AI in Sales
Twenty Seconds to a No: What the In-Network Question Costs Your Dental Group
Most dental groups treat the in-network question as a records lookup, and the caller is gone in twenty seconds with no record anywhere. Here is a five-step audit to find out how often it happens in your practice.
by Jerrod Anthraper
9:14 on a Tuesday. The phone rings and the caller asks one question before she gives her name: "Are you in-network with Delta PPO?" Your coordinator checks, says no, and offers to email a fee schedule. The caller says thanks and hangs up. Twenty-two seconds, start to finish.
No name in the practice management system. No note in the call log. No line on any report you will read this month. She was calling about a cracked molar and a bridge she had been putting off for two years, and she is already dialing the next practice on the results page.
The question is getting answered when it should be getting qualified
That call was not a records lookup. It was the opening line of a sales conversation, and your front desk treated it like a yes-or-no database query.
Here is what makes it expensive: 78% of sales go to whoever responds first. The instant your coordinator said "no, we're out of network," the race restarted at zero and someone else got the head start. Not because the other practice is better. Because they are the next one to actually engage.
Multi-location groups feel this hardest. A single practice might field a handful of these a week. A ten-location DSO fields them all day, across ten front desks, with ten different people improvising ten different answers — and not one of those calls creates a record anywhere. The leak is invisible by design.
Why good front desks do this
Nobody is being lazy. They are doing exactly what you trained and measured them to do.
Your coordinator is accountable for schedule accuracy, insurance accuracy, and not creating write-offs. Telling an out-of-network caller "come on in" and having that blow up at checkout is a real, punishable mistake in most practices. Saying "we're out of network" is technically correct, protects the schedule, and gets the next call answered faster. It is the rational move inside your incentive structure.
Second problem: the call leaves no trace. A patient who books shows up in production reports. A patient who hangs up at twenty seconds shows up nowhere. You cannot manage a number you have never seen, and most groups have never seen this one.
Third: benefits questions feel clerical. They arrive sounding like paperwork, so they get handled by whoever handles paperwork, at the exact moment the caller is most willing to be persuaded.
A one-week audit you can run without buying anything
This costs you about ninety minutes of attention spread over five days. Do it before you decide whether you have a problem.
Step 1: Pull every inbound call under 90 seconds from the last 30 days.
Your phone system almost certainly has this report — VoIP platforms, call-tracking numbers, and most practice phone vendors all expose call duration. Filter to inbound, external, answered, under 90 seconds. Do not filter out short calls as noise. Short calls are the dataset.
For a single location, expect somewhere between 40 and 120 of these in a month. For a group, run it per location so you can see variance between front desks. That variance is usually the most useful thing you learn.
Step 2: Listen to twenty of them and write down the first question the caller asked.
Just the first question, verbatim, in a spreadsheet. No analysis yet. You are looking for the frequency of four openers: "Are you in-network with ___?", "How much does ___ cost?", "Do you take ___?", and "Do you have anything today?"
If more than a third of your sub-90-second calls open with an insurance or price question, you have found the leak and you can stop looking for it elsewhere.
Step 3: Tag how each one ended, using four buckets.
Booked. Callback promised. Information given, no next step. Caller ended the call. The third and fourth buckets are your loss column. Add up the minutes — most owners are startled that a month of lost new-patient opportunity adds up to less than an hour of phone time. It is a small leak in duration and a large one in dollars.
Step 4: Check how many of those callers exist anywhere in your system.
Take the twenty calls. Search each phone number in your PMS. Count how many produced a patient record, a lead record, or even a note. In most practices that number lands between zero and three out of twenty. Those seventeen people asked you for something, told you their need, and left no footprint. They are not on a recall list. They are not in a campaign. They cannot be followed up with, because operationally they never happened.
Step 5: Call your own practice as a patient, twice.
Once during your busiest morning hour, once at 5:50 p.m. Ask the in-network question about a plan you do not accept. Time how long it takes to get to a no, and notice whether anyone asks your name, what is going on with your teeth, or whether you would like to hear what the visit would actually cost you.
Then do the same to the two practices you compete with most. The gap between their call and yours is your actual competitive position at the moment of first contact, and it has nothing to do with your clinical quality.
What a healthy version of that call sounds like
The fix is not "say yes to everyone." It is sequencing. Get the name, the callback number, and the problem before you deliver the benefits answer. Then answer honestly, and follow the honest answer with a specific next step — a real out-of-network estimate, a membership plan number, or the one plan you do accept.
Same information, same integrity, different order. The caller now exists in your system either way, which means a "no" becomes a follow-up instead of a disappearance.
The proof
This is the kind of work that suits a machine better than a person. It is repetitive, it happens at the worst possible moments, and it demands identical execution on the fortieth call of the day as on the first.
Tykon's framework has benchmarked above a 32% win rate while requiring only 27% of the work a human sales team puts in — near parity on performance, at a fraction of the labor, with none of the day-to-day variance.
That last part matters most for a group. The difference between your best front desk and your worst one is not a training problem you can permanently solve. It is a staffing reality. Consistency is the thing you are actually buying.
Where this goes next
James, our AI sales agent, answers that 9:14 call in under sixty seconds — including the ones that land while your coordinator is checking a patient out. He asks for the name and the reason for the call before he addresses benefits, gives the caller an accurate answer, and books the consult when there is one to book. Every call becomes a record, whether it ends in an appointment or not, so the twenty-second hangups stop being invisible.
If you are not ready to look at software, run the five-step audit anyway. Step 5 alone takes four minutes and tells you most of what you need to know.
And if the audit shows what it usually shows, the next step is small: send us one location's call log and we will tell you what we see in it. No pitch attached to that — we would rather you know the number than guess at it.