshop is an input to the release, not the finish line for the customer’s journey.

9. Equalize Digital: WordPress-focused auditing and remediation

Equalize Digital’s audit service describes an audit-and-remediation process, and its wider offering focuses on WordPress websites, applications and plugins. It is a more specific option than a general accessibility consultancy when barriers live in a WordPress theme, plugin or custom implementation.

Supply the actual component inventory and clarify which parts are controlled by your team. A third-party plugin may need a different repair path from custom code, and the proposal should identify that early. Ask how findings will be prioritized, who will implement changes and how updates will affect them. Include an agreed regression scope around shared forms and navigation. Compare the specialist services and software together where useful, but keep the source-code repair and final manual checks explicit in the contract.

10. Accessible Web: audit findings managed as repair tasks

Accessible Web’s audit service places findings into RAMP as remediation tasks with affected elements, severity and instructions. It also describes specialist guidance during the repair process. That fits an internal engineering team that needs a practical way to assign, discuss and resolve accessibility defects across a release.

Ask how the tasks will connect to the system your developers already use and which support level includes the technical help you need. The report should identify the underlying component or state, rather than leave the team to infer a repair from screenshots. Confirm retest terms and which findings need manual verification. This is an audit-and-guidance purchase with workflow software attached; direct implementation capacity needs its own agreement if your developers cannot take on the work.

11. UsableNet: managed services around auditing, fixing and monitoring

UsableNet’s services combine accessibility technology, specialist work and continuing support. Its published model discusses audit findings, what needs to change and verification of remediation for included application scopes. It is relevant when the buyer wants a managed relationship around the repair cycle rather than an isolated audit.

The proposal should explain how changes reach the product. Ask which fixes occur in your source, which rely on supplied technology and which remain with your internal developers. That distinction affects maintenance and what happens if the service ends. Name mobile-app work separately if it belongs in the engagement, because web coverage does not automatically include it. Agree the retest and regression process before choosing the ongoing service level, so the purchase has a concrete repaired journey as well as a monitoring plan.

12. AudioEye: automated and custom fixes with explicit source-repair boundaries

AudioEye’s remediation service distinguishes automated fixes, expert-written custom fixes and issues that need to be repaired at source by the customer’s team. That is the useful detail to examine when comparing a managed accessibility product with direct development services. Its model also includes ongoing monitoring and audits.

Require a proposal that classifies the barriers in your scoped journey by how they will be fixed. Ask which changes persist in your application code and which depend on the service running. Then verify the intended interaction with the agreed assistive technologies after the fix is active. AudioEye fits a technology-and-services comparison, but the presence of automation does not remove the need for manual evaluation or an internal owner for changes the supplier identifies as source-code responsibilities.

13. DigitalA11Y: manual evaluation with developer-facing remediation guidance

DigitalA11Y’s audit service describes manual and automated evaluation, including screen-reader and keyboard testing, with developer-facing remediation guidance. It is a useful comparison when the product team has engineers available but needs specialist diagnosis that can become implementable tasks. The public scope also includes mobile assessment.

Ask to see how a complex interaction is documented: the state, the action, the barrier and the intended behavior should all be clear. Code-level guidance helps a developer start, while the actual framework and shared component structure still need consideration in your implementation. Agree the follow-up support and retest separately from the initial audit. Select this model when your team can own repairs and wants an evaluator capable of discussing the technical details, rather than assuming every audit comes with an outsourced development squad.

14. BarrierBreak: audits linked to remediation support and regression testing

BarrierBreak’s WCAG-audit offer covers static, dynamic and transactional experiences. Its conformance-report process describes an initial audit, customer remediation and regression testing before documentation. That makes the division of labor visible: the evaluator checks and advises, while the customer resolves the issues under that described process.

Use the same separation in a product-repair contract. Ask how findings are discussed with developers, which regression rounds are included and what happens when a shared fix introduces a new issue. If the engagement includes testing by people with disabilities, define tasks and access needs alongside the technical evaluation. BarrierBreak is relevant to a team seeking repeatable assessment and support, with direct application-code implementation needing an explicit scope if you want to outsource that responsibility too.

