AI Legal Intake Software: What It Can and Can’t Do

Quick answer: AI legal intake software may take a caller’s name and contact details, the other parties’ names for a conflict check, the matter type and a short factual summary, then book a consultation. Florida Bar Opinion 24-1 (19 January 2024) says it must first say it is AI, not a lawyer, and applies three conditions to that interview. Merits, deadlines, fee-agreement questions and the conflict decision go to a lawyer.

Below, each intake task is mapped to ABA Model Rules 1.6, 1.18, 5.3 and 7.3, ABA Formal Opinion 512, the Australian Solicitors’ Conduct Rules 2015 (ASCR) and the Evidence Act 1995 (Cth), and Smith.ai, CaseGen and Lawmatics are compared. This is not legal advice; your own bar or law society’s rules bind you.

Where Zian fits, and where it does not

Zian AI is an autonomous AI sales-agents platform. It is not marketed for legal intake. Its named integrations are HubSpot, Salesforce, HighLevel and Zapier, alongside an API and connections to “calendars, knowledge bases, and more”; its home and features pages name no legal practice-management system and no conflict-check feature.

The overlap is the work after intake. Zian publishes phone, SMS, email and WhatsApp agents, a live web-chat widget and a live website voice chat agent in 30+ languages, with configurable knowledge-base scope and rules, human-in-the-loop approvals and audit trails. Private model deployment on your own infrastructure bears on Opinion 512’s question of who can access what callers say. SmartReach AI™ paces follow-up; PrecisionPitch AI™ split-tests scripts. That fits disclosed, opt-out-respecting follow-up to an enquiry already in hand. The hand-off line still applies: booking a consultation is fine; telling a caller whether they have a case is not.

What AI legal intake software is allowed to do on the call

Florida Bar Ethics Opinion 24-1 (19 January 2024) applies an older Florida opinion on human intake staff, Opinion 88-6, to AI chatbots. As 24-1 summarises it, a nonlawyer “may conduct the initial interview with a prospective client” but must:

1. clearly identify their nonlawyer status to the prospective client;
2. limit questions to the purpose of obtaining factual information from the prospective client; and
3. not offer any legal advice concerning the prospective client’s matter or the representation agreement, and refer any legal questions back to the lawyer.

For AI specifically, a chatbot that talks to clients or third parties “must comply with restrictions on lawyer advertising and must include a disclaimer indicating that the chatbot is an AI program and not a lawyer or employee of the law firm.” And “a lawyer may not delegate to generative AI any act that could constitute the practice of law such as the negotiation of claims or any other function that requires a lawyer’s personal judgment and participation.”

Two limits: it is a Florida advisory opinion, which states “Advisory ethics opinions are not binding”, and persuasive at most elsewhere; and it starts from permission (“Lawyers may use generative artificial intelligence”), describing AI intake done properly rather than banning it. ABA Model Rule 5.5 comment [2] explains why no national line exists: the practice of law “is established by law and varies from one jurisdiction to another.”

The quotable version: an AI intake agent collects facts and books time with a lawyer; the moment a caller asks what the facts mean, the agent’s job is to hand over.

The hand-off line: each intake task, and who must do it

Where no professional rule decides a row, the table says so.

