Top 10 AI Receptionists for Law Firms 2026:#1 Frenchy Digital

It's 9:40 on a Saturday night and someone has just been served with a temporary restraining order. They call three family law firms from the parking lot. Two go to voicemail. The third answers, asks their name and the other party's name, and tells them an attorney will call first thing Monday. Which firm do you think gets the consult?

That's the pitch every AI receptionist for law firms makes, and most of it holds up. What doesn't hold up is how these products get compared. The ranking pages I found lean on answer rates, "accuracy" and conversion figures the vendors publish about themselves.

So on September 30, 2026 I did the slow version. I read the law pages and pricing pages of every legal AI receptionist I could verify, wrote down only what each page actually says, and scored ten of them on six things a managing partner can re-check in about fifteen minutes per vendor.

A quick word on what this piece isn't. Our earlier law firm AI receptionist checklist covers the ten things to check before you buy any of them. This one names names and puts a score on each.

The mental model most firms bring to this is that an AI receptionist is a smarter voicemail. It picks up, takes a message, maybe books a slot. Judge it on price and how human it sounds.

That model works for a plumber. It breaks for a law firm, bc the first two minutes of a legal intake call carry professional obligations that a plumber's call never does.

Here's the one that bites. Under ABA Model Rule 1.18, a person who consults you about possibly hiring you is a prospective client, and what they tell you is protected even if you never take the case. Worse, if they tell you something "significantly harmful" and you already represent the other side (or later could), you can be disqualified from the matter, and that can spread to the whole firm.

Think of it like a restaurant host who seats two people at the same table before checking whether they're divorcing each other. The seating is quick. The consequences aren't.

So the better model is this: a legal AI receptionist is an intake clerk with a script, and the order of that script matters more than the voice. Names of every party first, conflict check second, the story third. An agent that lets a distressed caller pour out the whole narrative before asking who the other side is has created a risk, no matter how natural it sounds.

Therefore I scored the things that decide whether an AI receptionist can run that script safely: does it do legal intake at all, does it write into your practice management system, does it support a conflict check, does it publish its confidentiality terms, can you see the price, and does a person pick up when it can't cope.

Why bother answering at all hours? The evidence is thinner than the marketing suggests. Clio says its Legal Trends research found that 48% of firms don't respond to phone inquiries, and Clio sells an intake agent, so read that with the seller in mind. Martindale-Avvo, which sells legal marketing, reports that only 11% of legal consumers in its survey hired the first attorney they contacted. That second number cuts against the "first to answer wins" folklore. Answering fast still matters. It just isn't the whole contest; what happens after the answer is.

Six attributes, 0 to 2 points each, 12 points maximum, taken only from each vendor's own public pages (plus Clio's App Directory) on September 30, 2026.Here's the rubric in full, before the results, so you can disagree with it first.

Ties go to the lower published entry price, then alphabetical order. It's blunt, and I'd rather use a blunt rule you can see than a clever one you can't.

Judgment calls, disclosed.Smith.ai's conflict workflow is documented in a 2020 post updated in 2022 and conflict checks are listed on its legal page, which sells AI and human receptionists together. Posh and Abby Connect are hybrid human and AI services, scored on the whole offering. My AI Front Desk prices overage in credits without a dollar value, and its law page says $109 while its pricing page says $99, so it gets 1 on pricing. Clio Grow AI is Clio's own product, so "Clio plus another system" can't apply, and it gets 1 there.

Who didn't make the list. Ruby is in the Clio App Directory, but its AI is described as an enhancement layer for human receptionists, not an AI answering product. Lawmatics' Merlin Engage voice agent was marked "coming soon" on the day. LEX Reception and Answering Legal are human intake services. Goodcall, Rosie and other general small-business tools are covered in our ranking of the best AI receptionists overall, and I couldn't confirm a law-specific product page for them in time.

How to re-check.Every cell in the evidence table below paraphrases what the vendor's page said on September 30, 2026. Compliance lines are the vendor's own claims; I inspected no audit report, certificate or signed agreement for any of them. Pricing pages change without notice, so screenshot the page the day you sign.

Three vendors tied at 11 of 12, and Smith.ai takes first on the tie-break bc its free tier is the lowest entry price. Here are the scores.

Scores reflect the vendor's public pages checked September 30, 2026. A zero means the attribute wasn't disclosed on the pages we checked, not that the vendor can't provide it.

Three things jumped out when I laid this out.

