Key takeaways
- Humans win on judgment, empathy and handling the unexpected; AI wins on coverage, consistency and handling many calls at once.
- A receptionist's wage is only part of the cost: benefits are 30% of private employer compensation costs[1].
- Callers are often free when you are closed: only 30% of employed Americans work on an average weekend day[2].
- Supervision duties apply to both, and AI tools that talk to clients need clear disclosure[3][4].
The real question
Law firms rarely choose between a human receptionist and an AI one in the abstract. They choose how to cover specific hours and specific calls: the morning rush, lunch, court days, evenings, weekends and the moment when three lines ring at once. A human and an AI system are good at different parts of that problem. The useful comparison is task by task, with the costs and the professional duties counted honestly on both sides.
Side by side
| Factor | Human receptionist | AI receptionist |
|---|---|---|
| Hours | Scheduled hours, plus cover for breaks, illness and leave | Every hour, including nights, weekends and holidays |
| Several calls at once | One at a time; others wait or go to voicemail | Handles simultaneous calls |
| Judgment with unusual calls | Strong; can read tone and improvise | Limited to what it was configured to do |
| Empathy with distressed callers | Strong when well trained | Can follow calm, approved wording; cannot truly empathize |
| Consistency | Varies with mood, workload and training | Same questions and wording every time |
| Languages | Those the person speaks | Depends on the product; check quality per language |
| Office tasks | Greeting visitors, mail, scheduling, filing | Phone and possibly chat only |
| Records | Notes as good as the person takes | Recording, transcript and summary if configured |
| Start-up | Hiring and training | Configuration, testing and approval of scripts |
Where a human receptionist is better
- Visitors. Anyone who walks into the office needs a person.
- Callers in crisis who need someone to slow down, listen and stay with them until help is reached.
- Long-standing clients who expect to hear a familiar voice and ask about matters in progress.
- Calls that do not fit any category, where judgment matters more than consistency.
- Office work beyond the phone: mail, deliveries, scheduling rooms, chasing documents.
Where an AI receptionist is better
- Nights, weekends and holidays, when the alternative is voicemail.
- Overflow during busy hours, so a second or third caller is answered instead of waiting.
- Asking the same intake questions in the same order every time, which helps conflict checks and reporting.
- Producing a written summary and transcript of every call for staff to review.
- Covering sudden absences without a scramble for temporary staff.
Work out the true cost of a human receptionist
The wage is the visible part. The median hourly wage for US receptionists was $18.27 in May 2025[5], about $38,010 a year[6]. Employer costs run higher. Across US private industry, employer compensation costs averaged $46.89 per hour worked in June 2026, and benefits made up 30.0% of that, or $14.07[7][1]. Those are economy-wide averages, not receptionist figures, but the benefits share is a reasonable starting point for estimating what a hire really costs.
- Wages and salaries $32.82 70%
- Benefits $14.07 30%
Source: U.S. Bureau of Labor Statistics, 2026[7]
Show the data
| Item | Value |
|---|---|
| Wages and salaries | $32.82 |
| Benefits | $14.07 |
Start with the wage
Use your local wage or the national median as a starting point.
Add benefits and payroll costs
A share in line with the national benefits share is a reasonable first estimate; replace it with your actual figures.
Add cover
Count the hours you would need cover for breaks, holidays, illness and leave, and how you would fill them.
Add the hours you cannot cover
Evenings and weekends remain uncovered unless you add more staff or an answering service.
Compare like for like
Compare against the hours and call volume an AI system would actually cover, plus the staff time to review its summaries.
When callers can actually call
People contact a law firm when they have time and privacy, which is often outside business hours. In 2025, 81% of employed Americans worked on an average weekday, against 30% on an average weekend day[2]. Most working people are free on weekends and evenings, and that is exactly when a firm with only a daytime receptionist sends them to voicemail.
Turnover is a hidden cost
Front-desk roles turn over often. Even though receptionist employment is projected to decline, about 105,100 openings are projected each year on average from 2025 to 2035[8], largely to replace people who leave. Each departure means hiring, training and a period where intake quality drops. That is not an argument against people, but it belongs in the comparison, along with the time a firm spends keeping an AI system's scripts and settings current.
