Case studies

For a home care company, we conducted a concept validation of an AI-driven home care platform. Our AI Product Blueprint process confirmed project viability and produced a detailed implementation roadmap. We developed a comprehensive AI implementation budget, providing cost estimation for full-scale development and deployment. The process culminated in a clickable prototype, completed after extensive stakeholder workshops, offering an interactive demonstration of the platform’s potential.

For a client overwhelmed with emails and complex scheduling, we deployed an AI-based virtual assistant. It extracts crucial information from emails, analyzes their content, and presents an immediate overview of meeting statuses. By condensing lengthy, multi-threaded conversations into concise updates, our solution saved up to 50% of the time spent on scheduling meetings compared to manual methods.

For a client struggling to track model performance, we deployed MLflow to organize models, compare approaches, and monitor metrics. We refactored code, optimized queries, and covered nearly 30% of the code with business logic validation tests, improving its reliability and scalability. We established clear short-term and long-term goals and implemented a collaborative sprint-based workflow, enabling the client’s engineering team to collaborate effectively and achieve business goals.

An AI-based student assistant designed to act as a 24/7 personal academic advisor, providing personalized, accurate information to help students make informed educational choices.

• Optimizing architecture with a scalable, cloud-based solution;
• Building a search engine with FaceNet inspired architecture;
• Auto-scaling for 75% fewer tasks running in quieter parts of the day

A scalable and 100% reusable cloud solution that simplifies decision-making processes for 8,000 shops (supported with explainable models), reduces food waste and increases revenue.

An algorithm that recognizes brand logos in any online image, allowing our clients to monitor brand reach and engagement. The system works with 60% accuracy, making it the most effective algorithm of its type.

A solution that automatically detects the most engaging content from across the web for any given category based on a defined taxonomy.

A model for accuracy of popularity prediction model reaching over 0,7 (F1 score) when classifying the post to the specific category.

A solution that supports preparing TV program descriptions. We built a hybrid extraction method based on regular expressions, linguistic rules, and statistical language models, using dependency analysis to decompose templates.

A simple app powered by machine learning and neural networks to automate tax filing with accuracy over 99,2%

A platform that automates the process of invoice reconciliation. The tool removed the manual effort involved in classifying invoices, helping the accounting department better track costs.

Suguard is a mobile app that uses state-of-the-art AI to personalize diabetic therapy and help patients in their everyday management of this chronic disease. It aims to predict and recommend actions to stabilize blood glucose levels, decreasing the risk of complications, thus reducing treatment costs.

A web application dashboard that presents data from online marketing campaigns run on different platforms (inc. Facebook, Google AdWords, BDM, Adfrom).

Improved collection of child maintenance supported by  community of 6,000+ parents and legal experts.

A tool that reviews how specific target groups use Facebook. The algorithm analyzes the social network to glean information around what type of imagery a user group prefers. Then, it leverages the insights to deliver a hyper-specific look across different websites.