Live Coaching with Synthesia’s AI Training Platform 2026

Table of Contents

• 1. Synthesia introduces live coaching for employee training
• 2. Step 1: Introduction to Synthesia’s Roleplay Sessions
• 3. Step 2: Understanding the Interactive Training Product
• 3.1 How Roleplay Sessions Work
• 3.2 Feedback Mechanism and Performance Scoring
• 4. Step 3: The Broader Sessions Platform
• 4.1 Future Expansion Plans
• 4.2 Integration with Other Training Formats
• 5. Step 4: Importance of Practice and Feedback in Learning
• 6. Step 5: Positioning as a Performance-Management Platform
• 7. Step 6: Early Adoption and Customer Base
• 7.1 Notable Early Customers
• 7.2 Use Cases in Sales and Leadership Training
• 8. Conclusion: The Future of AI in Training and Development

---

Synthesia introduces live coaching for employee training

Measurable AI Training Outcomes

• What launched: Roleplay Sessions, an interactive roleplay product where employees practice high-stakes conversations with an AI avatar that responds and “pushes back.”
• Why it matters now: enterprises are increasingly asking whether AI spend produces measurable outcomes, not just faster content.
• What’s next: Roleplay is the first release under Synthesia’s broader “Sessions” platform, with expansion planned into formats like job interviews and candidate screening.

• Synthesia has launched Roleplay Sessions, letting employees practice high-stakes workplace conversations with an AI avatar that responds and “pushes back.”
• The product scores performance against a rubric and surfaces analytics, aiming to prove training effectiveness—not just generate content.
• Roleplay is the first release under a broader “Sessions” platform, with planned expansion into formats like job interviews and candidate screening.
• Early enterprise rollouts include customers described as a top-three European company by market cap, a top-five Fortune 100, and a major recruitment company.

Step 1: Introduction to Synthesia’s Roleplay Sessions

Practice, Score, Improve Conversations
1) Pick a scenario (e.g., objection handling, performance review, complaint resolution).
2) Run the conversation with the AI avatar; expect it to probe, challenge, and change direction.
3) Get scored against a rubric immediately after the exchange.
4) Review feedback (what worked, what didn’t) and repeat the same scenario to improve consistency.
Checkpoints: if the scenario prompt is vague, the “pushback” can feel generic; if the rubric is unclear, the score won’t be actionable for coaching.

For years, Synthesia’s enterprise pitch centered on speed and scale: use AI avatars and voice to produce training videos in minutes, at a fraction of traditional production cost. In 2026, the company is widening that scope with a different wager—one that reflects a shifting mood in corporate AI spending.

This overview is based on TechCrunch’s reporting on the launch and on statements attributed to Synthesia and its CEO in that coverage.

This week, the British startup launched Roleplay Sessions, an interactive training product designed for practice, not just presentation. Instead of watching a module, employees rehearse conversations—sales pitches, performance reviews, customer complaints—with an AI avatar that talks back, challenges them, and then evaluates how they did.

The strategic subtext is hard to miss. As enterprises scrutinize whether AI budgets translate into measurable outcomes, Synthesia is positioning Roleplay as a step beyond “content that looks good” toward training that can demonstrate behavioral change. CEO and co-founder Victor Riparbelli framed it simply: video beats text, but practice beats passive consumption.

Step 2: Understanding the Interactive Training Product

Four Layers of Roleplay Quality
Think of Roleplay Sessions as four connected layers:

• Scenario: the situation, roles, and constraints (what “good” looks like in this conversation).
• Interaction: the AI avatar’s back-and-forth, including “pushback” to create realistic pressure.
• Rubric: the scoring criteria (e.g., clarity, empathy, objection handling, policy adherence).
• Analytics: roll-ups across attempts/teams to spot patterns (strengths, gaps, coaching needs).

If any layer is weak (unclear scenario, mismatched rubric, noisy analytics), the “measurable outcomes” promise gets harder to trust.