Intake task	AI agent may	Lawyer must	The rule that decides it
Opening line	Say it is AI, not a lawyer or firm employee, before taking any details	Approve the wording	Fla. Op. 24-1 (disclaimer); Fla. Op. 88-6 condition 1, as summarised in 24-1
Name, phone, email, preferred contact time	Yes	No	Fla. Op. 88-6 condition 2 (factual information)
Names of the other side and related parties	Yes, and before the story	No	ABA MR 1.18 cmt [3] (information needed for a conflict check)
Already represented by another lawyer?	Ask, and narrow the call if yes	Decide what happens next	Fla. Op. 24-1 (“should consider including screening questions”)
Run the conflict search	Software may search and flag	Decide whether a conflict exists and whether consent or a screen is possible	ABA MR 1.18(c)–(d) and cmt [4]; ASCR rr 10.2, 11.3–11.4
Matter type and practice area	Yes, for routing	No	Fla. Op. 88-6 condition 2
What happened, and when	A short factual summary only	Take the detail once conflicts clear	ABA MR 1.18 cmt [4]; 1.18(d)(2) (“reasonable measures to avoid exposure”)
“Do I have a case?”	No. Log the question and book a lawyer	Yes	Fla. Op. 88-6 condition 3; Fla. Op. 24-1 (no delegation of the practice of law); AU joint statement (AI tools “cannot reason, understand, or advise”)
“When is my deadline?” (limitation date)	Record the incident date as a fact	Work out and advise on the deadline	Fla. Op. 88-6 condition 3
“What will it cost?”	Read out the firm’s published consultation terms, word for word	Explain the retainer or agreement and answer any question about it	Fla. Op. 88-6 condition 3 (“or the representation agreement”)
Book the consultation	Yes	No	No professional rule restricts scheduling; firm policy
Send the engagement agreement	Send the firm-approved document once a lawyer has accepted the matter	Accept the matter and clear conflicts first	ABA MR 1.18 cmt [4]; Fla. Op. 88-6 condition 3
Call back someone who enquired	Yes, unless they asked not to be contacted	Set the stop rules	ABA MR 7.3 cmt [1] (a response to a request for information is not solicitation); opt-outs: telemarketing and spam law, and 7.3(c)(1) if a call is a solicitation
Cold-call people known to need a lawyer	No	Not by live voice, human or AI, where 7.3(b) applies (pecuniary motive and none of its three exceptions)	ABA MR 7.3(b); in Australia, ASCR r 34.2 bars only oppressive or harassing approaches to people at a significant disadvantage
Record and store the call	Yes, under the firm’s safeguards and recording-consent law	Vet the vendor and its retention terms	ABA MR 1.6(c); ABA MR 5.3 cmt [3]; ABA Op. 512 (vendor checks); ASCR r 9.1
Caller in crisis or danger	Transfer to a person at once	Staff the transfer line	No specific conduct rule; firm policy

The “Lawyer must” column is all judgment: whether there is a case, a conflict, a deadline or a fee question needing advice. The transfer itself is covered in how to pass context when an AI agent hands a call to a human.

Why the conflict questions must come before the caller’s story

Under Model Rule 1.18, a person who consults a lawyer “about the possibility of forming a client-lawyer relationship with respect to a matter is a prospective client” (1.18(a)). Even if the firm never takes the matter, it “shall not use or reveal that information”, except as Rule 1.9 would allow for a former client (1.18(b)).

Under paragraph (c), a lawyer who received information “that could be significantly harmful to that person in the matter” cannot act for a client with materially adverse interests in the same or a substantially related matter, and no other lawyer in the firm may knowingly take it on or continue with it, except as paragraph (d) allows. Paragraph (d) gives two ways out:

• Consent: both the affected client and the prospective client give informed consent, confirmed in writing; or
• Screening: the lawyer who received the information “took reasonable measures to avoid exposure to more disqualifying information than was reasonably necessary to determine whether to represent the prospective client”, the disqualified lawyer is timely screened and takes no part of the fee, and written notice is promptly given to the prospective client.

Comment [4] says to “limit the initial consultation to only such information as reasonably appears necessary for that purpose.” An agent that invites the whole story before asking who the other side is gathers exactly the information that makes screening harder to claim.

Do not overstate it. Comment [6] says the firm is not barred unless it received information “that could be significantly harmful if used in the matter”, and comment [5] lets a firm make the consultation conditional on the caller’s informed consent that nothing disclosed will stop it acting for a different client in the matter. Script order is cheap insurance for the few calls that come near the line.

A worked script order, for your lawyers to edit (no bar has approved it):

Step	What the agent asks or says	Why it sits here	Stop or route if
1	“I’m an AI assistant for the firm, not a lawyer. I can take some details and book you with one of our lawyers.”	Disclaimer before any information is invited (1.18 cmt [2]; Fla. Op. 24-1)	Caller asks for a person: transfer
2	Name and best contact number	Needed for any outcome	None
3	“Is another lawyer already acting for you on this?”	Fla. Op. 24-1 screening suggestion	Yes: take contact details only and route to a lawyer
4	Names of the other people or businesses involved	Conflict data, gathered before the story (1.18 cmt [3]–[4])	Conflict flag: stop taking facts and route to a lawyer
5	Matter type, from a fixed list	Routing	Not a practice area the firm covers: give the firm’s approved referral wording
6	Date of the event, as a fact	A lawyer needs it to assess deadlines	Caller asks “am I too late?”: log it and route to a lawyer
7	One or two sentences on what happened	Enough to book the right lawyer, no more (1.18(d)(2))	Caller volunteers detail: the agent does not prompt for more
8	Book the consultation	The outcome the call exists for	None

Conflict questions come at step 4, before the story at step 7, and every “what does this mean for me” question has a fixed route to a lawyer.

