Synthesia launches Roleplay Sessions for live avatar coaching
The July 2026 release moves Synthesia beyond passive training video into real-time practice, automated coaching, skills scoring, and measurable performance improvement.
By WhatAI Editorial Team ·
Synthesia launched Roleplay Sessions on July 22, 2026 as the first product in its upcoming Sessions platform. Instead of watching a prerecorded training video, a learner has a live conversation with an Interactive Avatar that listens, responds, asks questions, and pushes back in real time.
Organizations can create scenarios for cold calls, discovery conversations, customer support, manager feedback, retail sales, leadership, and frontline interactions. Each scenario can be tailored to a business, industry, persona, personality, and language.
After the conversation, an AI coach evaluates the learner against a role-specific skills rubric. The learner receives targeted feedback, while managers can track performance and improvement across attempts, teams, and the wider organization.
This changes Synthesia's position in the market. The platform is no longer limited to creating and distributing avatar videos. It is beginning to compete in practice, coaching, assessment, and performance readiness. The important questions are whether the scoring is valid, whether the scenarios transfer to real behaviour, and whether organizations govern recordings, learner data, bias, and manager access appropriately.
ℹ️
WhatAI Decision Box
✓
Best for:
Learning and development HR internal communications sales enablement support and global teams that need governed avatar-led business video localization interactivity and roleplay.
✗
Not for:
Cinematic filmmaking unrestricted generative video complex live-action production or teams that need advanced editing without credit limits or enterprise governance.
Synthesia's strongest advantage is governed business-video production, localization, learning delivery, and interactive practice rather than cinematic generative freedom.
Measure comprehension skill improvement update speed and localization quality, not only minutes saved during production.
Synthesia has expanded beyond script-to-avatar video. The platform now combines Assistant, Express-2 avatars, Dubbing 2.0, motion graphics, interactive video, Roleplay Sessions, and emerging Video Agents for business communication, training, localization, practice, and skills measurement.
What makes Synthesia different from a general AI video generator?
Synthesia is designed around repeatable business communication rather than open-ended cinematic generation. Teams can convert approved source material into editable branded scenes, use consistent presenters and voices, translate the result, publish through an LMS or branded page, and measure engagement. Enterprise workflows add collaboration, Brand Kits, SCORM, identity controls, and roleplay analytics.
Where Synthesia still requires human design and governance
Professional-looking avatar delivery does not guarantee effective learning or accurate communication. Scripts need clear structure, translations need review, interactive assessments need valid scoring, and personal avatars need informed consent. Video Agents and roleplays also require privacy, bias, tool-access, and escalation controls.
About Synthesia
Synthesia is an AI video platform for business communication, training, localization, and interactive skill development. It converts prompts, scripts, documents, presentations, and web pages into branded videos using AI avatars and voices, while also supporting screen recordings, motion graphics, captions, stock media, translations, analytics, and team workflows. Assistant is a chat-based production agent that can plan, create, and revise scripts and visuals inside the editor. Express-2 avatars combine full-body gestures, facial expression, lip sync, and voice performance, while Personal Avatars can be created from an approved photo or video. Dubbing 2.0 localizes existing video with improved timing, voice quality, translation, glossary support, and lip sync. Synthesia 3.0 expands beyond static video through clickable interactions, branching, Interactive Avatars, Video Agents, and Roleplay Sessions that let learners practise real conversations, receive coaching, and track skills. Synthesia is strongest for learning and development, HR, internal communications, sales enablement, customer education, support, compliance, and global business communication. It is less suitable for cinematic filmmaking, complex live-action choreography, or unrestricted creative video generation. Script quality, instructional design, translation review, avatar consent, moderation, accessibility, and enterprise governance remain essential.