The ethics rules apply to both
A receptionist is a nonlawyer assistant, and ABA Model Rule 5.3 requires partners and supervising lawyers to make reasonable efforts to ensure a nonlawyer's conduct is compatible with the lawyer's professional obligations[3]. Firms apply the same thinking to AI tools. Florida Bar Opinion 24-1 adds that generative AI chatbots communicating with clients or third parties must include a disclaimer that the chatbot is an AI program and not a lawyer or employee of the firm[4]. Whichever you choose, the firm approves the script, the questions and the boundaries: no legal advice, no promise of representation and no decisions about conflicts.
Hybrid models that work
- Human during core hours, AI after hours and on weekends.
- Human as the first answer, AI for overflow when the line is busy.
- AI as the first answer for new inquiries, with warm transfer to a person for existing clients and urgent calls.
- Human receptionist plus AI-generated call summaries, so notes are consistent without changing who answers.
Most small firms end up with one of these. The right mix depends on call volume by hour, which you can measure from your phone system before deciding anything.
Decide in five steps
Measure calls by hour
Pull a month of call logs and chart calls by hour and day, including missed calls.
List the call types
New inquiries, existing clients, opposing counsel, courts, vendors, and urgent matters. Note which need judgment.
Decide who handles each
Assign each call type and time slot to a human, an AI system or a hand-off between them.
Price the options
Compare the full cost of each mix for the hours it covers.
Pilot and review
Run the chosen mix for 30 days and review missed calls, complaints and conversion of new inquiries.
What callers notice
Callers judge the firm on three things in the first minute: whether someone answered, whether they were understood, and whether they know what happens next. A rushed human receptionist juggling three lines can fail all three; so can an AI system with a confusing script. Whichever option you choose, end every new-inquiry call the same way, by telling the caller what will happen, roughly when and who will contact them, without promising what the firm cannot guarantee.
Questions to ask an AI receptionist vendor
- How does the system disclose that it is AI, and can we approve the wording?
- Can it follow our intake questions for each practice area, or only a generic script?
- How does it handle a caller who is upset, confused or in danger?
- How are urgent calls escalated, and what happens if nobody answers?
- What do we receive after each call, and how quickly?
- Are calls recorded, how is consent handled, and who at the vendor can access recordings?
Training and supervision for either option
A new receptionist needs training on intake questions, conflict basics, confidentiality, what not to say and when to interrupt a lawyer. An AI system needs the same knowledge written into its configuration, and someone to check that it applies it. In both cases, a supervising lawyer should listen to a sample of calls in the first weeks, correct the script or the training, and keep doing spot checks after that. The supervision is the same duty in a different form.
Write down what was checked and what changed. A short log of call reviews shows the firm took reasonable steps, and it makes the next review faster.
Example: a three-lawyer immigration firm
Example (fictional): a three-lawyer immigration firm has one receptionist working 9 to 5. Call logs show a third of new inquiries arrive after 5 p.m. or on Saturdays, and many daytime callers hang up when the line is busy. The firm keeps its receptionist for walk-ins, existing clients and complex calls, and adds an AI system for overflow and after-hours calls with approved intake questions and a clear AI disclosure. After a month, voicemails from new inquiries fall sharply, and the receptionist spends her first hour each morning returning calls from the AI summaries instead of listening to voicemail.
Frequently asked questions
Can an AI receptionist replace a human at a law firm?
For some hours and call types, yes. Walk-ins, callers in crisis and calls that need judgment still favor a person. Most small firms use a mix.
What does a human receptionist really cost?
Start with the wage, then add benefits and payroll costs, which average 30% of private employer compensation costs nationally[1], plus cover for absences.
Do ethics rules apply to an AI receptionist?
Firms should supervise it as they would a nonlawyer assistant under Rule 5.3[3], and AI chatbots that talk to the public need a clear AI disclosure under Florida's guidance[4].
What is the best hybrid model?
A common one is a human during core hours and an AI system for overflow, evenings and weekends, with warm transfer for urgent calls.
References
U.S. Bureau of Labor Statistics (2026). Employer Costs for Employee Compensation - June 2026 (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2025). American Time Use Survey - 2025 Results (opens in a new tab) โ
American Bar Association (2026). Rule 5.3: Responsibilities Regarding Nonlawyer Assistance, ABA Model Rules of Professional Conduct (archived copy) (opens in a new tab) โ
The Florida Bar (2024). Florida Bar Ethics Opinion 24-1 (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2025). Receptionists: Occupational Outlook Handbook (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2025). Receptionists: Occupational Outlook Handbook (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2026). Employer Costs for Employee Compensation - June 2026 (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2025). Receptionists: Occupational Outlook Handbook (opens in a new tab) โ