Smith.ai sells calls. Ruby sells minutes. LegalClerk sells seats. Your shortlist is uncomparable until you fix that.
Smith.ai is a legal answering and intake service that publishes its rates, and so, it turns out, do five of its alternatives. Its tiers are $300 a month for 30 calls, $810 for 90 and $2,100 for 300, with a custom Enterprise tier, no setup fee and a 10 percent discount for a 12 month commitment. Divided out that is $10.00, $9.00 and $7.00 per included call, and the number the pricing page does not print is the overage premium, which runs the wrong way: 15 percent above your included rate on Starter, about 17 on Basic and about 21 on Pro, so the tier demanding the largest commitment penalises you hardest for outgrowing it. The crossovers land early too, at 75 calls a month from Starter to Basic and 213 from Basic to Pro. The bigger problem with every comparison you will read is that these vendors do not sell the same unit. Smith.ai bills by the call, Ruby, AnswerConnect, Alert Communications and Dialzara bill by the minute, and LegalClerk.ai bills by the seat with calls unlimited. Normalise them to one firm doing 100 calls a month at five minutes each and the monthly spread runs from $199 to $1,725 for what buyers call the same job, with the AI answered options at the bottom and the human staffed ones at the top. None of the seven vendors we read publishes that its conflict check queries the firm's own existing matter records. For most firms arriving here the honest conclusion is to switch answering vendors rather than commission anything, because coverage is a solved, competitive, published-price problem and against a fixed fee starting at $45,000 the arithmetic is not close. The graduation question arrives later, when the expensive work has moved downstream of the call.
A firm shopping answering vendors is deciding how much of its highest intent contact to rent, and from whom. This page normalises six published price tables to one worked example, using each vendor's own pricing page, read on August 15, 2026.
What Smith.ai is, and the unusual thing it does.
Smith.ai sells call answering and client intake to law firms, and it does something a good deal of its category avoids: it publishes a complete price table. Starter, Basic and Pro are listed at $300, $810 and $2,100 a month for 30, 90 and 300 calls, with overage rates printed next to each and an Enterprise tier held back for quotes. The page states no setup fees, month to month billing, a 30 day money back guarantee, and a 10 percent discount for subscribing for 12 months. Its law firm page states that all calls on its Virtual Receptionist plans are answered by a real person, 24/7, by North America based agents working from prompts rather than scripts, while its pricing page separately offers a choice between an AI first and a human first handling model without publishing a price difference between them.
On integrations it names MyCase, Clio, Lawmatics and Zapier on the law firm page, Actionstep and Assembly Neos on the integrations page, and it publishes a Filevine page describing itself as that platform's preferred partner for answering and intake. Every figure on this page was read on the vendor's own pricing page on August 15, 2026. We have deliberately left out numbers that exist only on comparison blogs, because those go stale quietly and we cannot audit them.
We are not neutral and it would be silly to pretend otherwise, so the bias is declared up front. We build custom systems for mid market firms starting at $45,000, which makes us the wrong purchase for most people who arrive at this page. The rest of it is written accordingly.
The per-call rate its pricing page does not print.
A price table sold in calls invites one division the vendor does not do for you, and it changes how the tiers read.
| Tier | Monthly | Calls included | Per included call | Overage per call | Overage premium |
|---|---|---|---|---|---|
| Starter | $300 | 30 | $10.00 | $11.50 | 15 percent |
| Basic | $810 | 90 | $9.00 | $10.50 | about 17 percent |
| Pro | $2,100 | 300 | $7.00 | $8.50 | about 21 percent |
| Enterprise | Not published | Custom | Not published | Not published | Not published |
The volume discount is real: committing to Pro takes your included rate from $10.00 down to $7.00 a call. The part worth pausing on is the last column. The proportional penalty for exceeding your plan is largest on the tier that asks for the biggest commitment. Most buyers size a plan assuming the opposite, that a bigger commitment buys a softer landing when a busy month arrives.
The tier crossovers also land earlier than the included call counts suggest. Staying on Starter costs $300 plus $11.50 for each call past 30, so at 74 calls you pay $806 and at 75 you pay $817.50 against Basic's flat $810. Basic wins from 75 calls, well under its 90 call allowance. Staying on Basic costs $810 plus $10.50 for each call past 90, so at 212 calls you pay $2,091 and at 213 you pay $2,101.50 against Pro's $2,100. Pro wins from 213 calls, against a 300 call allowance. A firm whose volume is drifting upward is therefore usually overpaying for several months before the invoice is annoying enough to re-read.