Use Cases
Create employee onboarding and policy videos without filmingBuild technical and compliance training that can be updated quicklyConvert PowerPoint documents and web pages into narrated business videosLocalize training and communication through Dubbing 2.0Create product demonstrations with screen recordings and AI presentersProduce internal communications and executive messagesBuild customer education and support videosCreate sales enablement and product-marketing videosGenerate multilingual explainers from one approved scriptCreate branching learning scenarios and interactive quizzesRun sales customer-support and manager roleplay practiceMeasure learner performance and skill improvement through roleplay analyticsCreate avatar-led B-roll that demonstrates a task or processMaintain consistent presenters branding and terminology across a video libraryAutomate recurring video generation through the APIExport SCORM packages into an LMSCreate accessible captioned video for distributed teamsTest interactive Video Agent experiences for training and customer guidance
Key Features
✓Text-to-video creation from prompts scripts documents presentations and URLs
✓Assistant chat agent for planning creating and refining complete videos
✓Prompt-to-video workflows with cinematic and presentation delivery styles
✓Script import that preserves approved wording
✓More than 240 stock AI avatars
✓More than 160 supported languages and voices
✓Express-2 full-body avatars with facial expression lip sync and natural gestures
✓Customizable avatar outfits backgrounds environments and action footage
✓Avatar B-roll that shows an avatar performing prompted actions
✓Personal Avatars created from an approved photo or video
✓Express-Voice cloning that preserves dialect accent rhythm and tone
✓Workspace-wide pronunciation glossary and reusable pronunciation controls
✓Dubbing 2.0 with improved translation timing voice quality and lip sync
✓Segment-level dubbing corrections without regenerating the full video
✓Motion Graphics for animated timelines charts processes and infographics
✓Gemini Omni Flash public preview for text image audio and video generation or editing
✓AI Playground for experimenting with media-generation features
✓Generative assets for B-roll backgrounds images and supporting visuals
✓Interactive videos with clickable buttons hotspots quizzes and branching
✓Branded public video pages and embedded playback
✓Roleplay Sessions with real-time Interactive Avatar conversations
✓Roleplay coaching and automated skills scoring
✓Roleplay analytics for individual team and organization progress
✓Interactive Avatars and Video Agents for real-time conversation
✓Business knowledge and tool connections for agent workflows where configured
✓PowerPoint import with support for presentations up to current platform limits
✓AI screen recording and transcript-assisted editing
✓Automatic captions subtitles and multilingual playback
✓Templates stock media icons music and Getty integrations by plan
✓Multiple avatars in one scene on Creator and Enterprise
✓Brand Kits custom fonts colors logos and templates on qualifying plans
✓Live team collaboration on Enterprise
✓Comments guests reviews and version control
✓Viewer analytics including views completion and engagement
✓SCORM export for learning-management systems on Enterprise
✓API access on Creator and Enterprise
✓SAML SSO SCIM and advanced role controls on qualifying enterprise plans
✓SOC 2 Type II ISO 27001 ISO 27701 ISO 42001 and GDPR controls
✓Consent-based avatar creation and platform content moderation
✓Responsible AI governance and AI-generated content labelling support
Pricing
Basic
$0
• 1200 credits each month
• Up to approximately ten minutes of generated video each month
• Twenty-five AI-generated video assets
• Nine stock AI avatars
• More than 160 languages and voices
• More than sixty templates
• Three video-editor projects
• Synthesia watermark
• No video downloads
• No Personal Avatars
• Regular content moderation
Starter
$29 per month with annual discount available
• Everything in Basic
• Up to approximately ten minutes of video or dubbing each month
• Annual plans include an annual credit allowance
• Video downloads
• Synthesia watermark removal
• Assistant and AI Video Assistant
• AI Dubbing
• More than 125 stock AI avatars
• Three Personal Avatars
• Customizable on-brand avatars
• Getty Images Icons8 and expanded stock media
• One editor and three guests
• Chat and email support
Creator
$89 per month with annual discount available
• Everything in Starter
• Up to approximately thirty minutes of video or dubbing each month
• Annual plans include an annual credit allowance
• More than 180 stock AI avatars
• Five Personal Avatars
• Multiple avatars in one scene
• Interactive videos
• Branded video pages
• API access within plan limits
• One editor and five guests
• Priority chat and email support
Enterprise
Custom
• Everything in Creator
• Custom credits and video volume
• More than 240 stock AI avatars
• Unlimited Personal Avatars subject to fair-use terms
• One-click translation into more than eighty languages
• Live team collaboration
• Brand Kits and custom templates
• SCORM export
• SAML SSO SCIM and enterprise administration
• Custom editor and guest limits
• Tailored onboarding and implementation services
• Dedicated Customer Success Manager
• Priority moderation and support
• Roleplay Sessions and enterprise interactive workflows where contracted
