Frequently Asked Questions

ChatGPT Voice / Live supports natural spoken back-and-forth and can be prompted for supplier roleplay. Voice2Evolve adds a standardized procurement workflow: structured scenarios, repeatable pressure, negotiation-specific analysis, and tenant-scoped team reporting.

Official OpenAI source: Official OpenAI documentation: ChatGPT Voice

No. Voice models are one component within Voice2Evolve. The platform adds procurement-specific planning, scenario and counterpart logic, live-session orchestration, acoustic and negotiation analysis, tenant isolation, persistence, and aggregate team reporting.

A prompt can shape one conversation. By itself, it does not create a standardized scenario, comparable repetitions, a stable analysis contract, or tenant-scoped reporting. Procurement teams need the complete workflow, not only a stronger instruction.

For ad hoc conversation practice, brainstorming, or rough wording, ChatGPT Voice may be enough. Once the goal is training procurement behaviour under supplier pushback, it stops short.

No. It complements coaching, playbooks, and workshops. The point is to give teams live reps under realistic pressure so the training survives the next supplier conversation.

Voice2Evolve is a conversational sparring platform for critical procurement conversations. You train supplier negotiations, contract discussions, sourcing pressure, and internal escalations against an AI sparring partner that won't go easy on you. The system doesn't rescue you. It pushes back, escalates, and disengages when you lose control. After every session, you receive forensic performance data: exactly what happened and why.

Most AI tools optimize for comfort. We optimize for consequence. Behavioral science is consistent on this: skills must be rehearsed in conditions that resemble the target environment, including its pressure. Without the pressure, the rehearsal trains the wrong state. Your sparring partner is not helpful. It's realistic. It won't rescue you from poor preparation, soften pushback, or prompt you mid-session. Dead ends are allowed. This is the point: pressure changes performance, and we make that pressure available before it matters.

A custom GPT is a text interface that yields when you push. Voice2Evolve is a voice sparring partner that holds position when you go quiet. Three differences that matter: voice captures what text misses, silence is a functional tool here rather than an absence, and the system is built to stay adversarial under pressure instead of drifting toward helpfulness.

No. You train using your voice only. Think of it like a phone call: you speak, the AI counterpart responds in real time, and the session progresses through spoken exchange. There is no text field to type into and no written transcript to work from during the session. Voice is the only channel because voice is what matters in the conversations you are preparing for.

Because typing is not the skill you are training. The procurement conversations Voice2Evolve prepares you for happen in real time, with your voice. Typing introduces a buffer: time to compose, revise, and optimise. That buffer does not exist in a supplier negotiation or internal escalation. Voice removes it. What you train in is what you perform in. The absence of a text interface is the feature, not a limitation.

No. The AI counterpart stays in role throughout. It will not prompt you, offer suggestions, or acknowledge that it is an AI. If you go quiet, it waits or escalates - it does not rescue you. This is deliberate: prompting during a session trains reliance, not independence. The feedback comes after, in your analysis, with specific guidance on what to do differently next time.

Yes - and that is the point. Repetition under pressure is how skills consolidate. Each session with the same scenario is comparable but not identical: the AI counterpart adapts to your responses in real time and there are no fixed scripts. What you say determines where the conversation goes, which means two runs of the same scenario will feel different and expose different gaps. Track what changes run by run in your analysis.

Yes. In the session planner you can paste your own context directly: a negotiation brief, supplier information, contract excerpts, company background, or whatever the real procurement conversation requires. Every session is then built around that specific material. No scripts, no generic role-plays. For teams, admins can configure shared personas with defined characteristics and organisation-wide training context. Every team member trains against a consistent, realistic baseline that the admin controls, and admins can add, modify, or remove personas at any time.

Yes. Difficulty is set in the final step of the session planner. You control how much pressure the AI counterpart applies - from a challenging but controlled first run to a high-stakes scenario with full escalation. Start where it is useful, not overwhelming. The goal is productive discomfort, not paralysis.

The current focus is procurement negotiation training: supplier discussions, contract standoffs, price increases, internal escalations, and sourcing pressure. The platform is built for conversations where pressure changes performance. More scenarios remain possible, but they are not the focus of this website.

Turn-by-turn voice analytics: energy, pitch, tempo. Pattern detection: where your confidence dropped, where you recovered, where you lost control. Phase breakdown across opening, discovery, and close. The exact moment where leverage shifted. What you missed and what to focus on next: not "practice more", but specific guidance with rationale.