Roleplay Sessions is built around a familiar corporate reality: many of the most consequential moments at work are conversations. They’re also the hardest to standardize and scale through traditional training, especially when the goal is not knowledge transfer but improved performance under pressure.

Synthesia’s approach is to simulate those moments with an AI counterpart and then attach measurement. The avatar layer is proprietary, while the “reasoning intelligence” underneath is provided by OpenAI—a detail that matters because it pushes Synthesia to differentiate above the model layer.

In practice, the product is aimed at enterprise learning and development teams that want repeatable training experiences, and at executives who want visibility into whether training is actually working. Roleplay can be created by customers directly or built with help from Synthesia’s consultants, using a company’s existing training documents and context.

How Roleplay Sessions Work

At the center of the experience is an AI avatar that can hold a two-way exchange. Employees enter a scenario—such as handling a customer complaint or delivering a performance review—and the avatar responds in real time, including by “pushing back” the way a real counterpart might.

That pushback is the point: it forces the learner to adapt, choose language carefully, and manage tone—skills that are difficult to build by watching a polished example. The scenarios are framed as conversations where confidence and clarity matter as much as factual correctness.

Roleplay is currently positioned as an enterprise offering, consistent with Synthesia’s historical base in corporate training and communications. But the company’s stated direction is broader distribution as costs fall, suggesting the interaction model is meant to become a mainstream format rather than a niche add-on.

Feedback Mechanism and Performance Scoring

Roleplay Sessions doesn’t stop at the conversation. After the interaction, the system scores the employee against a rubric, turning a subjective exchange into structured feedback that helps companies prove training effectiveness.

The scoring also creates a data trail: performance data and analytics that can be aggregated across teams. Riparbelli described “huge interest” from the executive layer in mapping talent—something that has traditionally been difficult to do with soft skills, where managers rely on anecdote, observation, or inconsistent evaluation.

In other words, the product is not only a practice tool for individuals; it’s designed to become an instrument panel for organizations. If an entire sales team roleplays the same scenario, Synthesia argues, leaders can get granular insight into strengths, gaps, and patterns across the workforce.

Step 3: The Broader Sessions Platform

Roleplay is not being presented as a one-off feature. It is the first release under a broader “Sessions” platform, which Synthesia plans to expand into additional interactive formats. The naming is deliberate: “Sessions” implies repeated practice, not a one-time content drop.

This matters because Synthesia’s original moat—fast, localized video creation—faces a reality across enterprise AI: models and generation capabilities diffuse quickly. By building a platform that includes rubrics, analytics, and repeatable practice workflows, Synthesia is trying to anchor value in outcomes and measurement.

The company’s bet aligns with a broader enterprise shift: less interest in AI as a demo and more interest in AI as a system that can be audited, tracked, and tied to performance. Sessions is the container for that shift, with Roleplay as the first proof point.

Future Expansion Plans

Synthesia plans to expand Sessions into other formats, including job interviews and candidate screening, according to information shared with TechCrunch. That direction suggests the company sees interactive simulation as useful not only for training existing employees, but also for evaluating prospective ones.

If Roleplay is about practicing difficult conversations, interview-style sessions could be about rehearsing—or assessing—how candidates respond under structured prompts. Candidate screening, meanwhile, points toward a more operational use: standardizing early-stage evaluation with consistent scenarios and scoring.

Synthesia has not detailed timelines beyond indicating that more formats are planned. But the inclusion of hiring-related use cases signals ambition to move beyond L&D into adjacent HR workflows where measurement and comparability are prized.

Integration with Other Training Formats

Roleplay Sessions arrives alongside Synthesia’s established video training product, which has long been used to create and update corporate learning materials quickly. Synthesia itself argues that most corporate training tends to “inform and demonstrate,” while behavior change requires doing and feedback.

That framing positions Sessions as complementary to video, not necessarily a replacement. Video can introduce a concept, demonstrate best practice, and scale globally; Roleplay can then provide the practice loop that turns exposure into skill.