Extend that to a year and the shape of the decision changes again. Pro at exactly 300 calls a month is $25,200 a year before the 12 month discount. At 400 calls it is $35,400, at 500 calls $45,600, and at 600 calls $55,800. One caveat has to be stated plainly or the arithmetic becomes a straw man: Smith.ai publishes an Enterprise tier precisely for this situation, and a firm at those volumes should be negotiating it rather than paying list overage. The table above is what you pay if you never ask.
Three vendors, three different units, one uncomparable shortlist.
Here is the thing that breaks almost every Smith.ai alternatives list, including the ones that rank well. These vendors do not sell the same unit, and there are three units in play.
| Vendor | Sold by | Published plans | Answered by | Legal tools named |
|---|---|---|---|---|
| Smith.ai | Calls | $300 / 30, $810 / 90, $2,100 / 300, Enterprise custom | Human or AI first | Clio, MyCase, Lawmatics, Actionstep, Assembly Neos, Filevine |
| Ruby | Minutes | $250 / 50, $395 / 100, $720 / 200, $1,725 / 500 | Human | None named |
| AnswerConnect | Minutes | $350 / 200, $395 / 300, $575 / 400 | Human | Generic CRM only |
| Alert Communications | Minutes | From $240 / 100; pay as you go $2.50 per minute plus $100 retainer | Human | Clio, Filevine |
| Dialzara | Minutes | $29 / 60, $99 / 220, $199 / 500, $349 / 1,000 | AI | None named |
| LegalClerk.ai | Seats | $400 per seat, calls unlimited; Enterprise custom | AI | Clio, MyCase, Lawmatics |
Converting between calls, minutes and seats requires a number most firms have never measured, which is average handle time on an intake call. So normalise them. Hold one firm fixed at 100 intake calls a month averaging five minutes, which is 500 minutes, and ask what each vendor charges that firm.
| Vendor | Monthly at 100 calls / 500 minutes | How it is built |
|---|---|---|
| Dialzara | $199 | Business Plus covers exactly 500 minutes |
| LegalClerk.ai | $400 | One seat, calls unlimited |
| AnswerConnect | about $760 to $765 | Growth or Standard plus overage at $1.85 a minute, plus $49.99 setup on some tiers |
| Smith.ai | $915 | Basic at $810 plus 10 calls of overage at $10.50 |
| Alert Communications | $1,350 | 500 minutes at the published $2.50 pay as you go rate plus the $100 retainer |
| Ruby | $1,725 | The 500 minute plan exactly |
That is a spread of roughly nine times for what buyers describe as the same job. The pattern underneath it is not subtle: the cheapest options are answered by AI and the most expensive are staffed by humans, and the middle of the table is where you are actually choosing between them.
Now change the assumption. At two minutes a call the same 100 calls need 200 minutes, Ruby drops to its $720 plan and Dialzara to its $99 plan, while Smith.ai does not move at all, because Smith.ai never cared how long the call was. Push the other way to ten minutes a call and Smith.ai still does not move, while the minute priced vendors climb their tables or run off the top of them entirely: Ruby publishes nothing above 500 minutes and no overage rate, so at 1,000 minutes you are into a quote rather than a price. Nothing about any of these products changed. Only the assumption did, which is why the first action for anyone shopping this category is not to collect quotes. It is to pull ninety days of call records and find your real average handle time and monthly call count. Without those two numbers every quote you gather is uncomparable.
What "conflict checks" on a feature list does and does not tell you.
Smith.ai lists conflict checks among the legal features on its law firm page. That is a real and useful thing for an answering service to offer, and we are not going to pretend otherwise. What its published material does not describe is what the check runs against, and that gap is where a firm should spend its diligence.
We read seven vendors in this category looking for the answer. None of them publishes that it cross references a caller against the firm's own existing client and matter records. Where the language appears in any detail it describes scripted questions put to the caller during the call, covering opposing parties, prior representation and venue, which is caller self report captured by a trained person rather than a lookup against your database. LegalClerk.ai does not use the word conflict anywhere on its page at all.
