AI That Listens Like a Lawyer: Courts Are Exposing the Hole Between What AI Notetakers Promise and What Their Contracts Allow. Function-Constructed Authorized Conversational Intelligence™ Instruments, Akin to Querious®, Supply Attorneys a Defensible Path Ahead.
A spot exists between how general-purpose AI notetakers are marketed to authorized professionals and what their phrases of service allow. As courts and regulators start to check that hole, Authorized Conversational Intelligence™ platforms, comparable to Querious®, provide attorneys a defensible different constructed from the bottom as much as align structure, contractual phrases, {and professional} obligations.
Hilary Bowman, Querious, CEO & Founder, Creator
William Love, Managing Associate of EsqLove, Editor
A. Government Abstract
Legal professional-client conversations are essential moments that increase authorized points, uncover factual inconsistencies, and even set off disclosure obligations. Key phrases floor within the stream of dialogue and context might be terribly troublesome to reconstruct after the dialog.
This actuality makes stay dialog one of the worthwhile and untapped use circumstances for AI in authorized apply. In contrast to instruments that assist analysis, drafting, and knowledge evaluation behind the scenes, AI in a stay shopper dialog is a extremely seen utility of the know-how.
Till lately, attorneys who acknowledged this chance confronted an uncomfortable selection: undertake general-purpose AI notetakers that had been by no means designed for the authorized career or go with out. At present, a brand new class of purpose-built know-how, Authorized Conversational Intelligence™, affords attorneys a 3rd path that’s designed to advance each productiveness {and professional} accountability by means of real-time insights throughout a dialog and safe note-taking after the dialog.
The next is a side-by-side evaluation of those two classes of instruments from purposeful, design, contractual, authorized, and moral views.
B. Evaluation
1. What’s the objective of the product?
Basic-purpose AI notetakers are designed to seize all the things that occurs in a dialog. Instruments like Otter.ai, Fireflies.ai, Fathom, and Granola robotically be part of scheduled conferences, file audio and video,¹ generate real-time transcripts, and produce post-meeting summaries. They’re cheap, require no prompting, and combine with Zoom, Microsoft Groups, and Google Meet. For professionals outdoors of regulation, these instruments ship significant productiveness good points. For attorneys, they increase questions on privilege, consent, confidentiality, and biometric privateness that the instruments weren’t designed to deal with.
Querious is the primary Authorized Conversational IntelligenceTM device. Throughout a dialog, Querious accesses audio in real-time, analyzes it, and delivers to attorneys real-time insights about potential authorized points, follow-up questions, and related authorized content material. After the dialog, Querious generates a abstract e-mail and detailed notes, together with a draft billing entry. Querious is on par with the price of general-purpose AI notetakers. The product integrates into all main digital assembly platforms, in addition to legal-specific apply administration instruments comparable to Clio and Smokeball. Whether or not it’s an consumption, advisory or strategic dialog, this product opens up new prospects for attorneys to brainstorm and seize essential particulars in real-time throughout each key authorized dialog.
2. How is the product constructed?
Basic-purpose AI notetakers sometimes create and retailer full audio recordings and transcripts on the seller’s cloud infrastructure, typically indefinitely by default. Deleted content material could stay in a trash folder for weeks earlier than everlasting elimination.² When speaker recognition options are used to differentiate between individuals (a course of often known as speaker diarization), the ensuing voice-derived knowledge could represent biometric identifiers underneath state privateness legal guidelines.³
Querious processes assembly audio and by default, filters personally identifiable data (“PII”) comparable to names, telephone numbers, and identification numbers, earlier than the content material reaches an AI mannequin.⁴ The conversational knowledge is then analyzed by a handful of privately deployed LLMs.⁵ Querious neither creates, nor shops, full audio recordsdata or verbatim transcripts of the dialog.⁶ Any interim speaker diarization is deleted upon conclusion of the dialog.⁷ After a dialog ends, Querious generates the post-meeting summaries and billing entries. Any partial audio or transcript recordsdata created in the course of the evaluation course of is deleted. By default, the lawyer’s summaries are retained for sixty days, until the lawyer units a shorter time interval of their profile.⁸
3. Can the product use my conversational knowledge to coach AI?
Not less than one main vendor of a normal objective AI notetaker explicitly reserves the suitable to make use of machine studying on person content material and utilization knowledge for testing, tuning, and bettering its algorithms.⁹ Others grant themselves licenses to make use of de-identified or aggregated buyer knowledge to coach and enhance their merchandise.¹⁰ These permissions are sometimes buried in dense contractual language that bears little resemblance to the seller’s advertising, which usually emphasizes privateness and safety.