When does an AI intake call create a prospective client?

Comment [2] to Rule 1.18 says a consultation “is likely to have occurred if a lawyer, either in person or through the lawyer’s advertising in any medium, specifically requests or invites the submission of information about a potential representation without clear and reasonably understandable warnings and cautionary statements that limit the lawyer’s obligations, and a person provides information in response.” On that wording, an agent that opens with “Tell me about your case” is inviting exactly that. A caller responding to advertising that “merely describes the lawyer’s education, experience, areas of practice, and contact information” is not a prospective client.

Florida’s opinion warns that a prospective client or even lawyer-client relationship can be created “without the lawyer’s knowledge”, and against “an overly welcoming generative AI chatbot that may provide legal advice, fail to immediately identify itself as a chatbot, or fail to include clear and reasonably understandable disclaimers limiting the lawyer’s obligations.” The fix is step 1 of the script: say what the agent is before asking for anything.

Are AI intake call recordings confidential, and are they privileged?

They are different questions. Model Rule 1.6 comment [3] separates “the attorney-client privilege, the work product doctrine and the rule of confidentiality established in professional ethics”: privilege applies when evidence is compelled; the ethics rule covers “all information relating to the representation, whatever its source.”

Confidentiality (US). Rule 1.6(a) bars revealing that information unless the client gives informed consent, disclosure is impliedly authorised to carry out the representation, or one of the seven exceptions in 1.6(b) applies. Rule 1.6(c) requires reasonable efforts to prevent its inadvertent or unauthorised disclosure, or unauthorised access to it. Opinion 512 confirms Rules 1.9(c) and 1.18(b) “require lawyers to extend similar protections to former and prospective clients’ information”, so a recording of a caller the firm never signs is covered.

Privilege (Australia). The Evidence Act 1995 (Cth) applies to proceedings in federal courts (s 4); state and territory courts apply their own evidence law, so check which applies to you. Under its s 118, evidence is not to be adduced if, on a client’s objection, the court finds it would disclose a confidential communication between client and lawyer, between two or more lawyers acting for the client, or the contents of a confidential document, made or prepared “for the dominant purpose of the lawyer … providing legal advice to the client”. Section 117 makes a communication confidential if either party “was under an express or implied obligation not to disclose its contents”, and its definition of client is non-exhaustive (“client includes”). Whether a non-engaging caller is a “client”, and whether a booking call has a dominant purpose of legal advice, are questions for your counsel.

The vendor questions. Opinion 512 applies earlier outsourcing opinions to AI providers: check references and credentials, security policies, hiring practices, confidentiality agreements, the vendor’s conflicts check system and a forum to enforce the contract. Then four cloud checks: confidentiality and security that are configured in and enforceable, with notice of any breach or service of process; reliability, security measures, policies and liability limits; retention before and after the service ends, or claimed proprietary rights in the data; and server failure or attack. Rule 5.3 comment [3] names “using an Internet-based service to store client information” as needing this care.

Recording also needs consent under telecommunications law. In the US that varies by state; see two-party consent states and AI call recording. For Australia, see call recording laws in Australia for AI agents.

What ABA Formal Opinion 512 says, and what it does not

ABA Formal Opinion 512, “Generative Artificial Intelligence Tools”, issued on 29 July 2024, covers competence, confidentiality, client communication, meritorious claims, candour to the tribunal, supervision and fees. It is not an intake opinion: the word “intake” does not appear in its text, and footnote 4 points to the Florida opinion for chatbots. What matters for intake:

• Consent before self-learning tools. “A client’s informed consent is required prior to inputting information relating to the representation into such a GAI tool”, meaning self-learning tools whose output could lead to disclosure. Footnote 34 ties this to the risks “as of the publication of this opinion”.
• Boilerplate is not consent. General provisions added to engagement letters are “not sufficient”.

It also cuts against a blanket reading: “there are uses of self-learning GAI tools in connection with a legal representation when client informed consent is not required because the lawyer will not be inputting information relating to the representation.” It does not say every AI intake call needs written consent; it says a tool that learns from callers makes the analysis harder. Ask your vendor in writing whether it trains on your callers’ data.

