Bring the team building the agent: product and platform owners, CX leads, engineers, and growth or localization specialists. Working together keeps the exercises tied to a real project.

Is this specific to ElevenLabs?

Yes. The program covers voice design, knowledge and tools, turn-taking, testing and deployment on ElevenLabs. For broader agent architecture across vendors, choose our AI Agents training.

Does it cover contact-center strategy or telephony procurement?

No. It covers voice-agent design, conversation, evaluation and operations. We explain platform requirements that affect the build; your teams own procurement and infrastructure decisions.

How does voice cloning work?

Compare Voice Library, Voice Design, and Instant or Professional Voice Cloning. The module covers required consent, documentation and identity checks. We do not provide legal advice; your counsel handles impersonation and biometric law.

Do participants need to be engineers?

No. Most configuration is no-code. Product owners and CX leads work on the persona, knowledge and evaluation; tools, function calls and SIP deployment get a deeper technical track.

How long does it take?

Two days, about 12 hours, onsite or live online. Sessions can be split into half-day blocks.

Can you adapt it to our use case and languages?

Yes. A discovery call identifies your use case, languages and deployment channel. We adapt the exercises and evaluation criteria to your team’s scenario.

What affects the price?

Group size, delivery format, customization and location. Share your use case for a tailored quote.

AI Training · ElevenLabs Voice Agents

Voice AI Agents Training with ElevenLabs

Build a voice agent on ElevenLabs' Agents Platform, from the voice and system prompt to your knowledge base and tools. Tune turn-taking and latency, then test simulated calls before launch.

The training also covers multilingual, web and phone deployment, voice-cloning consent, and AI disclosure. Your team leaves with a working agent and a repeatable build process.

modules: 6
hours: 13

Some of the 500+ brands we've worked with

Selected brands we’ve worked with across industries.

See all references

Build an ElevenLabs voice agent with approved knowledge, tested conversations and a clear handoff to people. Voice adds timing challenges: interruptions, latency and callers who switch languages. A short demo does not test them all.

Design or clone the voice with documented consent, connect your knowledge and tools, then test before deploying to a web widget, phone number or SIP trunk. Disclose the AI voice to callers and monitor it after launch. For broader agent architecture, see AI training and AI Agents training.

Handle interruptions

Test accents, interruptions and out-of-scope questions before callers encounter them.

Keep conversations moving

Avoid half-second pauses and overlapping speech. Voice conversations need different timing from text chats.

Build in consent and disclosure

Keep written consent for each cloned voice and disclose AI on every call. Missing either creates legal and reputational risk.

Choose channels early

Plan web, phone or SIP deployment and language switching early to avoid rebuilding later.

Learn a workflow that lasts

Learn how to configure and test agents as ElevenLabs controls and integrations change.

01
Voice-agent basics and scope
90 minBeginner
Learn how voice, a system prompt, knowledge and tools work together on ElevenLabs. Compare voice agents with chatbots, IVR and general AI agents, and define where this program stops.
What builds a voice agent: a voice, a system prompt, knowledge base, and tools
A voice agent versus a text chatbot and IVR tree, then a general-purpose AI agent
Good first use cases, FAQ and triage, booking, reminders, multilingual front desk
What this program will not touch, identity checks and safety-critical instructions
Where this program stops
02
Voice and persona
120 minBeginner
Compare Voice Library, Voice Design and voice cloning, including consent requirements. Write a persona and system prompt for spoken conversations.
Voice Library versus Voice Design versus voice cloning
Voice Design controls: age, gender, accent, and tone for a voice that does not mimic anyone
Instant Voice Cloning vs Professional Voice Cloning, and the consent each one requires
Writing a system prompt for speech, with pacing and backchanneling built in
One consistent voice and persona across every language
03
Knowledge, tools and turn-taking
150 minIntermediate
Connect approved documents and tools for bookings or CRM lookups. Tune silence, interruptions and turn-taking, and plan human handoffs.
Grounding answers in a knowledge base instead of the model's memory
Wiring tools and function calls for bookings and lookups, including CRM actions
A narrow sub-agent with its own tools and knowledge
Turn-taking tuned by eagerness and silence timeout, plus interruption handling
Clean human handoffs
04
Testing before launch
150 minIntermediate
Build tests from live and simulated conversations. Define success criteria, check answers against your knowledge base, and review failed or dropped conversations.
Turning underperforming live conversations into simulation test cases
Simulation and next-reply tests, plus tool-call tests
Mocking tools safely
Success criteria and structured data collection, defined before launch
A hallucination check against your own knowledge base
Reviewing transcripts as a standing habit
05
Languages and deployment channels
120 minIntermediate
Deploy to a web widget, phone number or SIP trunk. Plan language switching and channel requirements before launch.
Web widget and phone deployment, including SIP
What automatic language switching covers, and what it doesn't
Native phone-number integration versus SIP trunking for scale
Encrypted signaling and static IPs for compliance
One human-handoff route across every deployed channel
06
Consent, disclosure and monitoring
120 minAdvanced
Document cloning consent, disclose the AI voice, and monitor post-call results. Finish with a rollout plan for one tested, governed agent.
Documented consent and rights before any voice clone goes live
Disclosing an AI voice to callers and users, plainly
Detection tooling that catches unauthorized use of a cloned voice
Live monitoring through post-call analytics and evaluation trends
Choosing a pilot use case and champion, with adoption metrics beyond call volume
One tested, disclosed, monitored voice agent, shipped as the capstone