The declare that knowledge is “de-identified” earlier than coaching deserves scrutiny. Distributors not often provide technical explanations of how de-identification takes place. Each courts and regulators have grown skeptical of such assurances. For instance, the FTC has taken an aggressive place on “hashing,” which is a technique that “us[es] math to show [personal data] right into a quantity (known as a hash) in a constant method.”¹¹ The FTC warned that “hashes aren’t ‘nameless’” and “employees will stay vigilant to make sure corporations are following the regulation and take motion when the privateness claims they make are misleading.”¹²
Querious’s structure differs in a number of respects. The phrases of service certify that manufacturing and coaching knowledge pipelines are bodily and logically separated on the community stage.¹³ Thus, it’s technically inconceivable for shopper communications to enter coaching pipelines.¹⁴ Querious makes use of privately deployed AI mannequin situations which can be remoted from the general public web.¹⁵ Because of this, the conversational knowledge isn’t routed to publicly accessible AI platforms. Querious’s agreements with third-party AI suppliers contractually prohibit these suppliers from utilizing shopper knowledge for coaching and require processing by means of personal, remoted situations.¹⁶
4. Does the product assist compliance with wire-tapping legal guidelines?
Basic-purpose AI notetakers that file assembly audio could expose attorneys to legal responsibility underneath federal and state wiretapping statutes.¹⁷ Fifteen states require all-party consent to file a dialog, together with California, Florida, Illinois, Massachusetts, and Pennsylvania.¹⁸ If any assembly participant is positioned in an all-party consent jurisdiction, that state’s regulation could apply and an AI device that data by default, with out prompting the host to acquire consent from every participant, might set off legal responsibility for the lawyer who enabled it.
In distinction, Querious’s default settings are designed to assist compliance with all-party consent necessities. For instance, when Querious joins a digital assembly, the default bot picture seen to each participant within the assembly clearly states that the device is getting used to investigate audio and if you don’t consent to the assembly, to inform the organizer to take away it or exit the assembly. An identical default textual content message is pushed to the assembly chat seen to all individuals with an analogous message relating to consent. Querious additionally permits customers to customise the bot picture, together with including their very own emblem and modifying the disclaimer language to satisfy their particular wants and danger tolerance. Equally, Querious and its “instantaneous assembly” function runs off a tool throughout an in-person assembly, the lawyer can’t begin the audio evaluation and note-taking course of till the lawyer has checked a field acknowledging that each one individuals within the dialog have been knowledgeable that Querious is getting used and the individuals have consented to its use.
5. Does the product assist compliance with biometric privateness legal guidelines?
Basic-purpose AI notetakers that distinguish between audio system by analyzing vocal traits face rising publicity underneath state biometric privateness legal guidelines. In Cruz v. Fireflies.AI Corp. and Basich v. Microsoft Corp., plaintiffs allege that speaker diarization creates voiceprints with out the written discover, knowledgeable consent, or publicly accessible retention and destruction coverage the statute requires.¹⁹ With the Illinois Biometric Privateness Act having statutory damages of $1,000 to $5,000 per violation and over twenty further states now classifying biometric knowledge as delicate private data, the compliance panorama extends nicely past Illinois.²⁰
Querious addresses biometric privateness compliance straight in its structure, phrases of service and devoted compliance paperwork.²¹ From an structure perspective, Querious by default does not distinguish between audio system and it robotically applies a personally identifiable data (“PII”) filter that removes names, telephone numbers, identification numbers and many others. from the dialog knowledge earlier than privately deployed LLMs analyze it.²² When speaker diarization is enabled, Querious analyzes voice traits to cluster audio segments by speaker. Speaker labels within the output are derived from the assembly platform’s participant metadata, not from biometric identification. Querious doesn’t create a mapping between voice attribute knowledge and participant identification. As soon as a dialog ends, this ephemeral speaker-differentiation knowledge is deleted and doesn’t persist as a voiceprint that’s created, saved, or carried throughout conversations.