First, the conflict column is where this category separates. Only three of the ten (Smith.ai, My AI Front Desk and Clio Grow AI) document anything I'd call a conflict-check workflow. Four collect party names for you to check. Three show nothing or say outright they won't do it. If you run a practice where conflicts are frequent (family law in a small town, say, or insurance defense next to plaintiff work), that column matters more than the other five combined.

Second, Clio integration is nearly universal. Nine of the ten name Clio somewhere on their own pages. That's great if you're on Clio, and it means "integrates with Clio" is no longer a reason to pick anyone. Ask how: native, through Zapier, or by email to an inbox someone reads on Monday.

Third, human escalation is also close to universal. Eight of ten document a live or warm transfer. That's the right default for law, where some callers are in crisis, and it's the one thing I'd refuse to buy without.

Each cell is what the vendor's page said on September 30, 2026, or "not publicly disclosed".Lines marked search-confirmed come from search results for the vendor's own pages that I couldn't load fully.

Sources: the vendor pages listed in the sources section. Clio pages block automated fetches, so they were read in a real browser. The Clio Grow AI per-matter price is as reported by LawSites.

Read the pricing column for units before you read it for dollars. Upfirst and Smith.ai bill per call. Dialzara and Posh bill per minute. NextPhone and LegalClerk.ai sell unlimited calls (LegalClerk.ai per seat). Abby Connect bills in its own minutes, with an AI minute counted as half. Clio Grow AI bills per hired matter. Those are six different bets on what your phone traffic looks like.

Here's each vendor in rank order: what earned the score, what cost it points, and who it suits.Every recommendation comes with its downside, because that's the part the vendor page leaves out.

Top ranked

1

Smith.ai

11 out of 12

Smith.ai is the only vendor here with a public, step-by-step description of receptionists running conflict checks. Its conflict-check post describes asking callers for opposing party names and searching the firm's Clio, Lawmatics or Google Sheet records. It was published in 2020 and updated in 2022, and it also disclaims responsibility for misspellings, which is a fair and useful admission.

The AI Receptionist pricing page is the clearest in the set: a Free tier at $0 with 25 calls and $3 per extra call, Pro from $150 a month at roughly $2 per call, and Enterprise from $500. Native integrations with Clio, MyCase and Lawmatics are named on its law firm page.

The point it dropped is confidentiality terms. Its legal page says it doesn't share your lead or client information with other customers, and also states that Smith.ai is not HIPAA-compliant. I found no SOC 2 or BAA statement on smith.ai. One warning if you search for it: a different company with a similar name publishes SOC 2 and ISO claims. Don't mix them up.

Who it suits: small and mid-size firms that want AI and human receptionists under one roof and a documented conflict process. The downside: per-call pricing adds up on busy lines, and the conflict workflow is documented for the human receptionist service, so confirm in writing whether the AI tier does the same.

Visit Smith.ai

2

Upfirst

11 out of 12

Upfirst's law firm page is written for the way firms actually take calls. You create intake scripts for different case types, it screens callers to confirm they're in your practice area, and it captures the opposing party's name for a conflict check. Clio Manage is native; MyCase, Lawmatics and Clio Grow connect through Zapier. It's also in the Clio App Directory.

Its warm transfer is the best-described handoff in the set: the call comes to you with context, so you see who's calling and why before you accept. Pricing on the Upfirst pricing page runs from Starter at $24.95 for 30 calls ($1.50 per extra call) to Scale at $299 for 600 calls ($0.70 extra), with a 14-day trial. BAAs are listed on the Custom plan, and a Data Processing Addendum is linked.

The point it dropped: it collects the opposing party's name but doesn't document checking it against your records. Your staff still run the check.

Who it suits: solo and small firms on Clio that want to start cheap and trial it properly. The downside: the BAA sits on the Custom plan, not the self-serve tiers, and per-call billing charges the same for a 20-second wrong number as for a 12-minute intake.

Visit Upfirst

3

My AI Front Desk

11 out of 12

My AI Front Desk (the site now brands itself Frontdesk) has the most detailed law pages in the set. Its Clio law page says the agent captures matter type, statute concerns, opposing parties and insurance, runs pre-built workflows for personal injury, family, criminal, estate and immigration work, and names Clio, MyCase, Lawmatics, Litify and Filevine.