Can an AI agent call leads back? Solicitation rules on outbound follow-up

Usually, yes. Model Rule 7.3(a) defines solicitation as a communication by or for a lawyer “directed to a specific person the lawyer knows or reasonably should know needs legal services in a particular matter” that offers services for that matter. Comment [1] excludes one made “in response to a request for information”. Calling back someone who filled in your form is a response.

Paragraph (b) bars soliciting “by live person-to-person contact when a significant motive for the lawyer’s doing so is the lawyer’s or law firm’s pecuniary gain”, unless the person is a lawyer, has a family, close personal, or prior business or professional relationship with the lawyer or firm, or routinely uses that type of legal service for business purposes. Paragraph (c) bars solicitation even then where the person “has made known to the lawyer a desire not to be solicited” or it “involves coercion, duress or harassment.” Paragraphs (d) and (e) carve out communications authorised by law or court order, and some prepaid or group legal plans.

Is an AI voice “live person-to-person contact”? The rule does not mention AI. Comment [2] covers “live telephone and other real-time visual or auditory person-to-person communications where the person is subject to a direct personal encounter without time for reflection”, and excludes “chat rooms, text messages or other written communications that recipients may easily disregard.” Our view, not a bar’s: a real-time AI voice call creates that encounter, so treat it as covered. Comment [6] adds that live contact with especially vulnerable people, such as the elderly, those whose first language is not English, or the disabled, “is ordinarily not appropriate”.

The ASCR has no rule worded like 7.3. Its closest, r 34.2, bars a solicitor, in conducting or promoting a practice, from seeking instructions “in a manner likely to oppress or harass” a person who, because of “recent trauma or injury, or other circumstances”, is or might reasonably be expected to be at a significant disadvantage in dealing with the solicitor at the time when the instructions are sought. Rule 36.1 bars advertising that is false, misleading or deceptive (or likely to be), offensive, or prohibited by law. Outbound AI calling also sits under telemarketing and spam law; see whether AI cold calling is legal in Australia. We have not checked state legislation on legal advertising.

Australian law firms: what is different

• No prospective-client rule in the ASCR text. Rule 9.1 protects information “confidential to a client and acquired by the solicitor during the client’s engagement”, and the glossary defines a client as a person “for whom the solicitor is engaged to provide legal services for a matter.” There is no counterpart to Model Rule 1.18. Ask your law society how a non-engaging caller’s information is protected; treat it as if r 9 applied.
• Conflicts. Under r 10.2, a practice holding a former client’s confidential information that might reasonably be concluded to be material to another client’s matter and detrimental to the former client if disclosed must not act “UNLESS” the former client gives informed consent or “an effective information barrier has been established.” Rule 11 covers current clients and is stricter: r 11.3 requires each client to be aware of the other engagement and to give informed consent, subject always to each solicitor’s duty to act in their client’s best interests; and where the practice holds information confidential to one client that might reasonably be concluded to be material to another client’s matter and detrimental to the first client if disclosed, r 11.4 lets the practice act or continue to act only if each client’s informed consent either permits that information to be disclosed and used for the other client, or requires an effective information barrier maintained at all times. Under r 11 a barrier is never an alternative to consent.
• Supervision. Rule 37.1 requires reasonable supervision “over solicitors and all other employees engaged in the provision of the legal services for that matter.” It does not mention AI.
• The regulators’ AI statement. The Law Society of NSW, the Legal Practice Board of WA and the Victorian Legal Services Board and Commissioner issued a joint Statement on the use of artificial intelligence in Australian legal practice on 6 December 2024. Lawyers “cannot safely enter confidential, sensitive or privileged client information into public AI chatbots/copilots (like ChatGPT), or any other public tools”, and must “carefully review contractual terms” of commercial tools. LLM-based tools “cannot reason, understand, or advise”. It does not mention intake, but it recommends limiting AI to “tasks which are lower-risk and easier to verify”. On our reading, collecting names and dates is; advice is not.

