Fielding a Smiles For Children / DC Medicaid call without turning away a new patient by accident

It's 12:40 on a Wednesday. Lunch, technically, except lunch is a myth at a two-chair practice — the hygienist is running behind on a scale-and-polish, the doctor's mid-crown-prep in the other room, and the front desk is one person who is currently holding a payment terminal in one hand and a sticky note in the other. Then the phone rings.

"Hi — I'm calling around, um, does your office take Smiles For Children? Or, my husband's on DC Medicaid, does that work too?"

And here's the thing about that specific question, asked at that specific moment: the two fastest ways to answer it are both wrong.

The two wrong answers, and why they're the same mistake

Wrong answer number one is the guess. The front desk is slammed, the caller sounds hurried, and "do you take X?" feels like a yes/no it's rude to sit on — so out comes a reflexive "I think so, yeah" or "no, sorry, we don't do Medicaid." Fast. Confident. Possibly wrong.

Wrong answer number two is the punt: "we're really busy right now, can you leave a message?" — which, for this caller, is a hang-up with extra steps. They will not leave a voicemail. They'll dial the next practice on the list, because that's what someone does when they've already been dialing practices.

Different-looking answers, same underlying error: both treat a coverage question like a scheduling question. A "can you do a cleaning next Tuesday?" call can absorb a wrong guess — worst case you reschedule. A "do you take my kid's Medicaid dental plan?" call cannot. Because of who is on the other end of it.

Who's actually calling

Let's be specific about this caller, because the specificity is the whole point.

Smiles For Children is Virginia's Medicaid and FAMIS (CHIP) dental program — the coverage a huge share of Virginia kids on Medicaid actually get their cleanings and fillings through. It's administered for the state's Department of Medical Assistance Services by DentaQuest (you'll also see it under the Cardinal Care Smiles name now), and while it started as children's coverage, Virginia has extended dental coverage to adults on the program too. DC Medicaid dental is its own separate animal across the river — comprehensive for kids, and, unlike most states, it covers adults as well.

I'm not going to tell you which of your patients qualifies for what — that's not my lane and it's not yours to freelance on the phone either. Here's the one thing that matters for how you answer the call: a family shopping a Medicaid or FAMIS dental plan has usually already been told "no" a few times. Not every practice participates. Networks are narrow. The person asking "do you take Smiles For Children?" often isn't casually comparing — they've been working the phone, getting turned away, and quietly bracing for another one.

So a wrong "no" doesn't just decline a patient. It confirms the fear and sends a family that you might genuinely have been able to help to keep dialing. And a wrong "yes" is arguably worse: they take the appointment, rearrange work and a bus schedule and a kid's school day to get there, and find out at the desk that the plan doesn't actually work here. That's not a scheduling hiccup. That's a family that trusted your answer and got burned for it — and dental offices run on exactly that kind of trust.

Why the correct answer is a non-answer

Here's the move that feels wrong and is right: don't answer the coverage question on the call at all. Not with a yes, not with a no. Answer it with a promise to confirm.

"Great question — I want to get you an answer I'm actually sure of instead of guessing, because plans vary and I don't want to send you in on a mix-up. Let me grab your name, the best number to reach you, and which plan you've got, and I'll have someone confirm your specific coverage and call you right back today."

Read what that does. It doesn't turn anyone away. It doesn't promise anything it can't keep. It signals — to a caller who's braced for a brush-off — that this office is going to take five minutes to get it right for them, which is the opposite of every "nope, sorry" they've already collected. And it moves the actual eligibility check to where it belongs: a calm, verified callback, not a guess shouted over a lunch-rush counter.

The whole discipline is: capture, don't conclude. The front desk's job on that first call is to hold onto the facts — the caller's name, callback number, which program they named (Smiles For Children vs. DC Medicaid vs. FAMIS vs. a specific MCO), and what they need — cleanly enough that someone can verify coverage properly and call back. It is not to render a coverage verdict between two patients in the chair. The verdict is the follow-up's job. The call's only job is to not lose the family before the follow-up can happen.

The honest part — where software fits, and where it doesn't

I'll say the part a vendor usually skips: an AI phone line should absolutely not tell your caller whether their Medicaid plan is accepted, and Dialkeep is scripted so it never tries. It doesn't guess at coverage, it doesn't improvise a "yeah I think we take that," it doesn't do eligibility. That's not a gap we're apologizing for — on this exact call, refusing to guess is the correct behavior.

What it's built to do is narrower and, at 12:40 on a Wednesday, more useful: pick up while your front desk is mid-checkout, and when someone asks about Smiles For Children or DC Medicaid, do the careful non-answer on purpose — say plainly that it isn't going to guess at coverage, and ask for the name, the callback number, which program they named, and what they need, in that order, every time. The family that's already been turned away twice this week gets a voice and a careful set of questions instead of four rings and a beep. And no front-desk guess that turns into a burned trip.

And the boundary, up front, because it's where most of the work still sits: that's the end of what it does. It doesn't hand your team anything afterward — no message, no summary, no note — and it can't pass the caller to a person mid-call. Verifying coverage and getting back to that family is your office's job exactly as much as it is today, and if that follow-up isn't already a routine somebody owns, nothing here creates one. (Dialkeep isn't live yet — the demo line is launching soon; the point today isn't to sell you the call — it's that the right way to field this question is the same whether a person or a script is the one answering.)

That's the honest fit. It doesn't decide who's covered. It just refuses to lose the patient before someone can find out.

The bottom line

The Smiles For Children / DC Medicaid call is the one call your front desk cannot answer on reflex — because the caller has likely already been turned away, a wrong "no" sends a coverable family down the road, and a wrong "yes" burns a trip and the trust that came with it. The right answer isn't faster, it's correct: don't guess coverage on the call. Capture the name, the callback number, and the exact program they named, promise a same-day confirmation, and let a real check — not a lunch-rush reflex — decide. Guessing is the only way to actually lose this patient, so build a script that never has to.

That's before you even get to what a new patient is worth in raw dollars once the coverage gets confirmed — I broke down that math separately here: what a missed new-patient call actually costs a DMV dental practice. For the broader math on what a single missed new-patient call is quietly worth to a small practice, see how many calls a small operation actually misses.