Your brand sounds unique.
Your own AI voice for your phone assistant. Trained with ElevenLabs on your brand persona — tone, pace, regional flavour. Live in 14 days. Premium add-on to voiceOne plan from £45/month.
Why your own AI voice?
Because generic voices are interchangeable. And because brand starts with the first hello.
Brand from second 1
The voice is the first thing customers hear. With Custom Voice you sound like YOU — not like 1,000 other AI providers.
Personality to spec
Friendly, professional, casual, polished — trained on your brief. Regional flavour (Yorkshire, Geordie, Scots) possible.
Consistency across channels
Phone, IVR prompts, voicemail footers, ad spots, podcast intros — one voice for everything. Brand identity continuous.
Fast to result
30 minutes of audio is enough. With a professional voice actor: 1 studio session. Self-recorded: 1 day. Training: 7-10 days.
Custom Voice technical capabilities
Powered by ElevenLabs — the leading platform for voice cloning and custom voices.
30 minutes of audio is enough
We record your speaker (studio in Bamberg or remote). Alternatively: existing files with clean quality.
- Studio session inc.
- Script templates (1,500 sentences)
- Multi-speaker brand library
Multi-lingual from one voice
An English voice automatically speaks German, French, Spanish, Italian in your tonality. 13 languages from 1 training.
- English primary + 12 follow-up
- Accent preservation possible
- Switch language per call
Adapted per use case
Greetings friendly, dunning matter-of-fact, emergency answer urgent. One voice, multiple emotional registers.
- Emotion tags in SSML
- Per-industry defaults
- A/B tests in dashboard
Static prompts pre-rendered
Greetings, queue messages, voicemail intro generated as WAV/MP3 in advance. No pay-per-call.
- MP3/WAV export
- Version control
- Hot reload on updates
Real-time TTS
In the call, the AI response is generated in your voice in real time. Latency under 800 ms.
- WebRTC optimised
- Cache for standard answers
- Fallback to standard voice on outage
You own the voice
Your speaker signs a voice cloning agreement. The voice may only be used for you. Theft protection via ElevenLabs watermark.
- GDPR-compliant workflow
- Voice locked to your account
- Audit log on every use
voiceOne Custom Voice vs. standard TTS
What sets Custom Voice apart from Google / Amazon / standard ElevenLabs?
| Feature | voiceOne Custom Voice | Standard voice in plan | Google / AWS standard TTS |
|---|---|---|---|
| Audio quality | Studio, individual | Gemini Native Audio | WaveNet / Polly |
| Brand character | 100 % individual | Standard | Generic |
| Languages from 1 training | 13 | 13 | Different per voice |
| Latency | <800 ms | <400 ms | <500 ms |
| Setup effort | 1 studio session + 14 days | Immediate | Immediate |
| Setup cost | £1,290 one-off | £0 | £0 |
| Monthly licence | £45 / add-on | In plan | $0.03/1,000 chars |
| Rights protection | Your voice | Standard licence | Standard licence |
| Coaching / re-training | Inc. 1x per year | n/a | n/a |
| Brand library (multiple voices) | Yes, £179/mo extra voice | No | Theoretically |
As of 05/2026 — all prices per the providers’ public pricing pages. We refresh this table quarterly.
How brands use Custom Voice
Three real configurations.
National insurer
Own voice ‘Sarah’, trained on the speaker from their current TV spot. Consistency from TV to radio to phone — same person. Aided brand recall up 18 % in surveys.
Boutique hotel group Cotswolds
Voice of chief concierge Maria — ‘Welcome to Hotel Black-Forest. My name is Maria.’ Same voice, 12 sites, 5 languages.
MP candidate 2025
The candidate’s own voice for citizens’ hotline. Personal addressing on city level. Instead of one central voice: one candidate voice in 4 languages.
Custom Voice pricing
Talk to a Custom Voice specialist
30 minutes with our voice architect. We discuss speaker, tonality, studio booking and send you samples.
Frequently asked questions
A method where an AI voice is trained to sound like a specific person. ElevenLabs is the leading platform. From 30 min of clean audio we build a voice model that speaks any text in that voice.
30 min audio recording (1 studio session or existing material), 7-10 days training and QA, 2-3 days integration into voiceOne. Total: 14 days to live.
You do. The speaker signs a voice cloning agreement that explicitly grants you use. The voice may only be used for voiceOne and only for you. Theft protection via ElevenLabs watermark.
Re-training with new material: £849. We keep the old voice as backup. Per voice in the brand set: £179/mo additional licence.
Yes, indistinguishable in 95 % of listening scenarios. We show you a sample before go-live and let you run a 30-sentence test. If unhappy: another training round or refund.
Yes. The voice is stored in ElevenLabs and via studio export usable for ad spots, podcast intros, explainer films. In your voiceOne plan: 10,000 characters/month studio generation included.
ElevenLabs has watermark technology: every generation contains an inaudible signature identifying the source account. On suspected abuse we can prove whether the audio came from your account — or didn’t.
Your own brand voice — live in 14 days.
Studio in Bamberg or remote. Consultation call free.