Issue	United States (ABA Model Rules template)	Australia (ASCR; Evidence Act 1995 (Cth))
Information from a caller who never signs	MR 1.18(b): may not use or reveal, except as MR 1.9 permits	No ASCR prospective-client rule; r 9.1 is framed around a client’s engagement
Disqualification from intake information	MR 1.18(c): only if “significantly harmful”; (d) consent or screening	ASCR r 10.2: consent or effective information barrier (former clients)
Supervising the agent	MR 5.3(a)–(b); Op. 512 applies vendor checks to AI providers	ASCR r 37.1; joint statement recommends “clear, risk-based policies” on AI
AI telling callers it is AI	Fla. Op. 24-1 (advisory): a lawyer “must inform” callers they are talking to an AI program; separate state AI-disclosure laws may also apply (see the US state AI call-disclosure patchwork)	The ASCR text does not mention AI; joint statement recommends being transparent with clients about AI use; r 36.1 bars misleading promotion
Outbound follow-up	MR 7.3: response to enquiry is not solicitation; live solicitation for pecuniary gain barred unless one of three exceptions applies	ASCR r 34.2: no oppressive or harassing approach to disadvantaged people
Privilege over recordings	Jurisdiction’s evidence law (MR 1.6 cmt [3])	Evidence Act 1995 (Cth) ss 117–118 (federal courts; state courts apply their own evidence law): dominant-purpose test

AI legal intake vendors compared, on one rule

Inclusion rule: a vendor meets the rule only if its own website, read on 10 October 2026, says an AI agent (not only a human receptionist) holds a conversation with a law firm’s prospective clients by phone or voice, and does not mark that agent as not yet available. Every cell comes from the vendor’s own page. Smith.ai and CaseGen meet the rule today. Lawmatics does not yet: its Merlin Engage page describes a “Voice Agent” that “Reaches out by phone”, but its AI page marks Phone Agents “Coming soon” as at 10 October 2026 (although that page’s FAQ lists “answering after-hours calls” among its agents’ work), so its row is shown for comparison and its voice intake should be treated as announced, not available. Ruby is shown as the human-answered contrast. These are the three we compared, not a sweep. Clio’s pages returned a bot-check page to us, so Clio is not assessed; LEAP and Smokeball’s Australian home pages describe AI research and drafting assistants, not phone intake.

Vendor	Product, in its words	Channels stated	Hand-off to a human, in its words	Conflict checking, as described
Smith.ai	“Your 24/7 AI Receptionist for the whole intake journey”; “24/7 AI Receptionists + Live Agents”	Phone, including outbound follow-up calls (“we don’t do cold calling”); “Text & email follow-up” on the legal page; web chat (“Live-staffed website chat powered by AI”, where “Our AI bot will transfer to one of our live agents”); AI in “English and Spanish”	Offers “AI-first”, Hybrid and “Human-first” options (Hybrid: “AI handles certain call types, and Smith.ai Live Agents handle the rest”); legal page says “Our AI and live agents deliver” intake	“Conflict checks” listed among legal features; its help centre describes a Conflict Checker that searches a firm’s Clio, Lawmatics or Google Sheets contacts and is “an add-on to our phone receptionists plans”; neither page says whether the AI or a live agent runs it
CaseGen	“AI Powered Intake & Receptionist”; agents “built to sound human”	Inbound phone; its intake agent “can even call leads directly to complete the intake”; its Clio integration page says a lead may connect “through web chat, text, or phone”	CaseGen “flags high-value leads in real time and alerts you for live transfer”	Its AI intake “runs conflict screening” and is “equipped to: Verify conflicts (spouse or opposing party check)”; CaseGen says attorneys “retain responsibility for legal advice, conflict review and accepting a client”
Lawmatics (Merlin Engage; phone agents announced, outside the rule today)	Handles “lead outreach and conversation on your firm’s behalf”; “requires Merlin Qualify”; its agent lineup has “some agents now live and new ones rolling out to the Lawmatics platform soon”	Engage page: “SMS, email, voice, and web chat”; its AI page marks Phone Agents, Email Agents and Website Chat Agents “Coming soon” (Text Agents carries no such label), as at 10 October 2026	“Escalates to a firm member whenever a situation calls for human judgment”; the firm controls “when a human reviews first”	Separate CRM feature: automations “can also be configured to approve, deny, or flag new clients as a possible conflict”
Ruby (outside the rule)	Live receptionists; AI “supports receptionists behind the scenes”	Phone and chat, answered by people	A person answers every call	Not assessed

Using the table: CaseGen’s agents are “built to sound human”, and its subscription agreement says the provider, MAF AI, INC, “will clearly identify itself as an AI assistant working on behalf of Subscriber when interacting with third parties”. CaseGen’s blog post on disclosure says “transparency is the right direction”, tells firms to “always check with your state law disclosure requirements first”, and suggests a script that introduces “the AI as a smart assistant working directly with the firm”, though the sample opening in its body starts with purpose rather than an AI disclosure. Check that the opening line your callers hear says it is AI before any details are taken. Set Lawmatics’ conflict automations to flag rather than approve, because 1.18 comment [4] and ASCR r 10.2 leave the decision with a lawyer. Ask Smith.ai which conflict-check steps its AI performs.