What you will learn

Build an ElevenLabs agent with voice, knowledge, tools and a conversation flow.
Choose a library voice, Voice Design or cloning, with documented consent before any clone.
Ground answers in approved knowledge.
Tune natural turn-taking.
Build a simulation test suite and define launch criteria.
Deploy multilingual agents to a web widget, phone number or SIP trunk.
Disclose AI to callers and keep written cloning consent.
Monitor live agents and plan a pilot-to-rollout process.

Who should attend

Product and platform owners building support, sales or booking agents.
CX and contact-center leads defining the use case.
Growth and localization teams working across languages.
Engineers and no-code builders configuring ElevenLabs.
Founders and product managers piloting a first agent.
Trust and safety staff reviewing consent and disclosure.

Spend two days configuring, testing and improving a working voice agent. Onsite and live-online sessions cover the same syllabus, with small groups so everyone can build.

Format: Onsite or live online
Duration: 2 days (about 12 hours), can be split into half-day sessions
Group size: Up to 12 participants per group (hands-on building)
Materials: Voice and persona brief, test-suite template, and a consent and disclosure checklist
Language: English or Turkish
Certificate: Certificate of completion

Video Production with Sora6 modules · 12 hoursSora and other text-to-video tools can produce a striking clip in seconds. Turning that into a repeatable production stage, from brief and storyboard through motion prompting, continuity, review, and a provenance-tagged handoff, is what this training builds.View training

AI Agents6 modules · 13 hoursNo-code AI agents training for business and operations teams: what an agent really is, when to use one, and how to build and govern one on your own workflows, from Copilot Studio to custom GPTs.View training

AI in Customer Support & CX6 modules · 13 hoursAI training for customer support and CX teams, built around the actual queue: faster response drafting, clearer knowledge content, consistent quality, and escalation that stays visible instead of buried.View training

AI Automation with n8n6 modules · 13 hoursAI automation training for ops and marketing teams: build, test, and govern workflows with n8n, keep a person in the loop where it matters, and hold clear data boundaries throughout.View training

Tell us your use case, languages and deployment channel. We’ll tailor the syllabus and labs to your team.

A colleague adjusting a laptop console while comparing a sample with its output

Who should attend?
Bring the team building the agent: product and platform owners, CX leads, engineers, and growth or localization specialists. Working together keeps the exercises tied to a real project.
Is this specific to ElevenLabs?
Yes. The program covers voice design, knowledge and tools, turn-taking, testing and deployment on ElevenLabs. For broader agent architecture across vendors, choose our AI Agents training.
Does it cover contact-center strategy or telephony procurement?
No. It covers voice-agent design, conversation, evaluation and operations. We explain platform requirements that affect the build; your teams own procurement and infrastructure decisions.
How does voice cloning work?
Compare Voice Library, Voice Design, and Instant or Professional Voice Cloning. The module covers required consent, documentation and identity checks. We do not provide legal advice; your counsel handles impersonation and biometric law.
Do participants need to be engineers?
No. Most configuration is no-code. Product owners and CX leads work on the persona, knowledge and evaluation; tools, function calls and SIP deployment get a deeper technical track.
How long does it take?
Two days, about 12 hours, onsite or live online. Sessions can be split into half-day blocks.
Can you adapt it to our use case and languages?
Yes. A discovery call identifies your use case, languages and deployment channel. We adapt the exercises and evaluation criteria to your team’s scenario.
What affects the price?
Group size, delivery format, customization and location. Share your use case for a tailored quote.

By team

By industry

By tool

Voice AI Agents Training with ElevenLabs