The reason this matters is the imputation rule. ABA Model Rule 1.10(a) provides that while lawyers are associated in a firm, none of them may knowingly represent a client when any one of them practising alone would be prohibited from doing so. A clearance decision therefore has to reach the whole firm's client and matter history, and Rule 1.18 extends the exposure to prospective clients who never signed anything. Screening a caller's name against a list you hand a vendor genuinely saves time at the front of the call. It is a different artifact from a documented, auditable clearance decision.
So ask three questions of whichever vendor you pick, and get the answers on the order form: what does the check query, what does it return, and what record does it leave behind. Our longer treatment of that whole workflow, including where the six steps between "we want to hire you" and a matter number actually lose days, is in the intake and conflicts guide.
How to tell which one you are.
Three profiles cover most firms arriving at this page, and only one of them should be talking to anybody about a build.
Your problem is coverage. Calls go unanswered after six, or during hearings, or in the fortnight your intake person is away. Buy the answering service, and buy on your measured call volume and handle time rather than on a headline rate. Six vendors in the table above publish real numbers, which makes them straightforward to size honestly. Decide the human against AI question first, because it moves the price more than any other choice you will make here, and it is the one your callers will notice. This is the large majority of firms, and commissioning anything would be a waste of your money.
Your problem is price at volume. You are answering well and the invoice has become a line item somebody defends every year. The fix is usually still not a build. It is a tier renegotiation, an Enterprise quote, a second vendor bid using your real numbers, or a shorter call, since on a minute priced plan an average handle time cut by one minute is a straight percentage off the bill. Get the Enterprise number before you conclude the market has failed you.
Your problem is downstream of the call. The calls get answered and the cost sits after that: re-keying what the service already collected, deciding whether you may act, assembling engagement letters, routing by matter type against rules only your firm uses. This is the graduation case, and the only one where our answer is worth the money. Note what it is not. It is not a cheaper way to answer phones, and if you are here for that, the table above is a better use of your afternoon than a call with us.
What we would build, and what we have not built.
The fee bands, published rather than quoted on request. A focused build over one clearly defined system, for example a party matching and conflicts search layer over a single matter database, runs $45,000 to $65,000 over four to five weeks. An operations rebuild covering capture through matter opening, with routing, escalation and engagement letter assembly, runs $75,000 to $120,000 over six to eight weeks. A multi system platform with a custom interface for the intake team runs $140,000 to $180,000 over ten to fourteen weeks. Set that against an answering line at $199 to $1,725 a month in the worked example above, and for coverage alone the arithmetic is not close. Most firms in this band should be paying both bills, for different problems.
What we have actually shipped in law firms is the capture and routing half, and the distinction matters because half this page tells you to buy from somebody else. The nameable reference is Jim Glaser Law, where we built five channel specific voice agents covering PPC, Organic, TV, Meta and LSA, giving per channel attribution on answered calls rather than form fills. Those agents have handled 3,787 calls and 5,514 minutes, and Jim Glaser Law will take a reference call. Across all clients our systems have handled more than 6,000 live calls. To hear one, our own receptionist answers at (617) 675-9067.
We have not shipped a firm wide conflicts clearance system, and we would rather say that here than have you discover it on a call. Source code, prompts, models and the architecture document transfer to the firm at handoff, which is what makes the data path auditable permanently rather than for the length of a subscription. The five phases from first call to handoff are on the process page, and the breakdown by scope is in AI consulting cost for law firms.
Smith.ai and its alternatives, answered.
How much does Smith.ai cost?
Smith.ai publishes four tiers, which is worth noting because much of this market publishes nothing. Starter is $300 a month for 30 calls with one free transfer destination and overage at $11.50 per call over 30. Basic is $810 a month for 90 calls with two free transfer destinations and overage at $10.50 per call over 90. Pro is $2,100 a month for 300 calls with ten free transfer destinations and overage at $8.50 per call over 300. Enterprise is custom and carries no published figure. The page states no setup fees, month to month billing, a 30 day money back guarantee, and a 10 percent discount for subscribing for 12 months. Read on smith.ai/pricing on August 15, 2026.
What does Smith.ai actually charge per call?
Divide each tier by its included calls and the real rate appears: $10.00 per call on Starter, $9.00 on Basic, and $7.00 on Pro. The more interesting number is the gap between the included rate and the overage rate on the same tier. Starter overage is $11.50 against a $10.00 included rate, a 15 percent premium. Basic is $10.50 against $9.00, about 17 percent. Pro is $8.50 against $7.00, about 21 percent. The proportional penalty for exceeding your plan is largest on the tier that asks for the biggest commitment, which is the opposite of what most buyers assume when they size a plan.