15. TestPros: independent testing and procurement documentation

TestPros’ accessibility services include manual assessment, prioritized findings and conformance documentation. It separately lists PDF and document remediation. That distinction is useful for a software buyer whose release includes both an application and downloadable material: repairing one does not automatically repair the other.

Ask the proposal to separate application evaluation, document work and verification of fixes. If procurement documentation is required, establish the exact product version and scope it will describe. The application team still needs an implementation owner for source-code changes, unless the supplier explicitly includes that work. Compare TestPros when independent evaluation and an organized record of the resulting state matter, while keeping the practical outcome centered on completing the agreed customer task. Documentation should describe that outcome accurately rather than substitute for it.

16. Accessibility.Works: manual audit followed by consulting and verification

Accessibility.Works’ audit process includes scope definition, automated and manual testing, reporting, remediation consulting and verification. It is worth comparing when the buyer wants a clear path from the first assessment to checking the repaired product, with specialist advice available along the way.

Bring the complete task and its important states to scoping, especially where authentication, dialogs or error handling change what a person can access. Ask which findings will be checked again, how many rounds are included and how adjacent regressions are handled. The public model describes consulting around remediation; confirm source-code implementation separately if required. A useful first engagement leaves your team able to reproduce the checks and maintain the repaired behavior, rather than depend on an unexplained pass status in a report.

17. Knowbility: actionable testing combined with practice-building

Knowbility’s testing and auditing service emphasizes actionable information and recommendations for remediation alongside building an accessibility practice. It is a relevant option when the team needs both a credible assessment of current barriers and a better way to prevent the same problems in subsequent releases.

Ask what the report gives a developer who must change a custom component. Then define the consultation and verification available after that report arrives. A scoped repair should still have named code owners, a release boundary and a retest, even when the wider goal is team learning. Compare Knowbility for evaluation and capability development, while assigning the actual implementation to your engineers or a separately scoped partner. The desired outcome is a usable journey and a team that knows how to preserve it.

18. Ablr: specialist evaluation with people and technology in the process

Ablr’s digital accessibility offering combines human expertise and technology, and its service FAQ describes reports of violations and a remediation plan. It is useful to compare when the buyer wants specialist testing to produce a concrete sequence of work for an internal product team.

Ask how the plan prioritizes barriers in your actual journey and who is available to explain difficult findings. Include the assistive technologies, devices and error states that matter, rather than rely on a generic website scope. The supplier’s broad accessibility language should become explicit deliverables for evaluation, support and retesting. If you lack engineers to make the repairs, require a separate answer about implementation capacity. The selection should turn on a workable division of responsibilities and an agreed verification process.

19. WebKey: technical auditing and consulting for digital resources

WebKey’s public service description identifies technical accessibility auditing, training and consultancy for websites and other digital resources. It is a candidate when the product organization needs an evaluator and a practical source of advice while its own people make changes. The training component may help teams that are taking on this responsibility for the first time.

Specify the resource types and journeys in the brief. A website audit, a document assessment and a native application test need different expertise and test conditions, so confirm coverage for the actual product. Ask what the technical report includes and how implementation questions and retesting are commissioned. WebKey’s described role supports a consulting-and-evaluation model; source repair needs an assigned engineer and should not be inferred from the availability of accessibility advice.

20. Converge Accessibility: manual evaluation with a named fix owner

Converge Accessibility’s evaluation service describes manual assessment of key pages and tasks, followed by findings that identify the relevant fix owner. Those owners may be the platform, content team, a third-party component or developers. It also identifies re-evaluation as an optional engagement, which is useful to know before comparing audit prices.

This model fits a product with mixed ownership, where finding the barrier is only the first part of getting it repaired. Ask how the evaluator will distinguish a configuration change from a component defect and help resolve disagreements over responsibility. Commission the re-evaluation explicitly if accepted repairs are part of your goal. The useful deliverable is a traceable route from each barrier to an owner and a checked outcome, without treating a prioritization score as proof that the journey works.

Define the finish line before choosing the partner

A scan result is useful input, but it is an incomplete acceptance criterion. W3C’s guidance on evaluation tools explains that tools cannot check every accessibility aspect and require human judgment. Buy a combination of evaluation and repair work suited to the product, with the standard, version and target level written down.

The table separates the two public offers from the additional terms a buyer should commission. The table contrasts two different service models; exact inclusions belong in each proposal.