Even when the evaluation of conversational knowledge by Querious would represent biometric knowledge (i.e., a voiceprint) and implicate sure state privateness legal guidelines, Querious reminds the lawyer to gather consent from all individuals of the dialog. Querious maintains a publicly accessible biometric knowledge retention and destruction coverage as required by 740 ILCS 14/15(a)²³ and helps attorneys in reaching their consent assortment by means of compliance assets within the platform.
6. Does the product assist an lawyer’s moral obligation to maintain shopper data confidential?
Beneath Mannequin Rule 1.6 and analogous adopted guidelines, attorneys should make affordable efforts to stop unauthorized disclosure of shopper data.²⁴
When an lawyer permits a general-purpose AI notetaker, delicate shopper communications are transmitted to a vendor’s cloud infrastructure the place they might be saved indefinitely, used to coach AI fashions, or made accessible in ways in which the lawyer can’t management. The “affordable efforts” framework of Mannequin Rule 1.6 requires attorneys to evaluate the sensitivity of the information, the probability of disclosure, and their means to guage a vendor’s safety practices. The default settings of many general-purpose AI notetakers current a troublesome case for compliance.
Querious doesn’t require attorneys to configure their technique to compliance. Privateness protections are embedded within the platform’s options and default settings. An lawyer evaluating Querious utilizing the “affordable efforts” framework would discover quite a few options to assist and even exceed such a conclusion, together with the PII filter, privately deployed LLMs, and ephemeral entry to partial audio and transcript recordsdata of the dialog. Moreover, the strategy taken by Querious has been validated by means of its profitable completion of SOC2 Sort II certification and integrations with Microsoft Groups, Zoom, Google Meet, authorized apply administration instruments, in addition to a authorized malpractice insurance coverage service.
7. Does the product protect an attorney-client privilege and create documentation coated by work product doctrine?
In United States v. Heppner, the primary federal choice to deal with privilege and work product claims involving AI platforms, Choose Rakoff held {that a} shopper sharing communications from his counsel with a client AI platform’s phrases allowing knowledge assortment, mannequin coaching, and third-party disclosure destroyed any affordable expectation of privateness required to assert safety underneath attorney-client privilege or work product doctrine. If the courtroom’s reasoning was utilized to general-purpose AI notetakers, a vendor’s phrases that allow it to retain, use, or disclose privileged communications, would probably not rise to stage of privateness essential to maintain attorney-client privilege for the dialog or shield the ensuing notes underneath the work product doctrine.²⁵
Notably, nonetheless, Choose Rakoff noticed that the result would possibly differ with an enterprise-grade device that includes contractual confidentiality protections, prohibitions on knowledge coaching, or zero-retention insurance policies.²⁶ All of those options are constructed into the structure of Querious.
Querious’s real-time options additionally strengthen the authorized protections accessible for the ensuing notes and summaries provided that it takes route from the lawyer all through the dialog. In contrast to a general-purpose AI notetaker that produces a verbatim transcript (i.e., a mechanical replica of a dialog that’s troublesome to characterize as something apart from a file of the communication itself), Querious generates authorized concern identification, instructed questions, and structured summaries which can be formed by and aware of the lawyer’s dialog, interplay with real-time prompts, and illustration of a shopper. Thus, the outputs are extra readily characterised as supplies ready to additional improve authorized providers, supporting safety underneath the work product doctrine as supplies reflecting the lawyer’s authorized judgment and impressions slightly than a passive recording of what was stated.
