Project Description
The Challenge
Realistic sales training requires conversation partners that behave like real customers - not like chatbots. Each customer type needs a consistent personality, comprehensible reaction patterns, and a natural conversation flow. At the same time, the subsequent feedback must be objective and conducive to learning, not subjective and demotivating.
Our Approach: Typology-Based Persona System
Scientific Personality Model
The personas are based on the TetraMap and DISC model. Four basic types form the foundation:
- Analytical-factual: Needs facts, data, logical arguments
- Expressive-emotional: Responds to enthusiasm, visions, personal stories
- Controlled-dominant: Wants control, quick results, clear advantages
- Harmonious-relationship-oriented: Seeks trust, personal connection, security
Additionally, 6 mixed types represent realistic personality combinations - a total of 10 distinct behavioral profiles.
5 Resistance Levels
Each persona can be calibrated across 5 levels - from open and interested to strongly resistant. Resistance affects not only objections but also tonality, willingness to engage, and body language cues. This creates 1,212+ unique conversation combinations.
Configurator: Persona, Type, Resistance
Through an intuitive configurator, users select their desired training setup:
- Industry and context of the persona
- Personality type (basic or mixed type)
- Resistance level (1-5)
- Training scenario (e.g., cold calling, reactivation, existing customer)
7+ Training Scenarios
From cold calling to reactivating inactive customers to existing customer conversations - each scenario has its own briefings that define the context and expectations of the AI persona. Scenarios are extensible via an admin interface.
Dual-AI Architecture
Two AI systems work in parallel:
- Conversation AI: Conducts the realistic customer conversation with consistent personality
- Analysis AI: Evaluates the conversation in real-time using our proprietary EQ-CORE model and provides discreet feedback
EQ-CORE Evaluation Model (Proprietary Development)
Feedback is based on 6 categories: Emotional Intelligence, Argumentation Quality, Customer Orientation, Objection Handling, Rapport Building, and Results Orientation. Each category is evaluated across 3 dimensions: completion, quality, and tonality.
20 Voices for Natural Conversations
The voice simulation uses 20 different voices that match the persona profiles. Analytical types speak differently than expressive ones - the voice supports training immersion.
Scope of Services
- 103 psychologically grounded AI personas
- 10 personality types (4 basic + 6 mixed types)
- 5 calibratable resistance levels
- Real-time voice conversations
- EQ-CORE scoring with 6 evaluation categories
- 7+ configurable training scenarios
- Dual-AI architecture (conversation + analysis in parallel)
- 20 voices for natural conversations
- 1,212+ unique conversation combinations
- Scientific foundation (TetraMap, DISC)
Challenge
Creating realistic, psychologically grounded customer simulations with natural speech interaction and objective feedback. Each persona must display consistent behavior - an analytical customer type at resistance level 3 reacts fundamentally differently than an expressive type at level 1. At the same time, the feedback system must be objective, conducive to learning, and scientifically grounded.
Solution
TetraMap/DISC-based persona system with 4 basic types and 6 mixed types, combined with 5 calibratable resistance levels. A dual-AI architecture separates conversation and analysis: the conversation AI conducts the realistic customer conversation while the analysis AI evaluates in parallel using our proprietary EQ-CORE model. 7+ training scenarios (cold calling, reactivation, existing customer) with dedicated briefings enable targeted training. 20 voices support immersion.
Results
103 personas with 1,212+ conversation combinations enable varied, realistic training. The scientifically grounded typology ensures comprehensible, consistent customer behavior. Our proprietary EQ-CORE scoring delivers objective, constructive feedback across 6 categories. The dual-AI architecture enables real-time feedback without interrupting the conversation flow.
Ready for your project?
Let's realize your digital vision together.