Collaborated with the psychology student founders to understand user psychology, corporate mental health requirements, and German compliance standards. Defined the AI conversation architecture and scoped the MVP to the features that would validate the core hypothesis.
Mindflex
A secure web & mobile solution helping companies support employee mental health through AI-driven insights, guided exercises, and real-time support.
Overview
Mindflex is a digital mental wellbeing platform built for German employers who want to support employee mental health without the cost or scheduling friction of one-on-one consultations. The platform combines GPT-4o-driven conversation flows, OpenAI Whisper voice-to-text, OpenAI TTS voice replies, AI-generated adaptive assessments, and guided stress-management exercises. An employer-facing analytics dashboard surfaces team-level insights without ever exposing individual employee data — a hard constraint for the German market. Built on React.js, Nest.js, CosmosDB, and Azure, Mindflex shipped both web and iOS apps on time and runs inside German data-protection requirements while keeping AI-driven voice and text interactions responsive in real time.
At a glance
Key facts
- GPT-4o + Whisper + TTS
- OpenAI conversational AI, speech-to-text, and voice replies in one experience
- Web + iOS
- Browser app for employers and admins, native iOS app for employees
- German privacy compliant
- Hosted on Azure with CosmosDB to meet strict German employee-data protection requirements
- Privacy-first analytics
- Employer dashboard surfaces team trends without exposing individual employee data
- 6 core modules
- Conversation flows, adaptive assessments, voice-to-text, exercises, recommendations, admin
About
About this project
Mindflex is a digital mental health platform designed for modern companies that care about employee wellbeing, productivity, and long-term resilience. The product combines AI-powered mental health support, evidence-based practices, and intuitive UX to help teams manage stress, anxiety, and emotional wellbeing in a scalable and confidential way.
The Challenge
The challenge
- 01
The German market lacked automated mental health solutions that could scale across corporate teams. Existing solutions were either expensive one-on-one consultations or generic wellness apps with no employer integration. The founders needed to validate a new product category while maintaining strict German data privacy standards on a startup budget.
- 02
Designing a mental health application that balances empathy with simplicity. Every interaction needed to feel supportive, never clinical or overwhelming. The AI-powered conversation flows had to feel natural while producing genuinely helpful assessments and recommendations. The admin dashboard needed to surface useful team insights without ever compromising individual employee privacy.
- 03
Integrating multiple AI services (OpenAI Whisper for speech-to-text, GPT-4o for conversational AI, TTS for voice responses) into a seamless, natural user experience. The platform needed to run securely on Azure with CosmosDB, meeting German data protection requirements while supporting real-time AI processing for voice and text interactions across both web and mobile.
Deliverables
What we delivered
-
AI-generated adaptive assessments
Mental-wellbeing assessments that adjust questions based on prior answers, producing personalized snapshots without static questionnaires.
-
AI-powered mental health conversation flows
Empathetic, non-clinical chat experience powered by GPT-4o, designed to feel supportive in the moment and route users to relevant exercises.
-
Admin & analytics dashboard for employers
Aggregated team-level insight into engagement and wellbeing trends, with a hard rule that no individual employee data is ever surfaced.
-
Guided exercises and stress-management tools
A library of structured exercises — breathing, reframing, grounding — that users can complete in minutes between meetings.
-
Personalized wellbeing recommendations engine
Suggests next-best content based on assessment results, conversation history, and usage patterns — replacing a generic content list.
-
Voice-to-text implementation (OpenAI Whisper)
Whisper transcribes spoken input into the conversation flow so users can talk through stress instead of typing it.
Results
The results
-
Zero
Critical security or privacy incidents during testing and early usage - meeting German data protection standards
-
3×
Higher engagement with guided mental health exercises compared to initial prototype benchmarks
-
100%
On-time launch of both web platform and iOS mobile app, covering employee-facing and admin use cases
-
Up to 40%
Reduction in time required for users to access relevant mental health support content
The Process
How we worked
Each engagement runs the same defined workflow — discovery, design, build, and handover — adapted to the project's specific scope.
-
Step 1 / 4 -
Step 2 / 4 Created a calm, accessible, and trust-driven interface designed for sensitive user journeys. Every screen prioritized emotional safety - soft colors, clear progress indicators, and gentle transitions. The admin dashboard was designed to surface insights without any individually identifiable data.
-
Step 3 / 4 Built secure web and iOS applications with multi-model AI integration - Whisper for voice input, GPT-4o for conversational flows and adaptive assessments, and TTS for voice responses. Deployed on Azure with CosmosDB for German-compliant data residency.
-
Step 4 / 4 Validated usability with target users in a corporate environment, stress-tested AI conversation flows for accuracy and sensitivity, verified data privacy compliance, and launched to production with monitoring for AI response quality.
Testimonials
What clients say
“I was impressed by the way LNOKS worked together — they also have fun at what they do and that's what I love to see. They were from the beginning very professional and delivered every milestone on time.”
Estimate Your Project
Let's build it.
Tell us about the project. A founder replies within one business day.