Workstream Deque’s described offer Level Access’s described offer Term the buyer should add
Specialist audit Accessibility audits and prioritized findings Manual evaluations and centralized findings Agreed journeys, states, standard and target level
Assistive-technology evaluation Assistive technology acceptance testing listed Manual evaluation offered; exact device/technology matrix needs scope Named combinations and test tasks
Source-code repair DevServices can work directly in code Build/fix tools and expert support; direct code implementation not specified on cited pages Named engineer, repository access and code-review owner
Automated checks Axe tooling accompanies the wider services Scanning and CI/CD feedback in the platform Which checks run before each release and who handles failures
Retest Agree the exact retest workstream Agree the exact retest workstream Original defects retested on the released version, with status and exceptions
Sustaining the repair Training and test-case support Governance, issue tracking and development tooling A component owner and regression coverage after handoff

The decisive question is the third row. If nobody on your team has capacity to change the code, an excellent audit and a useful tracking platform still leave you with an unfinished product. If your developers can do the work, specialist diagnosis and retesting may be the better use of the external budget.

A repair engagement for an account-recovery flow

Consider a hypothetical product whose password-recovery journey contains an email field, an error message, a confirmation screen and a modal for updating an expired recovery link. The proposed scope is that complete journey, including failed and successful states. It is not a whole-site conformance claim.

The product owner supplies representative accounts, access to the recovery emails and the app’s supported browser/assistive-technology policy. The accessibility evaluator tests the agreed journey manually and with appropriate automated checks. Each finding needs a reproducible state, the barrier encountered, the relevant criterion and a recommended outcome. A screenshot alone cannot describe a focus-order or announcement problem.

The remediation engineer then repairs the shared components and the journey-specific behavior. The internal engineer reviews and releases those changes. The evaluator retests the released version against the original findings and looks for regressions in the agreed adjacent paths.

Use the accessible-onboarding checklist to turn labels, focus, authentication and error recovery into task checks for this scoped repair.

Example defect in this hypothetical journey Repair acceptance
The email field’s visible label is not programmatically associated The input has an appropriate accessible name and its purpose is clear in the agreed manual checks
Validation moves the page but leaves the error undiscoverable The user can locate the error and understand how to correct the field with the agreed assistive technology
Opening the expired-link modal leaves keyboard focus behind it Focus enters the modal appropriately; keyboard navigation and closing behavior work as specified
Closing the modal loses the user’s position Focus returns to a sensible control and the person can continue the recovery task
Shared form changes break an adjacent sign-in screen The adjacent agreed regression cases pass before release

Close each defect with a retest result tied to the version that contains the fix. Keep unresolved exceptions visible with their owner and next action. A finding marked “developer complete” should not become “accepted” until the agreed retest has happened.

User testing and conformance answer different questions

People with disabilities can reveal barriers that an engineering checklist misses in practice. W3C recommends involving users early and throughout a project and combining that work with standards-based evaluation. A successful session with one participant does not stand in for evaluating the agreed criteria across the journey.

Commission participant recruitment, compensation, access needs and actual tasks as a distinct workstream if you need user testing. For the recovery example, the task is regaining access to an account, including an expired link and a correctable input error. The report should show where participants needed assistance and how the repaired experience changed those barriers. It should avoid treating participants as a pass/fail instrument for the entire product.

Keep repair ownership after the retest

The useful handoff is a functioning journey plus the knowledge to preserve it: repaired components, reproducible checks, remaining exceptions and an owner for future changes. Decide whether the provider will retest subsequent releases or train your team to do that work.

Choose a direct repair service when implementation capacity is missing. Choose evaluation and specialist support when your engineers can make the changes. For an ongoing program, define how future releases will be checked. Every model should lead to an accepted repair in the product your customers use.

Sources

Frequently asked questions

Does an accessibility audit include fixing the code?

Not automatically. An audit identifies barriers; source-code remediation, developer advisory support and retesting are separate responsibilities that should be assigned and priced explicitly in the proposal.

What should accessibility remediation acceptance include?

Tie each original finding to the repaired release and an agreed retest. Include relevant manual and assistive-technology checks, automated checks where useful, regression coverage and a clear record of unresolved exceptions.

Related reading

Explore more in Product & Design →