Pitcher makes multi-avatar AI roleplay generally available for life sciences

Pitcher, an enterprise field execution platform, announced general availability of multi-avatar AI Roleplay for life sciences. The feature lets sales representatives practise conversations with two or more AI personas simultaneously, in pairings such as a physician and a hospital pharmacist, an HCP and a payer, or a patient and caregiver.

Each persona has a photorealistic face, voice, role and agenda; the release gives no technical detail on how the avatars are built or who supplies the underlying technology. Personas respond in real time, stay in role, and push back on formulary status, prior authorization and competitive alternatives at a configurable difficulty level. Roleplays draw on the customer's CRM data, including post-call notes, prior interactions, advocacy level and coaching history, and scenarios can be generated from approved materials, documents or web pages. Conversations run in 25 languages and are scored in the language they were held in. A compliance rubric uses weighted scoring and pass thresholds to flag unsubstantiated claims and off-message statements, and managers get recordings, transcripts, scores and coaching insights in one dashboard. The service is hosted in EU and US regions. Pitcher gave no adoption numbers or performance metrics.

Why it matters for digital humans: Multi-party roleplay is a harder problem than the single-avatar practice partner most vendors ship. Two personas that interrupt each other and hold separate agendas demand turn-taking and speaker attribution that single-stream avatar stacks do not handle, which is why the research side has been publishing on multi-person and talking-listening generation all year. The compliance-rubric framing is also telling: in regulated selling, the scoring layer is the product, and the avatar is the delivery mechanism. No effectiveness data was released, so treat the training claims as unevidenced for now.