Key takeaways
- An answering service is built to take and relay messages; a virtual receptionist is expected to finish the task the caller rang about.
- Every message creates a callback, and only 19% of US adults generally answer unknown numbers[1], so the message model leaks most with new callers.
- Scripts suit high-volume, predictable calls; judgment suits callers who need answers, but judgment needs written limits, especially in regulated fields[2].
- Sort your callers by type before choosing. The mix, not the brochure, tells you which service fits.
One takes the message, the other does the task
Ring a classic answering service and a polite operator writes down who you are and why you called. Ring a good virtual receptionist and the thing you called about usually gets done: the appointment booked, the question answered, the right person put on the line. Both pick up on the second ring. The difference shows up an hour later, when one business has a message to return and the other has a booked customer.
The official job descriptions draw the same line. O*NET's core tasks for switchboard operators, including answering service, are answering calls, greeting callers, providing information, transferring calls or taking messages, then recording messages and relaying or routing them; answering simple questions about a client's business from reference files is only a supplemental task[3]. For receptionists, core tasks include scheduling appointments and keeping calendars up to date, and hearing and resolving complaints[4]. One job moves information. The other acts on it.
| Answering service | Virtual receptionist | |
|---|---|---|
| End of a routine call | A message for you to return | The task done, or a booked callback |
| Works from | A script shared across shifts | Your script plus calendar and records |
| New customer ready to book | Takes details, promises a call | Books the slot on the call |
| Questions about your business | Basic facts from a reference sheet | Answers from your notes, defers the rest |
| Urgent call | Pages or calls down a list | Transfers by your rules, then logs it |
Typical design of each service. Many providers now sell both, so judge a specific plan by what its routine call ends with.
Why the message model leaks
A message is a promise that someone will call back. Keeping that promise is harder than it sounds, because the callback comes from a number the caller has never saved. In a Pew Research Center survey from July 2020, 67% of US adults said their habit is not to answer unknown numbers but to check any voicemail, 14% said they generally ignore voicemail too, and only 19% said they generally answer[1].
- Don't answer, check voicemail 67%
- Answer the call 19%
- Don't answer, ignore voicemail 14%
Source: Pew Research Center, 2020[1]
Show the data
| Item | Value |
|---|---|
| Don't answer, check voicemail | 67% |
| Answer the call | 19% |
| Don't answer, ignore voicemail | 14% |
So the usual sequence runs like this: the caller leaves details at 10:15, you call back at 11:40 between jobs, it goes to voicemail, you leave a message, they hear it at lunch and ring the main line, which is busy. By mid-afternoon you have played three rounds of phone tag with a person who, in the meantime, has rung two of your competitors. Nobody did anything wrong. The model simply puts the work after the call instead of inside it.
First contact often decides who gets the work. In Zillow's 2026 survey of recent home buyers, 54% of those who contacted an agent contacted only one before hiring, up from 47% in 2023[5]. Buying a home is a big, slow decision; for a leaking pipe or a haircut, the window is shorter still.
That does not make the message model wrong for every call. It works well for vendors, for existing customers with a non-urgent question, and for jobs that always need a site visit before anything can be booked. The trouble starts when it is the only model, applied to new customers who wanted an answer today. If you keep messages, tighten the callback: the callback column in the two-week worth-it test shows how long yours really take.
Scripted versus judgment
Answering services run on scripts for a good reason: the operator on your call may be answering for dozens of other businesses that hour. The script is a decision tree, and the operator's job is to follow it accurately. That gives you consistency and limits risk. It also means anything off the tree becomes a message.
A virtual receptionist works from a script plus context: your calendar, your services, the wording you use for prices, who on the team handles what. That context is what lets them answer the caller who asks whether you can fit them in Thursday afternoon, or whether you work on older homes. It is also what lets them go wrong, so judgment needs written limits.