In effect, Sessions is meant to sit on top of a company’s current training assets and convert them into interactive practice.

Step 4: Importance of Practice and Feedback in Learning

Practice and Feedback Drive Change

• Reported rationale: Synthesia points to a meta-analysis of learning research to argue that corporate training often “informs and demonstrates” but doesn’t reliably change behavior without practice + feedback.
• Product implication: Roleplay Sessions is designed to create that practice loop (repetition under pressure) and make feedback immediate and consistent via a rubric.
• Practical constraint: the article does not name the specific meta-analysis, so readers should treat the claim as Synthesia’s reported interpretation rather than a directly verifiable citation in this piece.

Synthesia’s product rationale leans on a familiar critique of corporate training: too much of it ends at consumption. As reported, the company cites a meta-analysis of learning research to argue that training often stops short of changing behavior, even when the content is well produced.

Riparbelli’s argument is not that video is ineffective—he explicitly says video performs better than text or sending a document. The claim is that, for most skills, people learn best by practicing rather than just reading or watching. Roleplay Sessions is designed to operationalize that idea inside enterprise workflows.

The emphasis on feedback is equally central. Practice without feedback can reinforce bad habits; practice with structured feedback can accelerate improvement. By scoring against a rubric, Synthesia is attempting to make feedback immediate and consistent, rather than dependent on manager availability or coaching quality.

This is also why “soft skills” are a recurring theme. Sales conversations, leadership dialogues, and difficult employee discussions are high-impact but hard to teach through static materials. A roleplay format can create repetition and exposure to variation—two ingredients that traditional training often lacks.

Step 5: Positioning as a Performance-Management Platform

Standardized Practice: Benefits and Risks
Upside if it works:

• Comparable practice data across teams can reveal coaching needs faster than anecdotal manager feedback.
• Repeatability (same scenario, multiple attempts) can show improvement over time, not just completion.

What to watch carefully:

• Rubric risk: if criteria are poorly defined, scores can create false confidence or unfair comparisons.
• Behavior vs. score: people may optimize for the rubric rather than real-world outcomes.
• Manager interpretation: analytics can help, but only if leaders treat them as signals to investigate—not as a complete picture of performance.

Roleplay Sessions also functions as a strategic repositioning. Synthesia’s avatar and voice technology are proprietary, but the reasoning layer is OpenAI’s—meaning the company cannot rely solely on model capability as a durable advantage.

Instead, Synthesia is adding a layer of rubrics, performance data, and analytics, and using that to position itself “less as an AI avatar company and more as a performance-management platform with a video front end.” The phrase is telling: the video and avatar are becoming the interface, while the defensibility shifts to measurement and organizational insight.

From a management perspective, the promise is visibility. If employees practice standardized scenarios, leaders can compare performance across teams and identify where coaching or enablement is needed. Riparbelli described executive interest in “mapping out the talent” in a way that has been difficult before—particularly for skills that don’t show up neatly in spreadsheets.

This approach also matches the broader enterprise moment. Companies are increasingly asking whether AI spend is paying off. A platform that can show progress over time, benchmark teams, and connect training to observable performance signals is better aligned with that scrutiny than a tool that only produces content faster.

Step 6: Early Adoption and Customer Base

Scaled Enterprise Roleplay Deployments
What’s been publicly described so far:

• Synthesia says Roleplay has early customers with scaled commercial rollouts (not just pilots).
• Customer identities were not disclosed; a spokesperson described them as a top-three European company by market cap, a top-five Fortune 100, and a major recruitment company.
• Most common reported deployments: sales and leadership training, largely focused on soft skills.

Why this matters: scaled rollouts typically imply real integration work (enablement, reporting, and internal adoption), even when customer names aren’t shared.

Synthesia says Roleplay already has customers with scaled commercial rollouts, suggesting the product is not confined to pilots or innovation labs. While the company did not name the organizations, it described them in terms that signal large, complex deployments.