Voice audio is processed in real time to run the session and generate your analysis. It is not stored by Voice2Evolve or retained by our AI provider after processing (session data is not used for model training per API terms). Session transcripts are stored in your account, encrypted, and subject to your retention settings. You can delete your session data at any time.

All data is stored in the EU, encrypted at rest (AES-256) and in transit (TLS 1.2+). Each organisation's data is isolated at the database level through Row-Level Security. We don't sell data and don't use sessions to train AI models. Full details are on the Trust & Security page.

Application data is stored in the EU (Supabase, Stockholm region). All infrastructure providers are SOC 2 Type II certified. A full subprocessor list is published at voice2evolve.com/legal/subprocessors.

We build our Information Security Management System on ISO 27001 and SOC 2 principles. Formal certifications are not currently completed. Security reviews using automated tools and manual verification are conducted on an ongoing basis. A security whitepaper and pre-filled security questionnaire responses are available on request.

Yes. A customer-facing DPA (GDPR Article 28 compliant) is published at voice2evolve.com/legal/dpa and available in English and German. It covers data subject rights, subprocessor management, international transfers, and technical security measures.

No. You can start a session immediately without creating an account. Create one afterward to save your session history and access past analyses.

Yes. You can delete your session history and account data at any time from your account settings. Full account deletion cascades all personal data within 72 hours, in line with GDPR Article 17.

Sessions are priced per use. No subscriptions. Individual sessions and packs are available at different volume levels. For context: a negotiation coach costs €200–500 per hour. A sparring round costs a fraction of that, and you can repeat it as many times as you need. Full pricing is on the pricing page.

No subscriptions for individuals. Sessions are purchased in packs and are yours to use when you need them. Teams can add session packs as needed. See the Teams page for team pricing options.

Unlimited works for passive content. Sparring is active, resource-intensive training. Per-session pricing means every round is purposeful. It also preserves quality: high-fidelity real-time AI and forensic analysis have real costs, and unlimited models force quality tradeoffs we're not willing to make.

No free tier. A single session is available without a pack commitment, enough to experience a full round and receive your first analysis. See the pricing page for the current rate. If the session doesn't deliver what it promises, contact us.

Credits remain available during normal account use while the service remains available. For full terms - including what applies to unused credits - refer to your Master Service Agreement. If you have questions about your specific situation, contact us.

A team setup creates a dedicated environment (a tenant) for your organisation. It includes an Admin Center for user management, central credit allocation, full session history and analytics across the team, and ongoing usage options. Sessions are assigned by the admin and consumed by team members. See the Teams page for the full overview.

The Tenant Admin Center handles everything: user provisioning, credit allocation per user or team-wide, session monitoring, and aggregated performance analytics. Admins top up session credits as needed. Team members access sessions directly. No individual purchasing required.

Team access uses a two-stage model: a one-time tenant creation fee that includes an initial session allocation, followed by ongoing usage via additional session packs. Pricing and included sessions by tier are on the Teams page.

Yes. Individual sessions are available without a team setup. Decision-makers can run a full sparring round themselves before committing to a team licence. For a guided evaluation or larger pilot, contact sales.

Each team operates in a dedicated tenant with database-level isolation (Row-Level Security). Tenant context is validated at three independent layers before any data is returned. There is no shared data surface between organisations. Enterprise tenants can bring their own API keys for AI providers.

A modern browser, a microphone, and a stable internet connection. No app download required. Sessions run directly in-browser, best on Chrome or another Chromium-based browser. We recommend a quiet environment: the sparring partner responds to what it hears.

Sessions currently run in English and German. The platform interface and marketing site are available in English, German, Italian, Spanish, and French. Additional session languages are on the roadmap.

Sessions work best on desktop or laptop with a dedicated microphone. Mobile is technically possible but not optimized. Voice quality and session control are significantly better on desktop.

Each sparring round is 30 minutes. That's deliberate: long enough to develop pressure, short enough to be a focused training block. You can run multiple rounds in sequence for extended conditioning.

No. Sessions run continuously from start to finish. Pausing would break the pressure dynamic the training is designed to create. Real counterparts do not pause. The discomfort of not being able to stop is part of what you are conditioning. If you need to stop, you can end the session early — your analysis will be generated from what was recorded.

Under a minute. Analysis is generated automatically as soon as the session ends. Results are typically ready within 60 seconds.

Sessions recover gracefully. Stay on the session page - the connection will typically restore itself automatically. If it does not, press Start again within a 5-minute window to resume where you left off. Your session progress is preserved during this window.