Multi-channel simulation
Two channels live.
One framework.
Practise via chat and real-time voice with AI personas that react in real time, with dynamic emotional state, configurable patience and regional accents (video is coming soon). Every session is evaluated against your company's competency framework, with criteria locked the moment the roleplay starts.
Chat
Ready. Infinite volume.
Voice
Real-time voice, sub-second latency
Video
Avatar with video, coming soon
On call
Dr. Marcus Rivera
Cardiologist · Regional Medical Centre
03:42
Voice Simulation
Talk. Naturally.
Real-time voice, with infrastructure dedicated to voice roleplay. Sub-second latency, natural interruptions and turn-taking, like a real call. The voice service runs in a dedicated environment per company, with the option of bringing your own AI credentials.
- Dynamic emotional state, configurable initial patience; the persona "heats up" if you push a bad argument and the irritation meter shows up in the session chrome
- Append-only restart, retrying a call creates a new attempt linked to the previous one; roleplay history is never overwritten
- Synchronous transcription, Whisper streams in parallel: the trainee sees the text as they speak, the manager reviews word-by-word afterwards
- S3 recording, Egress records OGG/MP4; signed URLs expire in 1h by default; lifecycle Standard → Glacier → delete is configurable
Dr. Helena Nakamura
Endocrinologist · University Hospital
Chat Simulation
Conversations that feel real.
Text-based role-plays with AI personas configurable by specialty, scepticism level and difficulty. Every message is evaluated against your custom criteria.
- Configurable personas, define specialty, personality, objection patterns and scepticism level
- Progressive difficulty, start easy and ramp up the challenge as skills develop
- Session history, full transcripts saved for review, coaching and audit
- Templates by vertical, ready-made scenarios for pharma, banking, call centre and tech
Video Simulation (coming soon)
Face to face. With AI.
On our roadmap: an AI-generated humanoid avatar with lip sync, facial expressions and body language. Designed for decision-maker meetings, proposal presentations and filmed sales calls. Talk to us if this is on your wishlist.
- AI avatars, realistic digital personas with facial expressions and gestures
- Meeting scenarios, board presentations, customer pitches and negotiation simulations
- Body language analysis, the AI scores non-verbal communication cues
- Full recording, sessions recorded for playback, coaching and compliance evidence
Sarah Patterson
VP of Procurement · Fortune 500
Framework evaluation
No keyword heuristics.
No prompt bias.
The AI scores your company's competency framework, not a generic rubric. The moment the roleplay starts, every criterion from the current version is locked into the session. Editing the framework afterwards does not affect evaluations already in flight: each session stays pinned to the rubric in force when it began.
Why one provider per surface instead of consensus across 3 AIs? Because we measured. Cost, latency and output stability all degraded with ensembles; the gain in score agreement was marginal. We invest in what matters: prompts versioned per surface (chat, voice, drafter, evaluation) with one-click rollback.
Overall score · Pharma Sales framework · v3 (pinned)
87/100
Provider
Anthropic
Model
claude-opus-4-7
Prompt
evaluation v12
Latency
3.2s · $0.018
Strength
Excellent label compliance throughout the conversation. The rep accurately cited SUSTAIN study data and proactively addressed safety profile questions with approved information. "1.8% HbA1c reduction compared to placebo at 56 weeks", accurate and on-label.
Features
Everything you need to train and evaluate
Custom Criteria
Define evaluation criteria specific to each scenario, role or regulatory requirement. Weight criteria by importance.
Actionable Feedback
Strengths, areas for improvement and highlighted conversation excerpts. It's not just a score, it's a roadmap to grow.
Audit Reports
Export complete session records with timestamps, scores and AI feedback. Auditable evidence for regulatory compliance.
Template Library
Ready-made scenarios for pharma, banking, call centres and tech. Customise or build your own from scratch.
Progressive Difficulty
Start with cooperative personas and gradually ramp up scepticism, complexity and time pressure as skills develop.
Real-time Transcription
Every voice session is transcribed in real time. Review, search and export transcripts instantly.
Label Compliance
Automatic detection of off-label claims, missing safety information and unapproved efficacy statements.
Multi-language
Train and evaluate in 23 languages. Same scenario, same criteria, any language your team needs.
Session Recording
Full recordings of every session, chat logs and voice audio for coaching and compliance.
See it in action.
Your team practises with a dynamic persona and gets evaluated against your in-house framework in minutes, not weeks.