At what call volume should we move up a Smith.ai tier?
Run the two crossovers. Staying on Starter costs $300 plus $11.50 for every call past 30, so at 74 calls you pay $806 and at 75 calls you pay $817.50 against Basic at $810. Basic becomes the cheaper plan from 75 calls a month, well under its 90 call allowance. Staying on Basic costs $810 plus $10.50 for every call past 90, so at 212 calls you pay $2,091 and at 213 calls you pay $2,101.50 against Pro at $2,100. Pro becomes the cheaper plan from 213 calls a month, against a 300 call allowance. A firm drifting upward on volume is usually overpaying for several months before the invoice is annoying enough to re-read.
What does Smith.ai cost annually at higher call volumes?
At exactly 300 calls a month, Pro runs $25,200 a year before the 12 month discount. Add overage at $8.50 and the annual figure climbs quickly: 400 calls a month is $2,950 a month or $35,400 a year, 500 calls is $3,800 a month or $45,600 a year, and 600 calls is $4,650 a month or $55,800 a year. One caveat has to be stated plainly or the arithmetic becomes a straw man. Smith.ai publishes an Enterprise tier for exactly this situation, and a firm at those volumes should be negotiating it rather than paying list overage. The overage table is what you pay if you never ask.
Which Smith.ai alternatives publish a real price?
More than the category's reputation suggests. Reading each vendor's own pricing page on August 15, 2026, six publish real figures. Smith.ai sells calls at $300, $810 and $2,100 for 30, 90 and 300. Ruby sells minutes at $250, $395, $720 and $1,725 for 50, 100, 200 and 500. AnswerConnect's legal page sells minutes at $350, $395 and $575 for 200, 300 and 400. Alert Communications sells minutes from $240 for 100, and publishes a pay as you go rate of $2.50 a minute plus a $100 retainer. Dialzara sells minutes at $29, $99, $199 and $349 for 60, 220, 500 and 1,000. LegalClerk.ai sells seats at $400 per seat per month for unlimited calls. We deliberately do not reprint figures that exist only on comparison blogs or review aggregators, because those go stale quietly and we cannot audit them.
Why can I not just compare the headline prices?
Because these vendors do not sell the same unit, and three different units are in play. Smith.ai bills by the call. Ruby, AnswerConnect, Alert Communications and Dialzara bill by the minute. LegalClerk.ai bills by the seat with calls unlimited. Converting between them needs a number most firms have never measured, which is average handle time on an intake call. Until you have that, plus your real monthly call count, every quote you gather is uncomparable, and the vendor whose unit happens to flatter your call pattern will look cheapest for reasons that have nothing to do with quality.
What does the same firm actually pay across these vendors?
Take one worked example and hold it fixed: 100 intake calls a month averaging five minutes, so 500 minutes. Dialzara's Business Plus covers exactly 500 minutes at $199. LegalClerk.ai is $400 for one seat with calls unlimited. AnswerConnect works out to roughly $760 to $765 depending on whether you size on the Growth or Standard plan, before its $49.99 setup fee on some tiers. Smith.ai is $915, being the $810 Basic plan plus ten calls of overage at $10.50. Alert Communications on its published pay as you go rate is $1,350, being 500 minutes at $2.50 plus the $100 retainer. Ruby's 500 minute plan is $1,725. That is a spread of roughly nine times for what buyers describe as the same job, and the cheapest options are AI answered while the most expensive are staffed by humans. Change the five minute assumption and the ranking moves, which is precisely the point.
Is Smith.ai answered by humans or by AI?
Both are offered and the distinction matters for an intake call. Its pricing page invites you to choose how calls get handled between an AI first and a human first approach, and publishes no price difference between them. Its law firm page states that all calls on its Virtual Receptionist plans are answered by a real person, 24/7, by North America based agents working from prompts rather than scripts. Across the category the split is clearer: Ruby, AnswerConnect and Alert Communications position as live humans, while Dialzara and LegalClerk.ai are AI first, with LegalClerk listing escalation to a live agent as an optional Enterprise feature. If which one you are buying matters to you, get it on the order form rather than inferring it from the plan name.
Does Smith.ai do conflict checks?
