Smith.ai vs. Dialkeep for a solo DMV family-law practice: the overage-month math

Pull up your last Smith.ai invoice before you read this. Actually — pull up two: a normal month, and the month right after a motions term when your phone didn't stop. This post only works if you check its math against your own numbers, so have both in front of you.

Here's the comparison I want to walk you through: Smith.ai, whose AI receptionist tier is billed per call past a plan cap, against Dialkeep, which is flat and which I sell, so weigh everything below accordingly. I'm not going to pretend Smith.ai is a strawman, because it isn't one. The question I care about isn't whether it works. It's what happens to your bill the week your intake calls spike, because for a solo DMV family-law practice, they spike on a schedule you don't control.

Let me be fair to Smith.ai first

I'd lose you fast if I set them up as a punching bag, so let me say what they genuinely do well, because it's real. Smith.ai is an established company with years of track record. You can sign up this week and have your line covered by Friday. Their tier lineup includes live human agents, not only an AI, so a caller in crisis can reach an actual person who can slow down, listen, and handle a sensitive family-law call the way a good intake coordinator would. They offer a bilingual option. If a live human on the phone for that 9pm custody call is the thing you care about most, that's a legitimate reason to want them, and I'm not going to talk you out of it.

Ruby is another live-answering name solo firms weigh against this same tradeoff — a different company, also billed by the minute, and I broke down what a busy month does to their bill for a DMV practice in a separate piece.

So this isn't "them bad, me good." It's narrower and more honest than that. It's about how each one bills you, and what that does to your best month.

What Smith.ai's AI tier actually costs

Here's the structural thing. Smith.ai is metered. Its AI receptionist plans are tiered by call volume, not a flat rate: you buy a bucket of calls, and the meter turns on the moment you empty it.

These are Smith.ai's own published numbers, taken off their AI receptionist pricing page as of August 19, 2026 (pricing pages move — re-check it yourself if you're reading this later):

• Free — $0 a month, 25 real calls included, $3.00 a call after that.
• Pro — $150 a month, 75 calls included, $2.50 a call after that.
• Enterprise — $500 a month, 300 calls included, $2.17 a call after that.

Both paid plans have higher-volume rungs that cut the rate further (Pro at 150 and 300 calls, Enterprise at 500 and 1,000+); the figures above are the base rung of each. Their live-staffed plans, where a human answers rather than an AI, are a different and much more expensive product — Starter $300 a month for 30 calls then $11.50 a call, Basic $810 for 90 then $10.50, Pro $2,100 for 300 then $8.50. Those human rates are the ones worth knowing if what you want is a person on the line; they are not what the AI tier costs, and I'd rather set both in front of you than let the expensive one stand in for the cheap one.

I'll say the uncomfortable part now rather than bury it at the bottom. At a typical solo family-law call volume, Smith.ai's AI tier is cheaper than my $199. Take the cheapest Smith.ai AI plan at each volume and it stays under $199 a month until you're somewhere around 90 calls a month — 30 calls costs you $15 on the free tier, 60 calls costs $105, 75 calls costs $150. If that's your practice, the rest of this post is not an argument for switching, and I'm not going to write it as one.

On a quiet month, in other words, the meter barely registers. Thirty calls a month for a one-attorney family-law practice is a slow trickle: a couple of consult requests, a returning client, a wrong number. Five of them tick the meter, it costs you fifteen dollars, and you never think about it again.

Then a motions term hits.

Why a DMV family-law solo's calls don't arrive evenly

Family-law intake doesn't spread itself out across the month the way a plumber's service calls do. It clusters, and in Fairfax and Montgomery County specifically, it clusters on the court's schedule, not yours.

Fairfax County Circuit Court and Montgomery County Circuit Court both run family dockets and motions days on set calendars. When a motions term opens, and custody hearings, protective-order returns, and contested-divorce motions all get docketed in the same window, the phone activity around it isn't random. A protective-order hearing on the docket this week means intake calls about protective orders this week, not spread evenly across the quarter. That's the nature of family-law demand: it's driven by court dates and life events, both of which arrive in bursts, not a steady drip.

Add a decent month of marketing on top, an ad run or a blog post that finally ranks or a referral source that starts sending real volume, and the burst gets bigger, right when you least want a per-call meter running.

Run the two bills side by side

Take the Pro plan, $150 a month for 75 calls. On a normal month you land under the cap and pay $150. Fine.

Then a Fairfax motions term lands the same week your intake number is doing double duty from a referral push. You take 40 calls past your plan in that stretch: consult requests, a few tire-kickers, several real cases. At $2.50 a call, that's $100 in overage on top of your $150 base — a $250 month, against $150 the month before. The meter doesn't care that a chunk of those calls turned into signed retainers. It bills every one of them the same, real case or wrong number.

Dialkeep is flat. That same month costs $199, the same as the quiet month before it and the quiet month after it. Forty calls or four hundred, the price doesn't move.

So the honest gap on that month is $51, not the several hundred dollars I'd have told you before I re-checked Smith.ai's published rates. I'd rather hand you the small true number than the big false one. The crossover is the thing actually worth writing down: on Smith.ai's AI tier you're under $199 up to roughly 90 calls a month and over it past that, and the further past it you go the wider it gets — at 175 calls the Pro plan is $400, at 200 it's $462.50. A busier practice, or one where a motions term is a regular event rather than a surprise, is on the wrong side of that line most months. A quiet one is on the right side of it and should stay put.

Don't take that on my word, because I'm the one selling the flat plan. Pull your Smith.ai invoice from your last busy stretch, count the calls billed past your included allowance, multiply by your plan's own per-call rate, and set the total next to $199. That comparison is a two-minute job and it beats anything I can tell you about your practice from here.

I'll say the limitation in the same breath as the pitch

Because the honesty only counts if I say the catch out loud, here it is: The number I won't dress up is zero. Zero live family-law intake calls answered, because Dialkeep hasn't taken its first one yet — there's no demo number in this post for you to dial tonight. It's launching soon, and the day it does, the custody-call test is yours to run, not mine to claim. Smith.ai, by contrast, you can sign up for this week and have on your line by Friday, with a live human option for the calls that need one. That's a real advantage for them today, and I'm not going to wave it away to make my own pitch land cleaner.

So what I'm actually selling you right now is the model, not a testimonial I don't have: one flat number, an intake flow built around family-law intake specifically rather than a generic script with your firm's name dropped in, and one number that doesn't move on the month your marketing finally worked.

What it comes down to

Both can staff your phone after hours. One bills you per call once you're past your plan, and a solo DMV family-law practice's real call pattern, court-date-driven and bursty and clustered around motions terms, is exactly the kind of volume that empties a bucket early. The other is flat, so a motions-term spike costs the same as a dead week in August.

But at Smith.ai's actual published AI rates the difference is tens of dollars a month at a solo practice's volume, not hundreds, and below about 90 calls a month it runs the other way and they're cheaper. So this isn't a "you're being fleeced" post, much as one of those would suit me better. It's a narrow claim: past a certain volume a flat number is worth more than a low per-call rate, and where that line sits for you is on your own invoice. Go look at it rather than signing up on a slogan — mine included.

The only arithmetic I trust more than my own pitch is yours, so go do it. Run the free missed-call calculator against your own numbers to see what a missed intake call is modeled as worth for a practice your size (it's an illustration, not a study). And if the deeper cost of a call that isn't handled cleanly the first time is news to you, here's why phone tag quietly becomes a part-time job even when someone eventually calls back. Then pull your own Smith.ai invoice and do the real math. That's the only version of this comparison I'd stake anything on.