The federal relay rule is often compressed into “answer in 10 seconds.” That shorthand loses the denominator, compliance period, covered service, and receiving party. Under 47 CFR 64.604, the principal standard is at least 85% of calls answered within 10 seconds, calculated daily for covered telecommunications relay service (TRS) calls. Video Relay Service (VRS), Speech-to-Speech (STS), and IP Relay have category-specific provisions. These are minimum standards for certified relay providers, not a universal service-level agreement for every business telephone queue.
That distinction matters because the evidence describes three different things: how quickly a relay provider connects a communications assistant (CA), how many people report selected functional difficulties, and what a business must do to communicate effectively. None is a direct estimate of how many relay-originated calls a particular answering operation will receive.
The compliance unit is a distribution, not an average
The FCC's general threshold requires 85% of incoming covered calls to be answered within 10 seconds by a method that allows the call to be placed immediately rather than putting the caller in a queue or on hold. The clock starts when the call reaches the provider's network, abandoned calls are included, and performance is calculated each day. A daily percentage therefore describes a call-attempt cohort over one calendar day, measured in calls and seconds, across a provider's U.S. relay operation. It does not report mean speed, median speed, or the experience of the slowest 15%.
The category rules prevent a single number from representing all relay traffic:
| Relay category | Current FCC answer-time threshold | Compliance period and unit |
|---|---|---|
| General covered TRS | 85% within 10 seconds | Daily share of incoming calls |
| IP Relay | 85% within 10 seconds | Daily share of incoming calls |
| Speech-to-Speech | 85% within 20 seconds | Daily share of incoming calls |
| Video Relay Service | 80% within 120 seconds | Monthly share of incoming calls |
These thresholds are not interchangeable. VRS uses sign-language video and a monthly 80/120 measure; STS uses CAs trained to understand callers with speech disabilities and a daily 85/20 measure. Traditional text relay generally connects a TTY user and a voice telephone user. Captioned Telephone Service and IP Captioned Telephone Service provide captions while the user listens and speaks. The FCC consumer guide also distinguishes Voice Carry Over, Hearing Carry Over, shared non-English relay, IP Relay, and VRS.
A percentage can conceal different tails. Two providers can both pass 85/10 even if calls beyond 10 seconds wait for very different lengths of time. Monthly VRS aggregation can also offset a weak day with stronger days, whereas a daily standard cannot. Consequently, “average answer time” alone cannot establish compliance, and a monthly blended score cannot establish daily compliance for a service governed by the daily test.
FCC oversight contains more than the threshold. Section 64.606 governs certification and provider reporting, while the FCC publishes TRS program data and identifies certified internet-based providers. Fund-compensated conversation minutes are useful for program scale and category trends, but minutes are not call attempts. They cannot reproduce a speed-of-answer fraction without the underlying offered-call, answer-clock, abandonment, service, and period fields.
Disability counts are population denominators, not call forecasts
The 2023 American Community Survey (ACS) one-year subject table S1810 estimated 44,105,436 people with a disability, or 13.4%, among 328,315,123 civilian noninstitutionalized people in the United States. The period is calendar year 2023; the unit is people; the cohort excludes the institutionalized population; and the geography is the 50 states and District of Columbia represented by the U.S. estimate. The values and field definitions are available through the Census API result and S1810 metadata.
This is a defensible denominator for general disability prevalence, but a poor multiplier for relay staffing. ACS disability status is built from six functional domains: hearing, vision, cognitive, ambulatory, self-care, and independent-living difficulty, with domains asked only of relevant age groups. It does not ask whether a respondent uses TRS, which relay category they use, how often they call businesses, whether they prefer text or video, or whether accessible direct communication makes relay unnecessary. A person can report more than one difficulty, while many people counted in the all-disability total have no communication-related difficulty.
The ACS is also a survey estimate rather than an administrative count. Sampling error, response error, proxy response, question wording, and the civilian-noninstitutionalized scope limit interpretation; the ACS methodology explains its sample design. Even a hearing-difficulty estimate is not a count of Deaf sign-language users, TTY users, captioned-telephone users, or VRS callers. STS demand is especially poorly represented because the six-domain ACS measure has no standalone speech-difficulty category.
The correct evidence hierarchy is therefore narrow: ACS describes population prevalence; FCC relay records describe program activity; a business's own offered contacts describe its demand. Joining those datasets without a measured conversion rate produces false precision.
What disability-access law requires of the receiving business
The FCC standard governs the relay facility's leg of the communication. The receiving business is governed by a different legal and operational question: whether communication is effective. Under 28 CFR 36.303, covered public accommodations must furnish appropriate auxiliary aids and services when necessary, subject to the regulation's defenses and circumstances. Critically, a public accommodation receiving calls through a federally established relay service must answer those calls in the same manner as other telephone calls.
DOJ's effective-communication guidance emphasizes that the aid or service must fit the nature, length, complexity, and context of the communication and the person's usual method. That is broader than speed. Accurate turn-taking, enough time for typed or interpreted exchanges, accessible authentication, and an alternative when voice-only automation fails can determine whether a completed answer is meaningful.
The Access Board's ICT standards and telecommunications guide add technical evidence around real-time text, two-way voice communication, interoperability, and operable parts for federal and covered telecommunications contexts. They are not a private answering service's 10-second rule. They demonstrate why access cannot be reduced to one queue metric.
Staffing implications without importing the wrong standard
For relay providers, the daily 85/10 rule creates a tail-capacity problem: staffing must absorb within-day peaks because a strong week cannot cure a failed day. VRS capacity is evaluated differently because its 80/120 result is monthly and requires video-capable CAs with sign-language competencies. STS likewise requires specialized communication skill, not merely another generic seat. Combining categories in one occupancy or answer-speed result can hide a service-specific failure.
For a business answering operation, FCC thresholds can inform stress testing but not certify compliance. Its clock normally begins only after the CA reaches the business line; the relay user's total wait can include relay answer time, call placement, the business queue, and authentication. Business records generally observe only the latter portion. Labeling that result “TRS answer time” would conflate two systems and two cohorts.
Staffing evidence is strongest when relay-originated contacts remain visible as a service cohort without being deprioritized. Longer conversational duration may affect occupancy because interpreted or text-mediated turns can take more time, but no federal source supplies a universal relay-call handle-time multiplier for businesses. A valid capacity model therefore uses observed arrival intervals, handling time by channel, abandonment, transfers, and repeat contacts, while retaining the same answering treatment required by DOJ. Training affects capacity too: agents who mistake a CA's introduction for solicitation, interrupt interpreted turns, or let a short IVR timeout expire can create repeat work even when the initial answer was fast.
Limitations and conclusion
This analysis does not determine whether any organization is ADA-covered, whether an auxiliary aid would be an undue burden or fundamental alteration, or whether state law sets additional duties. It does not equate ACS disability with relay eligibility or use. FCC fund minutes are not offered calls, and published national program data cannot predict one firm's half-hour demand. The regulation's percentile thresholds reveal nothing about waits beyond the cutoff or communication quality after answer.
The measured conclusion is narrower: U.S. relay compliance is category-specific and denominator-dependent. The principal daily test is 85% of covered calls within 10 seconds; STS uses 85% within 20 seconds daily, and VRS uses 80% within 120 seconds monthly. The 2023 ACS estimated 44.1 million civilian noninstitutionalized U.S. residents with a disability, but that population count is not a relay-demand forecast. Businesses should treat public relay standards as evidence about access architecture, not as a borrowed business SLA; their operative evidence must connect equal answering treatment and effective communication to the demand actually observed on their own lines.