It also says it surfaces opposing parties and runs an initial conflict check against your matter management system, states SOC 2 Type II, and says BAAs are available. Those are the vendor's claims; I saw no report. Emergency keywords such as an arrest or a deadline route to on-call attorneys.

The point it dropped is pricing clarity. The pricing page shows $99 a month ($79 billed annually) with 200 voice minutes, but overage is priced in credits (25 per minute) with no dollar value for a credit, and the law page quotes $109 flat. I'm using the pricing page figure, dated today.

Who it suits: firms that want the deepest legal feature list at a low monthly price. The downside:you can't forecast a busy month in dollars from the public page, and an "initial conflict check" is only as good as the data it checks against, so test it with a known conflict before you rely on it.

Visit My AI Front Desk

4

Posh

10 out of 12

Posh is a hybrid: AI handles qualifying leads and streamlining intake, and its attorney answering page says sensitive intake calls transfer to its human virtual receptionists. It names Rocket Matter, Clio, Calendly, Acuity and Zapier, and says it supports your internal conflict process by organizing and relaying client details.

Pricing effective January 1, 2026 on the Posh pricing page runs across seven tiers, from $65 a month with no included minutes at $2.30 a minute to $1,900 for 1,000 minutes at $1.90. The AI plan isn't priced separately.

Who it suits: firms whose callers are often upset and need a person, with AI as the first layer. The downside: per-minute human rates make long intake calls expensive, and the legal page carries a confidentiality statement but no SOC 2 or BAA.

Visit Posh

5

NextPhone

10 out of 12

NextPhone has the most specific practice-area page I read: its family law page lists intake scripts for divorce, custody, TROs, support, prenups and adoption, collects jurisdiction and opposing party details, and routes domestic violence and TRO emergencies to on-call attorneys. Clio Grow and Manage are native; MyCase, Lawmatics, PracticePanther and Smokeball connect via Zapier.

Pricing is the simplest in the set. The NextPhone pricing page lists Pro at $199, Growth at $299 (with setup done by NextPhone's team) and Scale at $599, with unlimited calls and no overage fees. Pro and Growth allow up to 10 transfer recipients.

On confidentiality, search results show NextPhone claiming SOC 2 Type II and stating in its terms that it does not execute BAAs. I couldn't load those pages fully, so treat both as unconfirmed. It collects conflict fields for attorney review rather than checking them.

Who it suits: busy family and PI practices with lots of calls, where unlimited pricing wins. The downside: no BAA, so keep health-heavy matters off it, and ignore its self-reported resolution rates until your own trial confirms them.

Visit NextPhone

6

Clio Grow AI

9 out of 12

4.8/ on App Store (Clio for Law Firms app) (8,993 reviews)

Clio launched Grow AI on August 20, 2026. According to Clio's announcement, its agents capture contact and case details, assess leads on case type, location and conflicts, and check each lead against the firm's own lead scoring and conflict rules. That's a documented conflict workflow, from the company that already holds your matters. Phone agents are currently US only.

The pricing is the most interesting thing here. Per LawSites' launch coverage, it charges about $25 per hired matter, with a monthly spending target, and the first three hired matters free for existing Grow customers. You pay when a lead becomes a client, not when the phone rings.

It lost points in three places. It only flows into Clio (fine if you're a Clio shop, a dealbreaker if you're not). Pricing needs a Clio Grow subscription on top, which the per-matter fee doesn't show. And stalled leads get flagged for a person to close out rather than transferred live.

Who it suits: firms already on Clio Grow, especially lower-volume, higher-value practices where paying per hire beats paying per call. The downside:no live transfer is documented, and the conversion claims in the launch are Clio's own data about a product Clio sells.

Visit Clio Grow AI App Store

7

Abby Connect

9 out of 12

Abby Connect's legal page describes AI intake following firm-specific scripts and qualifying leads against your criteria, across personal injury, family, criminal, estate, immigration and corporate work. It names the widest set of practice management systems in the list: Clio, Clio Grow, Lawmatics, MyCase, PracticePanther and Smokeball. Complex matters escalate to human staff.

On the Abby pricing page, Starter is $165 for 50 Abby Minutes up to Growth at $1,380 for 500, an AI minute counts as half an Abby Minute, and overage bills at your plan's rate. It offers HIPAA compliance options.

Who it suits: firms on a less common practice management system that want human backup. The downside: it scored zero on conflicts, as I found nothing about conflict checks on its public pages, and its entry price is several times the AI-only tools.

