How to get a professional voicemail greeting and hold message

This page may contain affiliate links from which Blur Studio earns commissions.

A caller forms an opinion about your business in the first five seconds of your greeting, before anyone's even spoken to them. A flat, echoey recording on a phone's built-in microphone, or the default robotic voice that came with your phone system, both say "we didn't think about this." It's a small thing that's disproportionately noticeable, which is exactly why it's worth the 15 minutes to fix properly.

Write the script first

Keep it short. A greeting that runs past 15 to 20 seconds starts losing callers who just want to leave a message or get through. A solid structure: your business name, a one-line statement of what you do if it's not obvious from the name, your hours if that's relevant, and a clear instruction (leave a message, press a number, stay on the line). For a hold message, keep it even shorter and loop-friendly: thank them for holding, tell them roughly how to expect a response, and avoid specific wait-time promises you can't actually guarantee.

Write it the way you'd say it, not the way you'd write it. "Thanks for calling ABC Plumbing" reads more naturally out loud than "You have reached ABC Plumbing Incorporated."

A concrete example, for a small plumbing company: "Thanks for calling ABC Plumbing. We're open Monday through Saturday, 7 to 6. If this is a water emergency, press 1 to be connected right away. Otherwise, leave your name, number and a quick description of the job, and we'll call you back within the hour." That's four short sentences, covers the business name, hours, an urgent path, and a clear instruction, and it reads at a natural pace in about 15 seconds. A hold message for the same business, kept deliberately shorter since callers hear it on a loop: "Thanks for holding. All our lines are busy helping other customers right now, we'll be with you shortly." Notice it doesn't promise a specific wait time, since an AI voice stating "you will be answered in under two minutes" sounds worse than a vague promise if that number turns out to be wrong that day. For more wording ideas, seewhat to say on a business voicemail.

If your business name itself is unusual, has a non-English pronunciation, or includes a number or acronym (think "24/7 Fast Fix" or a name with a silent letter), say it out loud to yourself first and check whether the generated audio actually gets it right on the first pass. If it doesn't, a short pronunciation dictionary entry fixes it permanently rather than you having to catch and reject the same mispronunciation every time you regenerate.

Generate the audio

1. Paste your finished script into ElevenLabs' Text to Speech tool. If the tool is new to you, abeginner guide to ElevenLabs covers the first setup.
2. Pick a voice that matches how you'd want to sound answering the phone yourself, warm and clear rather than theatrical. ElevenLabs' voice library has dozens of pre-made options across accents and tones; a plain, neutral voice usually reads better for a business greeting than an expressive one.
3. Generate the audio and listen to it on a phone speaker, not just your laptop, since that's closer to how your actual callers will hear it.
4. Export the file from ElevenLabs, which exports standard MP3 or WAV files, covering most phone systems, but the exact format and bitrate your specific provider wants can vary.

Check your phone system actually accepts it

This is the step people skip, and it's the one that actually determines whether this works at all. Some phone systems and VoIP providers let you upload any audio file as a custom greeting. Others only let you record through their own app, or only accept specific formats or file-size limits. Before you spend time perfecting a script, open your phone system's greeting settings and confirm it has an "upload audio file" option, not just "record now." If your provider genuinely doesn't accept uploaded files, the practical workaround is to play the generated audio on a speaker and record it through your phone system's own recording feature, which is a real quality downgrade but still better than nothing.

Once it's uploaded, call your own number from a different phone and listen the way a customer would. Check both the greeting and, separately, the hold message if you have one, since they're often configured in different places in the same system.

Handling after-hours and seasonal variants

Most businesses eventually want more than one greeting: a different message outside business hours, a holiday closure notice, or a seasonal line ("call volumes are higher than usual this week"). The AI text to speech approach makes this genuinely easy compared to a human recording, since generating a second or third version costs you five more minutes and a fraction of a cent in credits, not a second trip to book studio time. The practical habit worth building: keep your script variants in one document (default, after-hours, holiday), and regenerate and re-upload the relevant one when it's actually needed, rather than trying to maintain one greeting that tries to cover every case with conditional language ("if it's a holiday..."), which reads awkwardly and confuses callers who hear it on a normal Tuesday. Setting updifferent phone hours for Fridays and holidays follows the same pattern.

If you operate in a market where customers call in more than one language, for example English and Arabic in the UAE, the same workflow extends naturally: write and generate a separate greeting in each language rather than trying to cram both into one file, and route callers to the right one either through your phone system's own language menu or two separate numbers if that's simpler for your setup.

What this costs

A phone greeting and a hold message together are typically under 100 words combined, well inside the 10,000 free credits ElevenLabs gives every account per month. You can realistically do this entire project, including a few retries to get the tone right, without paying anything. If you later want to clone your own voice so the greeting sounds like you specifically rather than a stock voice, that starts on the Creator plan ($22 a month standing rate, sometimes discounted for the first month only). Music can be added to the hold message too: seecustom on-hold music for your phone line.

Even a business running several locations, each with its own greeting naming the branch and its own hours, stays comfortably inside the free tier: a dozen 100-word scripts is roughly 6,000 to 7,000 characters total, well under the 10,000-credit free allowance in a given month. The only case where this starts costing real money is if you're regenerating the same scripts constantly while iterating on wording, which is exactly the habit the "finalize the script first" advice above is meant to head off.

What I'd tell a small business owner doing this

Don't over-produce this. I've seen business owners spend hours trying ten different voices looking for a "perfect" one, listening back to back until every option starts to sound the same and the decision gets harder, not easier. For a phone greeting, clear and professional beats interesting, every time, because the caller's attention is on what you said, not how you sounded saying it. Pick a voice in the first ten minutes and move on. The thing that actually matters more than voice choice is keeping the script short and specific about what the caller should do next.

One thing worth knowing honestly: if your phone provider won't accept uploaded audio at all, no amount of script polish fixes that, and it's worth checking before you invest time here.

A second thing I'd flag: once you've picked a voice you like, write it down (the voice name, and if you adjusted the stability or speed settings, those too), somewhere you'll actually find it again. Six months from now when you want to update the hours in your greeting, you want to regenerate with the exact same voice and settings, not end up with a slightly different-sounding greeting because you picked whichever voice looked closest from memory.

When to call me instead

If your phone provider doesn't support custom greeting uploads at all, or you want the greeting and hold message to change dynamically, a different message outside business hours, a seasonal update, a language switch based on caller location, that's no longer a one-off recording job, it's a small piece of logic worth building properly. That's the kind of connective work I do for clients: wiring AI-generated audio into whatever phone system you're actually running, so it updates itself instead of needing a new file uploaded by hand every time something changes.

A real example of where this becomes worth building: a business running several locations, each needing its own hours and closure updates, currently has someone manually regenerating and re-uploading files every time something changes at any branch. A small integration that pulls each location's current hours from wherever you already manage that (a spreadsheet, a scheduling tool) and regenerates the right greeting automatically removes that manual step entirely, and it's a genuinely small, well-scoped build rather than a big project.