Querious’s contractual commitments reinforce this safety at each stage. The phrases of service set up that Querious operates as a specialised know-how agent underneath the lawyer’s route and management, with a relationship “analogous to that of a paralegal, courtroom reporter, or litigation assist vendor,” which is language designed to protect privilege underneath United States v. Kovel.²⁷ If a subpoena or courtroom order seeks shopper communications or AI-generated outputs, Querious commits to notifying the affected lawyer inside 24 hours, asserting privilege on the lawyer’s behalf, and declining to supply absent a ultimate courtroom order after exhaustion of obtainable appeals.²⁸ Querious additional agrees to cooperate in searching for protecting orders underneath Federal Rule of Proof 502(d), which permits courts to order that disclosure of privileged data in reference to litigation doesn’t represent waiver, together with offering declarations and technical documentation relating to its ephemeral knowledge dealing with structure.²⁹
The result’s a layered protection. Privilege protects the underlying attorney-client communications. The work product doctrine independently protects the AI-generated outputs that stream from the lawyer’s authorized evaluation. And the contractual structure that features company relationship, subpoena protocol, and FRE 502(d) cooperation, offers the procedural infrastructure to claim and defend each protections if they’re ever challenged.
C. Conclusion
The query is not whether or not AI belongs in attorney-client conversations. It’s whether or not the instruments attorneys select are constructed to function inside the authorized and moral necessities that govern these conversations. Because the above evaluation demonstrates, the reply relies upon not on what a vendor guarantees in its advertising, however on what its product structure and phrases of service allow.
Basic-purpose AI notetakers and Authorized Conversational IntelligenceTM instruments could seem to overlap in some performance, however they diverge on each dimension that issues to a working towards lawyer: what knowledge is retained and for a way lengthy, whether or not shopper communications are used to coach AI fashions, how consent obligations are supported, whether or not confidentiality is preserved by default or by configuration, and whether or not the ensuing outputs are positioned to face up to scrutiny underneath each the attorney-client privilege and the work product doctrine.
The authorized panorama is shifting shortly. Consolidated class actions towards AI notetaker distributors are testing wiretapping, biometric privateness, and knowledge coaching theories in federal courtroom. The primary federal choice on AI, work product, and privilege in United States v. Heppner has drawn a transparent line between consumer-grade instruments whose phrases undermine confidentiality and enterprise-grade instruments whose structure preserves it. Attorneys who undertake AI assembly instruments with out conducting due diligence on structure, phrases, and compliance options aren’t simply accepting productiveness tradeoffs, they’re accepting authorized publicity that’s quantifiable, rising, and more and more examined in litigation.
Querious, as the primary Authorized Conversational IntelligenceTM device, was constructed to provide attorneys a defensible reply to every of those challenges, particularly by means of the assist of product structure, contractual commitments, and compliance options that align with the foundations {of professional} conduct. Each architectural declare is mirrored in an enforceable contractual provision. Each privateness safety operates by default. And the platform’s real-time authorized insights create outputs that strengthen each privilege and work product protections in ways in which a verbatim transcript by no means might. The know-how has arrived. The authorized frameworks are clear. The one query is whether or not the device you carry into your subsequent shopper dialog was constructed to satisfy them.
William Love, Managing Legal professional of EsqLove, an Illinois regulation agency centered on progress corporations, and bluechip corporations, the place he advises shoppers on SaaS contracting and privateness; worldwide M&A; asset and expertise acquisitions; and the authorized dangers of rising know-how. As a contract nerd, he has delivered over 150 shows on know-how and AI, managing negotiations and battle, and creating enterprise alternatives by means of well-executed authorized technique.
¹ See, e.g., Otter your conferences wherever they occur, Otter.ai, https://otter.ai/transcription (final visited Mar. 16, 2026); Learn how to seize video on your Fireflies assembly?, Fireflies.ai, https://guide.fireflies.ai/articles/1980499609 (final visited Mar. 16, 2026).
² Phrases of Service § 9.3, Otter.ai, https://otter.ai/terms-of-service (final visited Mar. 16, 2026).
³ Compl., Cruz v. Fireflies.AI Corp., No. 3:25-cv-03399-SEM-DJQ (C.D. In poor health. Dec. 18, 2025), accessible at https://commlawgroup.com/wp-content/uploads/2025/12/Fireflies.ai-Complaint-1.pdf; Compl., Basich v. Microsoft Corp., No. 2:26-cv-00422 (W.D. Wash. Feb. 5, 2026), accessible at https://www.classaction.org/media/mircosoft-teams-bipa-complaint.pdf.
⁴ Phrases of Service § 3.31(b), Querious.ai, http://app.querious.ai/terms-of-service (final visited Mar. 23, 2026).
⁵ Id. § 3.3.
⁶ Id. § 3.4.