Visit Abby Connect

8

LegalClerk.ai

9 out of 12

LegalClerk.ai is built only for law firms, and its personal injury screening is the most specific: accident type detection, injury severity classification and liability filtering, with family, criminal and immigration flows too. It names Clio, MyCase, Google Calendar and Calendly, and offers warm transfers and priority escalation.

Pricing is $400 per seat per month for unlimited calls, no setup fee, with a 7-day trial. It describes HIPAA-grade encryption, AES-256 and audit logging, but states no SOC 2 attestation or BAA.

Who it suits: PI firms that want screening logic tuned to their case criteria. The downside: no conflict-check mention at all, and per-seat pricing means the bill grows with your team, not your calls.

Visit LegalClerk.ai

9

Lexidesk

8 out of 12

Lexidesk qualifies personal injury, family, immigration and other matters and scores each lead from 1 to 10 on the firm's criteria. It connects to Clio Grow and Manage, Lawmatics, Actionstep, HubSpot and HighLevel, is in the Clio App Directory, and live-transfers urgent matters to an attorney.

It scored zero on conflicts for a reason I respect. Its LawNext directory listing says it doesn't automate conflict checks, as a deliberate choice to avoid unnecessary data exposure. That's a design position, not a gap, and some firms will prefer it. My rubric still scores what's documented, so it takes the zero.

The security language says "aligned" with ISO 27001 and SOC 2 controls, which isn't an attestation, and pricing starts at $700 a month on its Clio listing with a quote otherwise.

Who it suits: firms that want lead scoring and a vendor that keeps its hands off conflict data. The downside: the highest published entry price in the list, and a young company (it describes itself as founded in 2025 and bootstrapped).

Visit Lexidesk

10

Dialzara

6 out of 12

Dialzara is a general small-business AI receptionist with a law page. Per its for-lawyers page, it collects caller-provided intake details, routes callers by matter categories you define, and attempts a transfer under your priority rules when an approved recipient is available. It says the law firm must run and evaluate its own conflict check.

Pricing on the Dialzara pricing page is fully published: Lite $29 for 60 minutes ($0.48 over), Pro $99 for 220 ($0.45), Plus $199 for 500 ($0.40), Elite $349 for 1,000 ($0.35), with a 7-day trial.

Who it suits: a solo practitioner who mainly needs messages taken and urgent calls transferred, cheaply. The downside:no practice management system named on its law pages, no practice-area intake, and no conflict support. It's a receptionist, not an intake clerk.

Visit Dialzara

Frenchy Digital is my company. That's a conflict of interest, which is why this pick sits outside the scored table and is labeled as mine.

So let me score it the same way first, and not generously.

• Legal intake: 1: We write the intake script per engagement, practice area by practice area, but our receptionist product page doesn't document legal practice-area intake as a feature. On a public-evidence rubric, that's a 1.
• Practice management integrations: 1: It's a custom build, so it connects to any system with an API, and Clio publishes a developer platform for Clio Manage and Clio Grow. But we have no prebuilt, named integration and we're not a Clio App Directory partner. That's a 1.
• Conflict-check support: 1: We can design the agent to collect every party name before the narrative and write them to your system for your own check. That's a design pattern, not a shipped automated conflict check, so it's a 1.
• Confidentiality terms: 1: We sign a BAA with every healthcare client, which proves little for a law firm. We hold no SOC 2, ISO 27001 or HITRUST. Full source code and IP transfer gives you control of the data path, but it isn't a published data term. That's a 1.
• Published pricing: 2: $5,000 for the agent plus $5,000 setup, $10,000 initial, rising with complexity and the number of integrations. You pay model, telephony and API usage directly at provider rates. After launch, $2,500 a month. It's published on our receptionist page.
• Human escalation: 2: The receptionist identifies itself as AI, hands over when it doesn't know, escalates anything outside its rules by default, and passes the conversation so far to the person who picks up.

That's 8 of 12. It ties Lexidesk, and on the entry-price tie-break ($700 against $10,000) Lexidesk wins. On this rubric we would place 10th of 11, above only Dialzara. I'm printing that plainly bc hiding it would undo every other cell in this article.

So why is it still my pick? Because the rubric measures what a vendor publishes, and the thing that matters most to a firm with a specific intake process isn't publishable: whether the agent follows your script, in your conflict order, with your escalation rules, across your stack.