The Synthesia 2026 comprehensive review covers the core text-to-video capabilities in a way that makes the ideal user profile clear without stating it explicitly. The user-friendly drag-and-drop interface with extensive customisation options for backgrounds, branding and media libraries is the production accessibility that makes Synthesia specifically valuable for organisations that need regular professional video without video production expertise in-house. Transforming documents into videos, creating training materials and producing product demonstrations are the primary use cases, all categories where the content exists in written form and needs to be converted to video format efficiently rather than produced from scratch. The AI avatars covering diverse global representation being highlighted is the inclusivity consideration that matters for global organisations producing training or communications content for diverse employee or customer audiences. The integration with tools like PowerPoint and Canva being available changes the workflow for teams already producing content in those environments. A Synthesia… Read full discussion →
The Synthesia beginner tutorial covers the three project starting paths in enough detail to make the first session productive rather than confusing. Starting from templates for established content formats, the AI Video Assistant for guided creation from a brief, or a blank canvas for complete manual control are three genuinely different starting points rather than the same workflow with different levels of assistance. Knowing which starting point fits your current task changes the first decision correctly rather than discovering halfway through that you chose the wrong approach. The scene-based editor for scripting, adjusting pronunciation and adding gestures is the content detail layer that changes video quality. Pronunciation adjustment being specifically available tells you Synthesia is designed for professional content where incorrect pronunciation of brand names, technical terms or proper nouns is a quality problem rather than an acceptable rough edge. The custom avatar creation for brand-consistent presentation is the enterprise… Read full discussion →
The Synthesia team features video covers collaboration, version history and shared assets in the combination that changes the evaluation from individual tool to team platform. Time-stamped commenting for precise feedback being available means reviewers can attach their feedback to the exact moment in the video rather than writing "around the 2:30 mark" in a separate document. Precise feedback on the exact content being reviewed rather than approximate location is the review quality that changes how useful the feedback is for the editor. Version history for tracking changes and restoring previous states is the safety net that makes experimentation low-risk rather than requiring careful manual version management. Trying a different approach and being able to restore the previous version without manual work changes the willingness to iterate. The shared asset library centralising team media including custom avatars, branded assets and approved content is the brand consistency infrastructure that makes Synthesia viable… Read full discussion →
I run L&D for a mid-sized company. Our training video backlog had been growing for two years because producing them properly required booking a studio, coordinating with subject matter experts, editing, re-recording when something changed. The turnaround was weeks and the cost per video made us selective about what we even tried to produce. We switched to Synthesia six months ago and I want to give an honest account of what that change actually looks like. The AI Video Assistant is the starting point for most of what we produce now. You give it a web link, a PDF, a Word document or a text description of the topic and it generates a full script, selects appropriate scenes and produces a draft video. That draft always needs editing but it is a usable starting point in minutes rather than a blank slate. The avatar library covers over 150 options with… Read full discussion →
The promise of professional video in minutes usually involves compromises. Synthesia actually delivers: The quick tutorial shows the full process from text input to finished avatar video. The avatar quality at this speed is genuinely good enough for training content, internal communications and marketing materials. Would you use avatar videos instead of filming yourself for any category of content? Read full discussion →
Native-language and cultural review remain necessary before publication
Interactivity
👍 Pro
Supports branching quizzes roleplays and real-time Interactive Avatar experiences
👎 Con
Advanced interactive workflows are more complex and often enterprise oriented
Learning and Skills
👍 Pro
Roleplay Sessions connect practice coaching scoring and progress analytics
👎 Con
AI scoring needs validation for fairness accuracy privacy and genuine skill transfer
Collaboration
👍 Pro
Provides guests reviews version control and enterprise live collaboration
👎 Con
Brand Kits live editing and the strongest governance controls require Enterprise
Pricing
👍 Pro
Basic offers a real free tier and Starter provides a practical route to downloads and dubbing
👎 Con
Credits limit revisions and unused self-serve allowances do not roll over
Enterprise Governance
👍 Pro
Offers SSO SCIM SCORM compliance controls onboarding and responsible-AI frameworks
👎 Con
Deployment still requires internal policies for avatars moderation assessment and data handling
How to Get Results with Synthesia: Step-by-Step Workflow
Define the business outcome
Specify the audience, required behaviour, delivery channel, language, source material, approval owner, and evidence that will prove the video or roleplay worked.