Law firms show how to draw those limits. ABA Formal Opinion 506 lets a lawyer delegate parts of intake to trained, supervised staff who are not lawyers: collecting first details about the matter, running an initial conflict check, even answering general questions about the fee agreement, as long as the prospective client is always offered a chance to talk with the lawyer[2]. Questions about which legal services to get, fee negotiation and what the engagement agreement means stay with the lawyer. That split works for almost any business.
- Delegate: booking into open slots, rescheduling, hours and directions, intake questions, approved price wording, status updates your systems already show.
- Delegate with a rule: discounts within a set limit, same-day squeeze-ins, waitlist offers, choosing between two technicians.
- Keep with people: quotes that need a site visit or a file review, complaints about work done, anything legal, medical or safety-related, and any caller who asks for a person.
Which suits which caller
| Caller | What they need | Answering service | Virtual receptionist |
|---|---|---|---|
| New customer ready to book | A time, today | Weak: a message and a later callback | Strong: books on the call |
| Prospect comparing three firms | A quick, credible answer | Weak: competitors may answer first | Strong if it can answer from your notes |
| Existing customer, quick question | An answer or a status | Fair: message for the team | Strong if it can see the status |
| Urgent problem | A person, fast | Fair to strong: paging and call lists | Strong: transfers by your rules |
| Upset customer | To be heard by someone responsible | Fair: message flagged urgent | Fair: hear them out, then hand to a person |
| Vendor or sales call | To leave details | Strong: exactly what it is for | Strong: screens and takes a message |
Count a normal week against those rows. If most of your calls sit in the bottom half of the table, an answering service is a reasonable, simple choice. If the top two rows are where your new revenue comes from, a service that only takes messages is working against you. For the full range of virtual receptionist setups, from shared pools to dedicated remote staff and software, see virtual receptionist options compared.
What a message costs: a family law example
Family law makes the gap easy to see, because a person who has just decided to call a divorce or custody lawyer is rarely in the mood for phone tag, and they can ring the next firm on the list in seconds. Here the calculator measures messages whose callback never connects, using the five inputs of the missed-call model with relabelled fields.
| Input | Example |
|---|---|
| Calls that end as a message, per week | 20 |
| Callbacks that never reach the caller (%) | 35% |
| Messages from prospective clients (%) | 60% |
| Prospects who would retain the firm (%) | 30% |
| Fees from a new matter's first stage | $1,715 |
| Unanswered calls a month | 30 |
| From new customers | 18 |
| Bookings that would have followed | 5.5 |
| Monthly value at risk | $9,357 |
How it is worked out: Calls a week × 4.33 weeks × unanswered share × new-customer share × booking rate × average value.
Assumptions
- First-stage fees of $1,715, modelled as five hours at the $343 average hourly rate for family law lawyers in Clio's 2025 Legal Trends Report.[6]
- Most adults let unknown numbers ring out and check voicemail instead, which is why the failed-callback share is set high.[1]
- The message count, the failed-callback share, the prospect share and the retention rate are illustrative; a month of your own message log will give better ones.
Example only. Replace the inputs with a month of your own message log and callback outcomes.
In this example about 87 calls a month end as messages, 30 of them never connect on callback, 18 of those are prospective clients and five or six would have retained the firm. At first-stage fees alone that is roughly $9,400 a month. The firm in this example would not need perfect callbacks to recover most of it; it would need the consultation booked during the first call.
For the full modelled walk-through, including urgent safety calls and weekend intake, read the family law firm intake scenario. Law firms comparing the two services in more depth can use the legal answering service vs virtual receptionist guide at avaforlawyers.com.
Signs a receptionist service is really a message service
Plenty of providers use the word receptionist for a service that, in practice, ends most calls with a message. Five quick checks during the sales call and the first week tell you which one you have.
- The setup form asks for your message preferences in detail but never asks for calendar access or booking rules.
- The sample call recording ends with the operator promising that someone will be in touch, even for a simple booking request.