Who each option is wrong for

• A legal-specific AI intake vendor is wrong for a firm that has not written its own hand-off rules; only your lawyers can say what counts as advice.
• A human-answered service with AI assistance, such as Ruby, is wrong if you want an AI to complete intake, and right where callers are often in distress.
• A general-purpose AI voice platform is wrong if you want conflict checks and practice-management integration out of the box; it fits follow-up on enquiries already in.
• No AI at all is right while someone trained can answer every call the same day. No rule requires automation.

Frequently asked questions

Where every figure on this page comes from

Figure	Who published it	Link	Date read
Opinion 24-1 dated 19 January 2024; three conditions for nonlawyer initial interviews (from Op. 88-6)	The Florida Bar	floridabar.org/etopinions/opinion-24-1 (text read from the Internet Archive copy; live page body loads by script)	10 October 2026
Formal Opinion 512 dated 29 July 2024; six vendor checks and four cloud checks	American Bar Association Standing Committee on Ethics and Professional Responsibility	americanbar.org (PDF) (read from the Internet Archive copy; direct fetch blocked)	10 October 2026
Model Rule 1.18(a)–(d) and comments [2]–[6]	American Bar Association	Rule 1.18 (Internet Archive copy)	10 October 2026
Model Rule 1.6(a)–(c), seven exceptions in 1.6(b), comment [3]	American Bar Association	Rule 1.6 (Internet Archive copy)	10 October 2026
Model Rule 5.3(a)–(c) and comment [3]; Rule 5.5 comment [2]	American Bar Association	Rule 5.3 (Internet Archive copy)	10 October 2026
Model Rule 7.3(a)–(e), three exceptions in 7.3(b), comments [1], [2], [6]	American Bar Association	Rule 7.3 (Internet Archive copy)	10 October 2026
ASCR rules 9.1, 9.2 (six exceptions), 10.2, 11, 34.2, 36.1, 37.1 and the glossary definition of “client”	NSW Parliamentary Counsel’s Office (Legal Profession Uniform Law Australian Solicitors’ Conduct Rules 2015)	legislation.nsw.gov.au (Internet Archive copy showing amendments to 2024 No 47)	10 October 2026
Joint AI statement dated 6 December 2024; three issuing bodies	Victorian Legal Services Board and Commissioner (with the Law Society of NSW and the Legal Practice Board of WA)	lsbc.vic.gov.au	10 October 2026
Evidence Act 1995 (Cth) ss 117–118 (Compilation No. 35, 10 June 2025)	Federal Register of Legislation	legislation.gov.au	10 October 2026
Smith.ai: “24/7”, “English and Spanish”, AI-first, Hybrid and Human-first options, outbound follow-up but no cold calling, AI-plus-live web chat, “Conflict checks” feature and Conflict Checker add-on	Smith.ai	legal page; AI receptionist; plans; Hybrid; Conflict Checker help article; web chat	10 October 2026
CaseGen: “24/7”, live transfer, outbound intake calls, web chat and text (Clio integration page), conflict screening, attorneys’ retained responsibility, AI self-identification clause, disclosure blog post	CaseGen	casegen.ai; legal answering service; industries; legal intake; subscription agreement; Clio integration; disclosure blog post	10 October 2026
Lawmatics Merlin Engage: four channels, escalation, requires Merlin Qualify, agents “rolling out”; Phone, Email and Website Chat Agents “Coming soon”; conflict automation	Lawmatics	Merlin Engage; Lawmatics AI page; conflict check software	10 October 2026
Ruby: live receptionists with AI behind the scenes	Ruby	ruby.com/ai; ruby.com/industries/legal	10 October 2026
Zian: 30+ languages, channels, integrations, private deployment, human-in-the-loop approvals	Zian AI	Zian’s home page and Features + Integrations page, and Zian’s published brand facts	10 October 2026