An off-the-shelf tool gives you its intake template and its integration list. A custom agent gives you one agent across the whole stack (phone, web chat, Clio or whatever you run, your calendar, your CRM), built around the order you'd want a trained intake clerk to follow. Names first. Conflict check against your data second. Narrative third. Emergencies to the on-call attorney with the conversation attached.

And you own it. Full source code and IP transfer means the scripts, the escalation logic and the data path are yours, which is the strongest confidentiality control I know of that doesn't involve trusting a vendor's page. You can see what it does. You can change it without a support ticket.

Here's what that looks like as a scenario, not a client story (we haven't shipped a law firm receptionist we can name). A three-attorney family and estate firm on Clio Manage takes intake calls in English and Spanish and keeps a separate list of former adverse parties in a spreadsheet. A custom agent would read both before the caller explains why they're calling, flag a match to the attorney without telling the caller why, and book only into the calendar of the attorney who handles that matter type. None of the ten documents that exact combination.

The price reality, stated out loud. Year one is $10,000 initial plus twelve months at $2,500: $10,000 plus $30,000 is $40,000, plus usage. Later years run $30,000 plus usage. Against NextPhone Pro at $199 a month ($2,388 a year), that's roughly 17 times the cost ($40,000 divided by $2,388). You should only pay that if the fit matters that much. You can see what's included on our AI receptionist page.

No bar has banned AI intake, but several have said how it should behave.This isn't legal advice, and your state's rules decide. Here's what I found, and how it maps onto the ten tools.

Confidentiality (Rule 1.6).The ABA Model Rule is short: a lawyer "shall not reveal information relating to the representation of a client unless the client gives informed consent, the disclosure is impliedly authorized in order to carry out the representation or the disclosure is permitted by paragraph (b)." Your AI receptionist vendor is handling that information. The confidentiality column in the rubric is a proxy for how much the vendor tells you about what it does with it.

Prospective clients (Rule 1.18).This is the one that decides your script order, as covered above. Collect names, check conflicts, then take the story. If a tool can't be configured to do that, it's the wrong tool.

Supervision (Rule 5.3).Managing lawyers must have measures giving reasonable assurance that a nonlawyer assistant's conduct is compatible with the lawyer's obligations. An AI receptionist is supervised like staff. Read its transcripts. Test it with hard calls. Keep a record that you did.

ABA Formal Opinion 512 (July 29, 2024).It applies existing duties to generative AI. Two lines matter for intake: for self-learning tools, "a client's informed consent is required prior to inputting information relating to the representation into such a GAI tool," and "merely adding general, boiler-plate provisions to engagement letters purporting to authorize the lawyer to use GAI is not sufficient." Ask every vendor whether caller data trains any model.

Florida Bar Ethics Opinion 24-1 (January 19, 2024).It addresses intake chatbots directly and says they "must include a disclaimer indicating that the chatbot is an AI program and not a lawyer or employee of the law firm." It's advisory, but it's the clearest statement of the disclosure norm I found.

Alabama State Bar Formal Opinion 2026-01.Its best-practices opinion on AI gives the example of a website chatbot offering specific legal advice without flagging that it isn't an attorney and without attorney review as potential unauthorized practice. An intake agent that starts answering "do I have a case?" is the risk.

State disclosure statutes.California's bot law covers bots used online to mislead people about being artificial; a web chat intake is inside it, and a plain phone call is likely outside it on my reading. Utah requires regulated occupations, lawyers included, to disclose generative AI verbally at the start of high-risk interactions, currently until July 1, 2027 under SB 332.

Outbound callbacks.The FCC ruled in February 2024 that AI-generated voices count as "artificial or prerecorded voice" under the TCPA, so outbound AI calls generally need prior express consent. Rule 7.3 separately limits live person-to-person solicitation. Answering inbound calls is fine; having the agent chase cold leads by phone is a different project with different rules.

Put together, the design that satisfies all of it is boring: say it's AI in the first sentence, never give legal advice, take names before the story, hand off to a person on anything sensitive, and don't dial out without consent.

The cheapest headline is rarely the cheapest bill, and in law it's rarely the right measure anyway.Here's a rough comparison at 300 calls a month, using only published prices.

Let me show the Dialzara line, since per-minute maths is where people get surprised. If an average intake call runs three minutes, 300 calls is 900 minutes. Pro includes 220, so 680 minutes over at $0.45 is $306, plus the $99 base: about $405. Upfirst Pro at $159.95 for the same 300 calls is cheaper, by roughly 2.5x.