Decision point: Use Synthesia for structured business communication and training. Use a cinematic production workflow when storytelling, physical performance, or visual effects are the main objective.
Choose the correct content type
Decide whether the outcome needs a standard video, interactive video, roleplay, dubbing workflow, or emerging Video Agent experience.
Prepare approved source material
Collect the current policy, process, product information, visuals, terminology, pronunciation, legal wording, and learning objectives before using Assistant.
Start with Assistant or an exact script
Use Assistant when Synthesia should plan and draft the video. Use script import when wording has already been approved and must remain unchanged.
Choose an appropriate avatar
Match the avatar, voice, clothing, environment, and delivery style to the audience and message. Use a Personal Avatar only with documented consent.
Build short readable scenes
Break long paragraphs into concise spoken sections. Use screen recordings, motion graphics, diagrams, or avatar B-roll when showing a process is clearer than explaining it.
Test pronunciation and pacing
Preview names, technical terms, numbers, abbreviations, pauses, and emphasis. Save approved pronunciation in the workspace glossary.
Add interaction deliberately
Use hotspots, branching, quizzes, or Roleplay Sessions only where interaction supports a real decision, skill, or learning objective.
Decision point: Do not add interaction merely to make a passive video look more advanced.
Localize with review
Use Dubbing 2.0 or translation workflows, then verify terminology, timing, cultural meaning, pronunciation, and accessibility with a qualified reviewer.
Apply governance and moderation
Confirm avatar consent, content policy, brand approval, privacy, roleplay data access, AI disclosure, and publishing permissions.
Publish to the intended system
Download, embed, share a branded page, export SCORM, or use the API according to the audience and tracking requirements.
Measure the real outcome
Track completion, comprehension, roleplay skill scores, behaviour change, support deflection, update time, localization cost, and credit usage.
Synthesia Gotchas and Limits to Know Before You Start
Each second of generated video currently uses two credits
Unused self-serve credits do not roll over into the next billing period
Editing and regenerating only changed seconds can reduce credit consumption
Basic videos include a watermark and cannot be downloaded
Brand Kits are restricted to Enterprise under the current plan structure
Assistant is available on Starter Creator and Enterprise rather than Basic
PowerPoint imports do not preserve every animation transition or editable object
Personal Avatars require live consent from the same person shown in the uploaded photo or video
Studio avatar creation can require a separate paid annual add-on
Stock and Personal Avatars remain subject to content moderation rules
Avatar action footage is generated and may not accurately demonstrate safety-critical physical procedures
Dubbing translates speech but still needs terminology and cultural review
Dubbing corrections and credit behaviour vary by plan and workflow
Interactive videos and multiple avatars in one scene require Creator or Enterprise access
Roleplay Sessions and Video Agents may require enterprise sales access or staged availability
Automated roleplay scoring can be biased incomplete or poorly aligned with real performance
Learner recordings transcripts and skill scores require privacy and retention governance
Gemini Omni Flash is a public preview and should not be treated as a stable guaranteed feature
AI-generated motion graphics can contain factual visual or accessibility errors
Synthesia is not designed for advanced cinematic editing compositing or highly emotional acting
Which Synthesia Feature Fits Your Use Case
Feature
Good for
Common mistake
Fix
Assistant
Turning approved documents URLs and prompts into structured video drafts
Generating from outdated source material without a clear audience or objective
Provide current sources the required behaviour and approval constraints before drafting
Express-2 Avatars
Consistent professional business presenters with natural gestures
Using long formal paragraphs that make delivery feel artificial
Write conversational sentences with clear pauses and one main idea per scene
Personal Avatars
Executive messages expert-led training and consistent personal presentation
Creating or sharing an avatar without explicit informed consent and lifecycle rules
Document consent usage ownership access review and retirement procedures
Dubbing 2.0
Scaling approved video into many languages while preserving timing and speaker delivery