The early traction also reflects where demand is strongest: roles where conversation quality directly affects revenue, retention, or organizational health. Synthesia says the two most popular use cases are sales team training and leadership training, much of it focused on soft skills.

Roleplay’s current enterprise focus is also pragmatic. Large companies have the budget, the training infrastructure, and the need for consistency across distributed teams. But Synthesia’s stated goal is to broaden access to small businesses, prosumers, and potentially educational institutions as inference costs fall.

Notable Early Customers

According to a company spokesperson, early Roleplay customers with scaled rollouts include:

• One of the top three companies by market cap in Europe
• One of the top five Fortune 100 companies
• One of the biggest recruitment companies in the world

The lack of names limits external verification, but the descriptors indicate Synthesia is targeting—and winning—large enterprise accounts where training standardization and measurement are perennial challenges.

These kinds of customers also tend to be demanding about governance, reporting, and integration into existing processes. That aligns with Synthesia’s broader push to make Sessions about outcomes and analytics, not just an engaging AI interaction.

Use Cases in Sales and Leadership Training

Synthesia says the most common Roleplay deployments so far center on sales and leadership—two domains where performance often hinges on how people handle nuanced, sometimes tense conversations.

For sales teams, roleplay can simulate pitching a product, responding to objections, and maintaining composure when a prospect pushes back. The value is repetition under realistic pressure, paired with scoring that can highlight patterns across a team.

For leadership training, the scenarios skew toward difficult conversations: performance reviews, feedback delivery, and handling complaints. These are moments where managers often lack practice, and where inconsistent execution can damage trust. Roleplay’s promise is a safe environment to rehearse, receive feedback, and improve—without waiting for the next real-world incident to learn the hard way.

Conclusion: The Future of AI in Training and Development

Key Signals to Monitor
What to watch next as Sessions expands:

• Whether companies use Roleplay scores as coaching inputs (development) vs. evaluation outputs (performance ranking).
• How transparent and customizable the rubrics are for different roles and cultures.
• Whether analytics can show improvement over repeated attempts, not just one-off scores.
• How well Sessions integrates with existing training assets so video “inform/demonstrate” content reliably feeds into practice.
• If interview/screening formats launch, how Synthesia handles consistency vs. nuance in hiring-related scenarios.

Synthesia’s move into Roleplay Sessions signals a broader evolution in enterprise AI: from generating artifacts to demonstrating impact. In training, that means shifting from “we produced content faster” to “we improved performance measurably.”

Roleplay also reframes what an AI training platform can be. Instead of a library of modules, it becomes a practice environment—one that can be repeated, scored, and analyzed. If Sessions expands into interviews and screening as planned, Synthesia’s footprint could extend beyond L&D into how companies assess and develop talent more broadly.

Embracing AI for Enhanced Learning Experiences

Synthesia’s core insight is that engagement alone is not the finish line. AI avatars and polished videos can make training easier to produce and more consistent to deliver, but the company is now emphasizing the next step: interactive experiences that require employees to respond, adapt, and perform.

That approach fits the enterprise climate described by Riparbelli: companies care less about whether AI looks impressive in a demo and more about whether it produces measurable business outcomes. Roleplay is designed to meet that demand with a format that is inherently performance-oriented.

The Role of Continuous Feedback in Skill Development

The most consequential element of Roleplay Sessions may be its feedback loop. By scoring conversations against rubrics and generating analytics, Synthesia is trying to make soft-skill development trackable—something organizations have long struggled to do consistently.

If that loop works as intended, it could change how companies think about training: not as a one-time event, but as continuous practice with measurable progress. In a world where AI tools are increasingly commoditized, the differentiator may be less about generating the next piece of content—and more about proving, with data, that people got better.

Perspective: This analysis reflects how enterprise teams typically evaluate training systems in practice—especially where measurement, repeatability, and governance matter—drawing on Martin Weidemann’s background building and scaling technology products in regulated, multi-stakeholder environments across Latin America.