It lists conflict checks as a legal feature on its law firm page. What its published material does not describe is what that check runs against, and that is the question to ask. We looked at seven vendors in this category and none of them publishes that it cross references a caller against the firm's own existing client and matter records. Where the language appears at all it describes scripted questions put to the caller during the call, which is caller self report captured live rather than a lookup against your database. LegalClerk.ai does not use the word conflict anywhere on its page. That gap matters because under ABA Model Rule 1.10(a) one lawyer's disqualification is imputed to the whole firm, so the check that actually clears a matter has to reach the firm's entire history, including prospective clients who never signed. Screening a name against a list you supply is genuinely useful and it is not the same artifact.
Which vendors write intake data into a practice management system?
Two of the six make a claim beyond sending a lead notification, and both name the systems. LegalClerk.ai publishes CRM integrations naming Clio and MyCase in its plan features, and Clio, Lawmatics, Google Calendar and Calendly in its integrations copy. Alert Communications makes the more concrete claim, describing call data flowing into Clio or Filevine without human handoff so that a qualified lead already exists as a matter record by morning. Smith.ai names MyCase, Clio, Lawmatics and Zapier on its law firm page and Actionstep and Assembly Neos on its integrations page, where the described action for Actionstep is to create contacts from new calls and chats. Ruby and Dialzara name no legal practice management tool on the pages we read, and AnswerConnect lists only a generic CRM integrations feature without naming a product. Read the verb rather than the logo: creating a contact is different engineering from writing structured intake fields into an open matter.
Are there setup fees?
They vary, and one vendor is inconsistent with itself. Smith.ai states no setup fees, month to month billing, a 30 day money back guarantee and a 10 percent discount for a 12 month subscription. Dialzara states no setup fees and no long term contracts. LegalClerk.ai states a $0 setup fee on its $400 seat. Ruby states no additional or hidden fees for activation, onboarding, setup or customization, but publishes no overage rate, so ask for it. AnswerConnect's legal page charges $49.99 setup on its Entry and Standard plans and nothing on Growth, which is worth noticing since Growth sits between them. Alert Communications publishes only a disclaimer that set up fees and overages may apply, with no figure. Nobody publishes the number that usually costs more, which is what it costs to leave: notice periods, minimum terms, and what happens to your recordings and transcripts on the way out.
When does an answering service stop being the right purchase?
When the constraint moves downstream of the call. Answering has no firm specific logic in it: greeting a caller, taking a name and booking a slot is the same job at every firm in the country, which is exactly why a product does it well and cheaply. The moment the expensive work is deciding whether you may act, assembling an engagement letter, routing by matter type against rules only your firm uses, or writing structured data into a system in a shape only your firm uses, you are no longer buying answering. Two signals are worth watching: your monthly answering invoice becoming a line item somebody defends every year, and your team re-keying by hand what the answering service already collected.
Should a law firm commission a build instead of buying an answering service?
For coverage alone, almost never, and we will say that on our own comparison page. Six vendors publish real prices in this category and the cheapest credible option in our worked example is $199 a month. Against a fixed fee starting at $45,000 the arithmetic is not close, and most firms in this band should be paying both bills for different problems. Commissioning earns its place when the work that costs you money sits after the call is answered. What we have actually shipped in law firms is the capture and routing half. The nameable reference is Jim Glaser Law, where we built five channel specific voice agents covering PPC, Organic, TV, Meta and LSA to give per channel attribution on answered calls rather than form fills; those agents have handled 3,787 calls and 5,514 minutes, and Jim Glaser Law will take a reference call. Across all clients our systems have handled more than 6,000 live calls. We have not shipped a firm wide conflicts clearance system, and we say so before you ask.
Book the diagnosis call.
Forty five minutes on where your intake actually loses hours, and an honest answer about whether that is a different answering vendor, a tier renegotiation, or something worth building. Often it is one of the first two, and that call costs you nothing.
Read the law firm offering → Or book directly →Related reading.
The neighbouring decision, in full: an AI receptionist against a custom build runs the cost crossover across the category, and law firm intake and conflicts automation covers the buy or build decision on the workflow underneath. On the money, AI consulting cost for law firms and the pricing page. On the rules, bar rules and malpractice exposure. For choosing a partner, how to choose an AI consultant for a law firm, what a law firm automation consultant does, and the consultants law firms actually shortlist. On the systems themselves, the Clio playbook, the Litify playbook and the iManage playbook. On ownership, renting AI against owning it and hiring against commissioning. To size your own numbers first, the law firm diagnostic takes a few minutes. The rest of this series sits on the comparisons hub.