AI Deal Watch 4.27.23
AI for health emergency detection, measuring staff morale, and investing in private credit.
Regular updates on the latest VC-backed AI startups. Follow along to stay informed!
OSO-AI, a France-based developer of AI sound recognition tools, raised $10M in Series A funding. The round was led by Innovacom, Novinvest Partners, and Breizh Up, with participation from Cemag Invest Partners.
Problem to be Solved
Some people in retirement homes or hospitals find it difficult to call caregivers, especially if they have fallen or are experiencing a health emergency. OSO-AI has developed a device and sound recognition algorithms that can detect critical situations of vulnerable patients and relieve the workload of healthcare workers.
How They Use AI
The core technology of OSO-AI’s business is their algorithm that has been “trained on several million samples that can identify over 150 sounds like falls, shocks, snoring, vomiting, and words”. No details have been released on their model architecture. Maybe transformer based? For reference, the transformer architecture is the same architecture used to create ChatGPT. Transformers accept a set of inputs (e.g. audio samples) and learn to generate the desired output (e.g. classify the type of sound).
Business Model
The company sells its solution as a service - “a simple all-inclusive offer with no installation costs and no commitment”. OSO-AI’s devices have so far been deployed in around 20 institutions in France and in private homes as part of experiments with key players in remote assistance.
Loopin, an England-based virtual coach for workplace morale monitoring, raised $1.9M in Seed Round funding. The round was led by undisclosed angel investors.
Problem to be Solved
Employers struggle to retain great talent and adopt high-performance practices because they don’t have good insight into how their teams are feeling. Loopin aims to solve this with a combination of AI software implementation, data-led consulting, and expert performance coaching.
How They Use AI
Loopin uses a combination of burnout prediction and OpenAI’s GPT models to provide morale monitoring and coaching. Most likely, they rely on a few indicators of burnout (e.g. sentiment in slack channels) and use OpenAI’s GPT models (probably ChatGPT) to generate a respective coaching message for the managers or individuals.
Business Model
Classic SaaS. The company reports an average cost of $11 per user per month.
Martini.ai, a California-based AI-powered provider of private credit risk solutions, raised $6M in Seed Round funding. The round was led by Neotribe with participation from Rocketship.vc.
Problem to be Solved
Credit liquidity providers are not optimally matched to companies that are great credit risks because risk profiles are difficult to model and the market is relatively illiquid. Martini.ai uses machine learning and quantitative techniques to model the pricing of illiquid assets (specifically, corporate bonds and credit spreads), which helps lenders unlock cheaper credit.
How They Use AI
Martini mentions that they use knowledge graphs and deep learning models. Deep learning models are artificial neural networks that have been trained to perform tasks, e.g. ChatGPT or AlphaGo. Judging by the resume of their CTO, they probably have some pretty sophisticated stuff going on under the hood.
Business Model
The company makes money by charging corporate lenders and companies fees for access to its platform, but pricing details are unclear.