Answer
Can you use AI voice tools with Illinois customers?
Only with care other states do not require. Analysing a voice to identify a speaker is biometric processing, and individuals can sue.
Yes, if the tool does not analyse voices to identify or distinguish speakers. Illinois treats a voiceprint as a biometric identifier requiring written release, and unusually, individuals can sue directly.
Illinois is the state that changes this answer, and it changes it through a statute written well before any current tool existed. Its biometric privacy law lists a voiceprint among biometric identifiers, and it requires a private entity to inform the subject in writing that biometric information is being collected, to state the purpose and the retention period, and to obtain a written release, before collection. It also requires a published retention and destruction schedule. None of that is unusual in the abstract; what makes Illinois different is that individuals may sue for violations directly, which is why the case law here is far ahead of anywhere else.
The trigger is the analysis, not the recording, and this is the distinction most businesses get wrong in both directions. A call recording is governed by the state's recording consent rules and is a separate question. What the biometric statute reaches is the extraction of the characteristics of a voice in order to identify or distinguish a person — which is precisely what speaker identification, voice authentication, and the speaker separation inside many transcription and meeting tools perform. A tool that transcribes without distinguishing speakers is in a different position from one that labels who said what.
Which means the practical question is technical and specific: does this product build a representation of the speaker's voice, and does it use it to identify or distinguish. That is a question for the vendor's technical documentation, and it is answerable. It is not answered by the vendor's assurance that the product is compliant, by the absence of the word biometric in the marketing, or by the fact that audio is not retained — a voiceprint derived and discarded within the session has still been collected.
Coverage follows the individual rather than the business, which is the part that catches companies with no Illinois presence at all. A contact centre in another state taking calls from Illinois residents, a meeting tool joining a call with an Illinois participant, or a receptionist product answering an Illinois customer are all reaching people the statute protects. Where the business happens to be located is not the question the statute asks.
The compliant path exists and it is procedural rather than technical. Written notice before collection stating purpose and retention, a written release, a published schedule for destruction, and a record that the release was obtained. For an inbound call that is genuinely awkward, which is why several businesses take the other route: disable speaker identification for Illinois numbers, or select a product that does not perform it, and treat the feature as one that has a jurisdictional cost.
The wider point is worth taking from this case rather than only the Illinois answer. Several states have biometric provisions inside broader privacy statutes, and the direction of travel is that voice is treated as sensitive data requiring consent rather than notice. A business designing a voice product now is better served by making speaker analysis a feature that can be switched off per jurisdiction than by establishing whether it is currently permitted in each.
The exposure attaches to the analysis rather than to the recording, which is why a tool that never keeps the audio can still be the problem.
Siddharth Sharma, Context Theory
Related questions
Does an announcement at the start of the call satisfy it?
It handles the recording consent question in many states and does not satisfy the biometric requirement, which asks for written notice and a written release before collection. A verbal notice on an inbound call is difficult to reconcile with that, which is the honest reason so many businesses disable the feature for Illinois rather than build a compliant flow around it.
What if the tool only identifies our own employees?
The statute protects individuals, and employees are individuals. Employment-context biometric claims have been among the most common, and consent obtained as a condition of employment is a fraught basis. The analysis is the same one applied to customers, with the additional complication that an employee's ability to decline is questionable in a way a customer's is not.
METHOD
Every figure below carries its source and the date it was verified. Nothing on this page is asserted.
The numbers on this page.
| What | Value | Specific to |
|---|---|---|
| All-industry average search CPC | $5.42 | Category-wide |
| Sub-15-minute compliance — automated routing vs manual only | 62.5% vs 39.1% | Category-wide |
LocaliQ / WordStream Search Advertising Benchmarks 2026 · Google + Microsoft Ads, 20 industries · Apr 2025–Mar 2026 · verified
2026 speed-to-lead benchmark · verified
What is specific to this page.
| Kind | Claim | Check it against |
|---|---|---|
| Regulation | Illinois lists a voiceprint among biometric identifiers and requires written notice of collection, a stated purpose and retention period, and a written release before a private entity collects one, together with a published destruction schedule. | The notice, release and retention provisions of the Illinois Biometric Information Privacy Act. |
| Constraint | The statute is triggered by extracting voice characteristics to identify or distinguish a speaker rather than by recording audio, so speaker separation inside a transcription product engages it while plain transcription does not. | The vendor's technical documentation on whether speaker diarisation or voice identification is performed, and on what representation it derives. |
| Software | A voiceprint derived and discarded within a session has still been collected, so a vendor's assurance that audio is not retained does not answer the question the statute asks. | Asking the vendor specifically whether any voice-derived representation is computed, separately from whether audio or transcripts are stored. |
| Response | Coverage follows the individual, so a contact centre, meeting tool or receptionist product operating entirely outside Illinois reaches the statute whenever an Illinois resident is on the call. | Checking whether the business can identify inbound calls originating from Illinois before the speaker analysis feature engages. |
Each row would be wrong on another industry's page. Where a sourced figure exists it is in the table above instead; these are the constraints that shape the work and do not happen to be numbers.
Start with the measurement.
Reading about a benchmark is not the same as knowing your own number. The audit produces yours, measured rather than estimated.
$497 · delivered in 5 business days · credited against month one