- Answers to common questions come from a one-page fact sheet, and anything beyond it becomes a message.
- Reports count calls answered and minutes used, but not appointments booked or questions resolved.
- Changing an instruction means emailing an account manager and waiting for every shift to be briefed.
None of these is a scandal; it is simply a different product. If three or more apply, price it and judge it as an answering service, and keep your own callback process tight. If you want the task done on the call, keep looking, and ask each provider to show a recording of a routine booking from start to finish.
Where software fits between the two
Software now blurs the old line. An SI Employee keeps the hours of an answering service, picking up day and night without a queue, but works like a receptionist: it puts callers into open appointment times, answers from the notes you give it, takes a message when that is the right outcome and transfers urgent calls by your rules. Each call is saved with its recording, transcript and summary, and the owner can listen in live or take the call over.
Its limits are the same ones you would set for a human receptionist. It should not improvise on fees, give legal or medical advice, or argue with a caller who wants a person. Those calls go to your team with a clear summary attached. The hybrid setups in AI vs live answering service show how to split the work so nobody duplicates it.
Choosing between them in a week
Tally a week of callers by type
Use the six rows in the table above. Mark which calls brought in new business and which were routine.
Define a finished call for each type
A booking, an answer, a transfer or a message. Write it down so every provider is judged against the same outcome.
Track your callbacks for five days
For every message, note how many attempts it took to reach the caller and whether they had already gone elsewhere.
Test both kinds of service
Make the same five calls, including a new customer who wants Thursday and an urgent problem at 8 pm. Compare what the caller hears and what you receive.
Choose by the caller mix
Mostly routine and vendors: an answering service is enough. New business on the phone: pick the service that finishes the call.
If you stay with messages, make each one complete enough to act on without a second call: name, callback number, reason, urgency and the best time to reach them. If you move to a receptionist model, write down your new-customer intake questions first so the first call collects everything the job needs, and see what a virtual receptionist does for the full task list.
Frequently asked questions
Is a virtual receptionist just an answering service with a nicer name?
Sometimes, which is why you should test rather than trust the label. A true receptionist service finishes routine tasks on the call, such as booking or answering from your notes. An answering service is built to take accurate messages and route urgent calls.
Can an answering service book appointments?
Some can, often as an add-on, and many only take a message. Ask whether the operator can see your open times and book one before hanging up, then test it with a real booking request.
Which is better for after-hours calls?
It depends on what callers want at night. If they need a person for emergencies, either can route them. If they want to book or get an answer, a service that does the task avoids a morning of callbacks, which often go unanswered[1].
Is a virtual receptionist suitable for a law firm or medical office?
Yes, with written limits. Law firms can delegate intake to trained, supervised nonlawyers if the prospective client can always reach a lawyer[2]. Medical offices need a business associate agreement with any service that handles patient information.
How should I compare the cost of the two?
Compare cost per finished call rather than per minute. A cheaper message service can cost more overall once you add the staff time spent on callbacks and the new customers who never pick up.
References
Pew Research Center (2020). Most Americans don't answer cellphone calls from unknown numbers (opens in a new tab) ↑
American Bar Association Standing Committee on Ethics and Professional Responsibility (2023). ABA Formal Opinion 506: Responsibilities Regarding Nonlawyer Assistants (opens in a new tab) ↑
O*NET OnLine (U.S. Department of Labor, Employment and Training Administration) (2025). 43-2011.00 - Switchboard Operators, Including Answering Service (opens in a new tab) ↑
O*NET OnLine (U.S. Department of Labor, Employment and Training Administration) (2025). 43-4171.00 - Receptionists and Information Clerks (opens in a new tab) ↑
Zillow Research (2026). Buyers: Results from the Consumer Housing Trends Report 2026 (opens in a new tab) ↑
Clio (copy hosted by the Alaska Bar Association) (2025). 2025 Legal Trends Report (Appendix A: Hourly rates by practice area) (opens in a new tab) ↑