DeepBrain AI adds emotional expression to AI STUDIOS roleplay training avatars

DeepBrain AI announced on 15 September 2026 that it has expanded AI Roleplay, the training module of its AI STUDIOS platform, so that the simulated AI customer expresses emotion through facial expression and tone of voice rather than delivering flat responses. The company says the AI customer responds dynamically to the trainee based on a configured personality, interests and conversational attitude.

The product provides real-time feedback on the trainee's speaking pace, intonation and facial expression, and a post-session review identifying strengths, areas to improve and suggested alternative phrasing. Scenarios can be configured around an organisation's own products, customer profiles and sales guidelines. DeepBrain names sales, customer service, wealth management, insurance, financial advisory, retail, automotive, telecommunications and education as target sectors.

Eric Jang, CEO of DeepBrain AI, said in the release: "AI Roleplay allows employees to practice with customers who express different emotions and behaviors, review their own approach and improve before the real conversation."

The announcement names no customers and gives no efficacy figures, pricing or adoption numbers.

Why it matters for digital humans: Roleplay training is where avatar realism has an actual job to do — a difficult customer who never looks annoyed is a poor rehearsal — so emotional range is a defensible product direction rather than a demo flourish. It also puts DeepBrain into a segment that Synthesia moved into in July 2026 and that a wave of non-avatar voice-only roleplay vendors already occupy, which raises the question of whether the face earns its cost. The release is a vendor announcement with no named deployments and no published evidence that expressive avatars improve training transfer.