Web Accessibility Vendors — Cost-Benefit Analysis

Our Place · fromourplace.com · prepared July 6, 2026 · sources: vendor quotes + Granola meeting notes. No recommendation — criteria and mapping only.

A.Feature & capability matrix

Per your direction: overlay / DOM-level remediation is excluded — only code-level (source/PR) capability counts. Allyant kept in as a real (if disfavored) option. UsableNet modeled on its AaaS ("Aqua") plan (your team remediates in code); its Fully-Managed DOM-script tier is excluded under the no-overlay rule.

Criterion TestParty accessiBe Level Access Allyant
(incumbent)
UsableNet
(AaaS / Aqua)
Remediation model Agentic Agentic* Non-agentic Non-agentic Non-agentic
Code-level fixes (not DOM) Yes — automated code fixes Yes — Access Code / Solve Yes (overlay excluded) No — advisory only Yes — AaaS, your team fixes source
Automated remediation / agent swarming Yes — only vendor live today Access Solve — July '26 target Inline fix suggestions only No No — your team remediates
Shift-left: CI/CD + PR scanning Yes (PR reviews) Yes — GitHub Actions, ~85% coverage Yes — GitHub Actions thresholds No IDE linter + CI/CD (GitHub) + browser extension
MCP Not mentioned Yes — Cursor/Claude/IDE Yes — IDE-level (top differentiator) No Not mentioned
SDK / API Not mentioned SDK (QA + agents) + API SDK + REST API (mid tier+) No Ticketing/Shortcut API only
Manual expert audit included Monthly, IAAP-certified Via Access Flow 1 audit + iterative Its core strength Annual + 3 QA rounds
Coverage cadence Continuous (weekly scan + ongoing PRs) Continuous (per-PR + biweekly rescans) Monitoring continuous; audit = snapshot Snapshot (~3 rounds / 3 yrs) Continuous automated + quarterly blind-user testing
Feature readiness Live Access Code live; Solve + legacy scan pending Live (SDK/IDE detail was open) Live AaaS live incl. CI/IDE tooling
Legal defensibility / warranty $15K/yr credit warranty + legal reports Monitoring + audit trail Audit + docs Conformance letters, affidavits, rebuttals Litigation support included (indemnity only on Fully-Managed)
Pricing predictability / coverage exposure Flat monthly, unlimited licenses + 12 hrs/mo support — predictable Flat annual, Access Solve included — predictable Mostly flat, but only 1 audit; extra audits billable Hourly buckets ($155–165/hr) — has historically under-covered; extra audits from page changes went unbought Flat — unlimited runs/pages; priced by domain + users

* accessiBe is agentic once Access Solve launches (July '26 target; September floor if delayed). Until then Access Code handles new-code PRs; legacy code needs an MCP workaround that adds overhead.

UsableNet deck (Jul 6): confirms AaaS has IDE linter + CI/CD (GitHub), continuous + quarterly blind-user testing, and included litigation support — reflected above. Cost/effort unchanged: UsableNet's own slide rates AaaS engineering "Medium–High" with a 3–6 month ramp, consistent with the non-agentic $17K eng line. Fully-Managed "Assistive" — excluded (confirmed): despite the deck's "fixes the underlying code" framing, Assistive remediates via runtime-injected script. It is an overlay — more intentional than accessiBe's (its injected changes come from a human code audit rather than purely machine-read DOM fixes), but overlay-based nonetheless. Excluded under the no-overlay rule. AaaS is therefore the only in-scope UsableNet option.

B.Cost model & assumptions

TCO = vendor license + engineering cost

C.Total Cost of Ownership

1-Year TCO

3-Year TCO (cumulative)

Vendor license Engineering cost (dev-weeks)

VendorTypeLicense / yrEng cost / yr1-yr TCO3-yr TCO

Allyant note: the modeled $10K reflects the fixed hour bucket that has historically under-covered — one basic audit + remediation/yr, with needed extra audits unbought. It is the cheapest line only at that reduced coverage level; buying enough hours to match continuous coverage would move Allyant materially higher, so this row is a coverage floor, not a like-for-like comparison.

D.Legal risk exposure & risk-adjusted TCO

Risk-adjusted TCO = license + engineering + expected annual settlements

Risk-adjusted 1-Year TCO

Risk-adjusted 3-Year TCO

License Engineering Expected settlements

VendorLicenseEngResidual exposure / yrRisk-adj 1-yrRisk-adj 3-yr

Do-nothing reference: ~$35K/yr (~$105K over 3 yrs) in expected settlements at current trajectory. Residual figures are assumptions tied to the coverage-cadence criterion, not vendor promises — adjust the per-letter cost ($10K), frequency (3–4/yr), and reduction levels to test sensitivity.

E.Reading the numbers (neutral)

What the dollar model says

  • On raw dollars, Allyant is cheapest in both horizons — but only at reduced coverage. Its $10K is a fixed hour bucket that has never funded full coverage; matching continuous coverage would raise its real cost. It is a coverage floor, not a like-for-like line.
  • Among non-agentic options, Allyant undercuts Level Access (~$40K → higher, driven by license + full eng).
  • The two agentic options (accessiBe ~$31.4K, TestParty ~$32.6K) sit close together and below Level Access.
  • The agentic-vs-Level-Access comparison confirms your thesis: a higher sticker can still be lower all-in once engineering time is priced.
  • Once expected settlements are priced in, the ranking reverses. Allyant's raw-dollar edge disappears — its under-coverage carries the highest residual exposure (~$27K/yr), pushing its risk-adjusted TCO (~$54K/yr) above the agentic options (~$41–43K/yr). Continuous/agentic solutions become the lowest risk-adjusted cost.
  • Because the do-nothing exposure (~$105K/3yr) exceeds most solutions' license, mitigation effectiveness outweighs license price in the total picture.

What the dollar model does not capture

  • Recurring vs. avoidable labor: Allyant's $17K/yr is real engineer time diverted from product; agentic tools redeploy it. PR volume also declines over time, pushing agentic eng cost below $1,400.
  • Coverage cadence: snapshot vendors leave undetected drift between audits — unpriced legal exposure — and produce the ~200-issue batches that caused the backlog.
  • Time-to-remediate: continuous/agentic fixes reach and hold conformance far faster than an audit-then-fix cycle, shrinking the window of exposure.
  • Readiness risk: accessiBe's agentic advantage depends on Access Solve landing on schedule.
  • Pricing predictability: usage-based models (Allyant hourly, UsableNet per-page) force a trade-off — pay more when scope exceeds the bucket, or cap spend and accept thinner coverage. The $17K eng estimate for Allyant assumes buckets that may not hold. Flat/unlimited models (TestParty, accessiBe) fix cost and remove the coverage-vs-budget tension.

Prepared from: TestParty POC & Annual SOWs, accessiBe proposal, Level Access proposal, Allyant 2025 renewal SOW, internal legal-exposure history, and Granola meeting notes (Jun 4–30, 2026). UsableNet: Jun 24 meeting notes + verbally quoted AaaS pricing (~$20K/yr, 1 domain, <5 users, unlimited runs/pages); written proposal still pending.