Publishing the first automated translation without native-language review
Use glossary controls then review terminology culture timing and pronunciation
Motion Graphics
Explaining timelines processes comparisons and structured information
Turning dense source material into an overloaded animated scene
Reduce each graphic to one message and verify data labels hierarchy and contrast
Interactive Video
Branching guidance quizzes knowledge checks and product exploration
Adding clicks without connecting them to a learning or user decision
Define the intended behaviour and what each branch should measure or enable
Roleplay Sessions
Practising sales support leadership and frontline conversations
Using generic scoring rubrics that do not reflect the actual role
Validate scenarios competencies coaching and scoring with experienced managers and learners
Video Agents
Real-time guided training knowledge access and interactive customer experiences
Allowing the agent to access tools or capture data without clear authorization
Use scoped tools consent notices logging escalation and human review
SCORM Export
Delivering Synthesia content through an existing learning-management system
Assuming SCORM export proves learning effectiveness or compliance
Validate completion tracking assessment evidence versioning and reporting requirements
API
Generating repeatable personalized or data-driven business videos at scale
Automating publication without script moderation translation and brand review
Add approval gates templates validation logs retries and spending controls
Consider instead: Colossyan for another learning-focused avatar-video workflow
Multilingual business communication — 5/5
Dubbing 2.0 and broad voice support make one approved source easier to distribute globally
Consider instead: HeyGen for marketing-oriented avatar and localization workflows
Sales and manager roleplay — 5/5
Roleplay Sessions combine live avatars coaching skill rubrics and progress analytics
Consider instead: A specialist AI coaching platform or human facilitator
Repeatable internal and customer education — 5/5
Scripts presenters templates and updates can remain consistent across a large library
Consider instead: Traditional filming for high-emotion or brand-critical flagship content
Interactive support and product guidance — 4/5
Interactive video and emerging Video Agents add guided visual explanation and conversation
Consider instead: A web knowledge base or chatbot for text-first support
Personalized video at scale — 4/5
Creator API access and Enterprise controls support repeatable data-driven production
Consider instead: HeyGen or a custom video-generation pipeline
Creative social and cinematic advertising — 2/5
Synthesia is optimized for structured avatar communication rather than cinematic visual invention
Consider instead: Runway Veo Seedance or a production team
Complex live-action filmmaking and advanced post-production — 1/5
Synthesia does not replace physical performance detailed compositing or full professional editing
Consider instead: A camera crew and DaVinci Resolve Premiere Pro or Final Cut Pro
Starter Prompts for Synthesia
Training Video from Policy
Create a three-minute training video from the attached policy for frontline staff. Preserve every mandatory instruction, remove legal repetition that does not affect behaviour, use short scenes, add one process motion graphic, and finish with a three-question knowledge check. Flag any ambiguous policy language before generating.
Product Walkthrough
Create a concise customer onboarding video from this script and screen recording. Use an approachable presenter, keep the approved wording, highlight the exact interface areas being discussed, add captions, and place a clickable support link at the end. Do not invent product features.
Dubbing Review Brief
Dub this English training video into Indonesian. Preserve the speaker's calm instructional delivery, use the attached terminology glossary, keep product names in English, and flag any segment where timing or translation meaning changed.
Sales Roleplay Scenario
Create a roleplay for account executives practising a first discovery call with a skeptical operations director. Score agenda setting, open questions, listening, problem diagnosis, evidence use, objection handling, and next-step agreement. Do not reward aggressive pressure or unsupported claims.
Video Agent Governance Plan
Plan a customer-guidance Video Agent for product setup. Define the approved knowledge, permitted tools, data collected, consent notice, prohibited actions, escalation triggers, transcript access, retention period, logging, and tests required before the agent can interact with customers.