But flip the unit of value. A PI firm doesn't earn money per call, it earns it per signed case. If 300 calls produce, say, five signed matters, Clio Grow AI's roughly $25 per hire is about $125 in fees, before the Grow subscription. A practice with few calls and valuable matters might find per-hire pricing the most sensible model in the list; a high-volume practice might not.

For deeper vendor-by-vendor pricing across the whole category, our AI receptionist pricing comparison works the numbers for twelve vendors.

One thing to check that no pricing page shows: what happens to spam and robocalls. On a per-call plan, a spam wave costs money unless the vendor filters it before it counts. Ask.

These are the things that would make me walk away from a legal AI receptionist demo, whoever is selling it.

• The agent takes the story before the names: If the demo script asks what happened before it asks who the other party is, the default flow creates Rule 1.18 exposure. Ask whether you can reorder it, and test that you can.
• It won't say it's AI: An agent that dodges the question, or is configured by default to pass as human, runs against Florida's opinion and Utah's statute and will embarrass you when a caller finds out.
• It answers legal questions: Ask it 'do I have a case?' in the trial. Anything more than 'an attorney will review that' is a problem.
• 'Integrates with Clio' means an email: Ask exactly what lands in Clio: a contact, a matter, a note, custom fields? And in which product, Manage or Grow?
• Accuracy figures without a method: A resolution rate or satisfaction score with no sample, period or definition is marketing. Every vendor in this list publishes at least one.
• Security by adjective: 'Bank-grade', 'HIPAA-grade' or 'aligned with SOC 2' isn't a report. Ask for the SOC 2 report under NDA, or accept that there isn't one.
• Caller data trains the model: Opinion 512 makes this a consent question. Get a written answer on whether recordings or transcripts are used for training.
• No named person on escalation: If the warm transfer goes to a general queue at the vendor rather than to your on-call attorney, a TRO call at midnight may get a message and nothing else.

This ranking measures disclosure, not performance. A vendor that publishes everything could still run an agent that mishears names. A vendor that publishes little could be excellent. The trial is where you find out.

I did not call any of these products as a test caller for this article, and I inspected no SOC 2 report, BAA or data processing agreement. Security and compliance cells are the vendors' own claims.

Some cells rest on search results rather than a full page load: NextPhone's SOC 2 and BAA position, Dialzara's BAA on request, and Clio's SOC 2 and ISO 27001 status. Clio's own pages block automated fetches, so I read them in a browser. Clio Grow AI's $25 per hired matter comes from LawSites' reporting.

Smith.ai's conflict workflow post dates from 2020, updated in 2022. It's still live and linked from its legal offering, but a four-year-old post describes a process, not a promise for today's AI tier.

The scope is ten vendors. Legal intake tools launch every month, and some I screened out (Lawmatics Merlin Engage, for one) could qualify within weeks. I also scored hybrid services and AI-only tools on one rubric, which flatters hybrids on escalation.

Finally, the ethics section is a map, not advice. I'm not a lawyer. Your state bar's rules and opinions govern.

You'll see these in other articles about legal AI receptionists. They're not here, and here's why.

The claim that a large share of legal clients hire the first firm that responds: I found no primary source. The nearest real survey I found, from Martindale-Avvo, points the other way, with most consumers contacting several firms.

A widely repeated survey saying callers want a real person: it comes from a human answering service's marketing, with no methodology I could find.

"AI costs a tenth of a human service": a marketing line with no source. Our AI versus human receptionist cost breakdown does that comparison properly.

Every vendor-reported resolution, sentiment, satisfaction, time-saved or conversion figure. They might be right. They're also the seller grading its own homework.

You can get from this article to a decision in about a week.Here's the order I'd do it in.

1. 1.Write your intake order on one page: which names to collect, where your conflict data lives (Clio, a spreadsheet, both), what counts as an emergency, and who is on call. That page is your test script and your filter.
2. 2.Pick two or three vendors from the scoreboard whose conflict and integration cells match that page, and start trials on the plan you'd actually buy.
3. 3.Call each trial yourself five times: a new PI lead, a caller whose opposing party is an existing client, a crisis caller, someone asking 'do I have a case?', and a spam call. Screenshot the pricing page the day you sign.

If none of them runs your intake the way you'd train a person to, that's when a custom build earns its price. Otherwise, pick the best trial and get back to your cases.