⁷ Id.
⁸ Id.
⁹ Phrases of Service §§ 9.9, 18.4, Otter.ai, https://otter.ai/terms-of-service (final visited Mar. 16, 2026).
¹⁰ See, e.g., Phrases of Service, Part 4: Your Information, Coaching AI, Fathom.ai, https://www.fathom.ai/terms (final visited Mar. 16, 2026) (enabling the seller to make use of deidentified Consumer Content material to coach its proprietary mannequin); Granola Platform Phrases § 6.2.2, Granola.ai, https://www.granola.ai/static/Granola-Platform-Terms-2025-12-19.pdf (granting vendor license to make use of and entry Buyer Information to create aggregated and de-identified knowledge to enhance, check, prepare, and function Granola’s services).
¹¹ No, hashing nonetheless doesn’t make your knowledge nameless, FTC.gov (July 24, 2024), tech-at-ftc/2024/07/no-hashing-still-doesnt-make-your-data-anonymous”>https://www.ftc.gov/coverage/advocacy-research/tech-at-ftc/2024/07/no-hashing-still-doesnt-make-your-data-anonymous.
¹² Id.
¹³ Phrases of Service § 3.2, Querious.ai, http://app.querious.ai/terms-of-service (final visited Mar. 23, 2026).
¹⁴ Id.
¹⁵ Id. § 3.3.
¹⁶ Id.
¹⁷ Compl., Brewer v. Otter.ai, Inc., No. 5:25-cv-06911-EKL, Class Motion Grievance (N.D. Cal. filed Aug. 15, 2025), accessible at https://www.fisherphillips.com/a/web/x27EBgcvus2uFdfXMJiyCk/aAQ5CP/brewer-v-otterai.pdf.
¹⁸ Justia, Recording Telephone Calls and Conversations: 50-State Survey, https://www.justia.com/50-state-surveys/recording-phone-calls-and-conversations/ (final visited Mar. 16, 2026).
¹⁹ Compl., Cruz v. Fireflies.AI Corp., No. 3:25-cv-03399-SEM-DJQ (C.D. In poor health. Dec. 18, 2025), accessible at https://commlawgroup.com/wp-content/uploads/2025/12/Fireflies.ai-Complaint-1.pdf; Compl., Basich v. Microsoft Corp., No. 2:26-cv-00422 (W.D. Wash. Feb. 5, 2026), accessible at https://www.classaction.org/media/mircosoft-teams-bipa-complaint.pdf.
²⁰ Illinois Biometric Data Privateness Act, 740 ILCS 14/20.
²¹ Phrases of Service § 5.2(c), Querious.ai, http://app.querious.ai/terms-of-service (final visited Mar. 23, 2026).
²² Id. at § 3.1(a-1).
²³ BIPA Coverage, Querious.ai, www.querious.ai/bipa-policy (final visited Mar. 23, 2026).
²⁴ Mannequin Guidelines of Prof’l Conduct r. 1.6, Americanbarg.org, https://www.americanbar.org/groups/professional_responsibility/publications/model_rules_of_professional_conduct/rule_1_6_confidentiality_of_information/ (final visited Mar. 16, 2026).
²⁵ United States v. Heppner, No. 25-cr-00503-JSR, slip op. at 6–8 & n.3 (S.D.N.Y. Feb. 17, 2026), accessible at https://www.akingump.com/a/web/ssTGsd5NHbtZ1onzXQMTye/1_25-cr-503-27-memorandum.pdf.
²⁶ Id. at 7 (noting that “had counsel directed [the defendant] to make use of Claude, Claude would possibly arguably be stated to have functioned in a fashion akin to a extremely skilled skilled who could act as a lawyer’s agent inside the safety of the attorney-client privilege”).
²⁷ United States v. Kovel, 296 F.2nd 918 (2nd Cir. 1961), accessible at https://law.justia.com/cases/federal/appellate-courts/F2/296/918/131265/; Phrases of Service § 2.2, Querious.ai, http://app.querious.ai/terms-of-service (final visited Mar. 23, 2026).
²⁸ Phrases of Service § 4, Querious.ai, http://app.querious.ai/terms-of-service (final visited Mar. 23, 2026).
²⁹ Id. §14.8.