Prompt pattern: Create [video roleplay or agent] for [audience] to achieve [behaviour or communication outcome]. Use [approved sources and brand]. Preserve [mandatory wording terminology and consent]. Include [avatar visuals interaction and channel]. Verify [accuracy translation accessibility scoring and governance] before [generation publication or deployment].
Iteration tip: Approve a representative thirty-second scene and one translated segment before generating the full video or localization set.
WhatAI verdict on Synthesia
Synthesia is strongest when an organization needs a repeatable system for clear business video, not when it needs an AI filmmaking playground. The editor is built around approved scripts, presenters, templates, screen recordings, motion graphics, captions, localization, analytics, and distribution. That makes it particularly useful for training, onboarding, internal communication, sales enablement, product education, and support.
The platform's 2026 direction goes beyond passive video. Roleplay Sessions let employees practise sales, support, leadership, and frontline conversations with an Interactive Avatar. The system can score performance against skills, provide coaching, and track progress. Video Agents extend the idea further by talking, listening, using business context, collecting information, and potentially taking approved actions.
Basic is a genuine free tier but keeps the watermark and prevents downloads. Starter is the practical individual plan because it adds downloads, Assistant, dubbing, more avatars, and Personal Avatars. Creator adds greater volume, interactive videos, multiple avatars, branded video pages, and API access. Enterprise is where Synthesia becomes an organizational platform through Brand Kits, SCORM, live collaboration, SSO, SCIM, custom limits, implementation support, and interactive training products.
Choose Synthesia over Runway, Veo, or Kling when business consistency, presenters, translation, LMS delivery, and governed collaboration matter more than cinematic freedom. Compare HeyGen for short marketing and avatar content, Colossyan for learning-focused workflows, and dedicated e-learning platforms when assessment logic and course management matter more than video creation.
Evaluate Synthesia with one real business workflow. Convert an approved document into a video, include a screen demonstration, apply pronunciation rules, dub it into another language, add an interaction, publish it to the intended channel, and measure script correction, translation review, viewer completion, and learning impact.
Synthesia — Frequently Asked Questions
What is Synthesia best used for in 2026?
Synthesia is best used for repeatable business video, including training, onboarding, internal communication, sales enablement, product education, multilingual localization, and interactive practice.
How much does Synthesia cost?
Synthesia has a free Basic plan, Starter at twenty-nine dollars per month, Creator at eighty-nine dollars per month, and custom Enterprise pricing. Discounted annual billing is available for paid self-serve plans.
How do Synthesia credits work?
Credits are shared across usage-based generation. Each second of generated video currently uses two credits. Self-serve credits do not roll over into the next billing period.
What is Synthesia Assistant?
Assistant is a chat-based production agent that can turn prompts, scripts, documents, and URLs into videos, then revise scripts and visuals directly inside the editor.
What are Express-2 avatars?
Express-2 avatars are Synthesia's newer full-body presenters with facial expression, lip sync, voice performance, and natural hand and body gestures. Some can also appear in prompted action footage.
Can I create an avatar of myself?
Yes. Paid plans can create Personal Avatars from an approved photo or video, with live consent verification. Enterprise customers can also use additional avatar-request and studio-avatar workflows.
What is Dubbing 2.0?
Dubbing 2.0 is Synthesia's updated localization system with improved translation timing, voice naturalness, glossary support, editing, and lip sync across fast cuts, transitions, and multi-speaker scenes.
What are Roleplay Sessions?
Roleplay Sessions are live practice conversations with Interactive Avatars. Learners can rehearse sales, support, leadership, and other scenarios, receive coaching, and be scored against role-specific skills.
Does Synthesia support interactive video?
Yes. Creator and Enterprise workflows can include clickable elements, branching, quizzes, branded pages, and other interactive experiences. Advanced Roleplay and Video Agent features may require enterprise access.
Can Synthesia replace a normal video editor?
It can replace much of the production workflow for structured business video, but it is not designed for cinematic filmmaking, advanced compositing, complex choreography, or detailed long-form post-production.
