AI & Automation / Platforms & SaaS

Meeting Metrics

An in-house SaaS product for AI-supported meeting documentation. Transcription, structured summaries, and flexible processing on Swiss infrastructure or external models.

Client

Internal SaaS Product

Year

2024

Services

Product Design, Web App, Backend, AI Infrastructure, Desktop & Mobile

Website

meetingmetrics.ai/ →

Bild und Video mit Hilfe von KI erstellt.

Overview

With Meeting Metrics, we built our own SaaS product for transcribing, structuring, and documenting meetings in a privacy-conscious way.

The starting point was a very concrete internal need: we wanted reliable meeting documentation without manually reworking notes every time or sending sensitive content to external services by default. That led to a product now used in healthcare, public administration, education, and SMEs, with both registered users and paying customers.

What we built

Meeting Metrics is not a single AI feature. It is a full product with multiple surfaces and a flexible processing architecture behind it.

Transcription and speaker recognition

Audio files are processed automatically, transcribed, and enriched with speaker attribution so recordings become structured working material.

Structured summaries with template control

Summaries and meeting records can be shaped through templates for different meeting types and teams instead of forcing a rigid default output.

Web, desktop, and mobile

In addition to the web app, we built and released a desktop app for macOS and Windows as well as mobile apps for iOS and Android.

Product, brand, and content end to end

From product design, logo, and website to backend, knowledge base, and blog, we built the full product foundation ourselves.

Technical implementation

The real complexity was an end-to-end pipeline that stays reliable from upload to processing to useful output across multiple platforms.

Processing on Swiss infrastructure

The web server runs in Switzerland. We also operate our own infrastructure for transcription and speaker recognition to keep data flows under deliberate control.

Local or external model choice

For summaries, we support both external models such as OpenAI and Gemini and open-source LLMs running on our own infrastructure.

Reliable audio and AI pipeline

The challenge was not just the AI itself, but a robust pipeline from upload, queueing, and processing to a clear output that real users can trust.

Product development beyond a prototype

Meeting Metrics has been developed over time as a real product and expanded so it remains useful across user groups and recurring workflows.

The Results

Meeting Metrics started from an internal need and became a production SaaS product with registered users and paying customers across multiple industries.

1,000+

Users

3

Platforms

B2B

Customers

CH

Hosting

Technology

Laravel Livewire PostgreSQL pgvector OpenAI Gemini Open Source LLMs Electron React Native Tailwind CSS Stripe

Next Project