# DNAi

> DNAi (DNAI s.r.o.) is a Prague AI company that builds digital twins and AI platforms for manufacturing and supply chain, healthcare and life sciences, energy and utilities, banking and insurance, and telecommunications — from research to production. Founded in 2018 by Pavel Wimmer and David Kadleček, PhD. Part of DNAi Group (with Tekies): 260+ people, 50+ AI experts, 15 PhDs in AI. ISO/IEC 27001:2022 certified delivery.

A digital twin is a software component that mimics the behavior of any physical system or product. Unlike physical systems that are closed, have hidden states and give partial, noisy and delayed information, the twin is open and transparent, suppresses measurement noise, estimates hidden states, predicts future development, keeps data available during outages and issues alerts on its own.

DNAi's platforms are built on this method, not on a wrapper around a public language model. The core is a mathematical model of the client's system (Bayesian data assimilation, Kalman and particle filters, stochastic processes, probabilistic forecasting, optimization) trained on the client's data; language models are an optional layer for interpretation. Deployment: cloud, hybrid or on-premise.

## Digital twin

- [Digital Twin — principles and implementation](https://www.dnai.ai/digital-twin): what a twin is, how DNAi builds one in iterations, worked examples (patient, production line, building, bank customer), technology stack, FAQ

## Platforms

- [Platforms overview](https://www.dnai.ai/platforms)
- [MyIO](https://www.dnai.ai/myio): demand forecasting and inventory optimization; 35% better forecast accuracy for a food distributor with 20,000+ products; available on Microsoft Azure Marketplace (https://marketplace.microsoft.com/en-us/product/dnaisro1733320242033.myio)
- [COGS](https://www.dnai.ai/cogs): predicts manufacturing cost in seconds instead of 3–12 months, with <0.1% deviation from official calculations; built and proven with a European generics manufacturer, applicable to any manufacturer
- [Legacy Modernization](https://www.dnai.ai/legacy-modernization): AI agents reverse-engineer AS/400, COBOL, PL/I, LANSA and mainframe systems into a complete specification with tests and migrate them with provably the same behaviour; 50–90% lower migration cost in early results
- [Computer Vision](https://www.dnai.ai/computer-vision): AI visual quality control for manufacturing; 99% of defects detected at Dormer Pramet, defects down to 10 μm, scrap rate down 60%+
- [Digital Assistant](https://www.dnai.ai/digital-assistant): answer clients, augment your workforce, automate the work with agentic AI — chat and voice on one knowledge base; powers EVA for the Czech Ministry of Labour (up to 6,000 calls a day)
- [Floxgen](https://floxgen-platform.com/index.html): genomic data platform — processing, aggregation and variant interpretation; 4,000+ genomes aggregated for the Czech Genome Portal (IMTM); Floxgen Aggregator is available on Microsoft Azure Marketplace

## Where to start

- [AI Capability Assessment](https://www.dnai.ai/ai-assessment): a few weeks for one team, up to a few months for a large division; use cases mapped and scored, value × feasibility portfolio, deep dives with PoCs. Reference: a European automotive manufacturer, 130+ use cases in Sales
- [Digital Twin Readiness Check](https://www.dnai.ai/digital-twin#readiness): one production line or process, up to three data sources, one month, fixed price

## Industries

- [Industries overview](https://www.dnai.ai/industries): healthcare and life sciences, manufacturing and supply chain, energy and utilities, banking and insurance, telecommunications

## Case studies

- [All case studies](https://www.dnai.ai/case-studies)
- [Dormer Pramet — computer-vision quality control](https://www.dnai.ai/case-pramet): 99% defect detection (92% required), defects down to 10 μm
- [Honeywell — AI for turbocharger design](https://www.dnai.ai/case-honeywell): compressor efficiency predicted from the design geometry in seconds, with accuracy comparable to CFD
- [Honda (with BellaDati) — IoT monitoring of lawn mowers and heavy equipment in Japan](https://www.dnai.ai/case-honda): operating results, maintenance and theft alerts, abnormality detection from vibration and position data, remote control
- [Food distributor — demand forecasting for 20,000+ products](https://www.dnai.ai/case-food-distribution): 35% better accuracy, 15% lower inventory
- [European parcel carrier — volume forecasting](https://www.dnai.ai/case-parcel-forecasting): 45 million parcels a year forecast down to postcode level, 3.9% average global error
- [EVA — AI assistant for the Czech Ministry of Labour (MPSV)](https://www.dnai.ai/case-eva): voice and chat on one knowledge base, up to 6,000 calls a day; Digital Future Awards winner
- [Generics manufacturer — manufacturing cost in seconds (COGS)](https://www.dnai.ai/cogs): <0.1% deviation; Digital Future Awards 2026 finalist
- [IMTM — Czech genome database with Floxgen](https://www.dnai.ai/case-imtm): 4,000+ genomes, development time cut by up to 90%
- [AKESO — healthcare shift planning](https://www.dnai.ai/case-akeso): 21 entities, 2,000+ staff; schedule creation from 24 hours to 1
- [Pandemic simulation with CERGE-EI](https://www.dnai.ai/case-pandemic): agent-based model on a synthetic population

## Company

- [About DNAi](https://www.dnai.ai/about)
- [Careers](https://www.dnai.ai/career)
- [Contact](https://www.dnai.ai/contact): dnai@dnai.ai, +420 721 201 219, U Nikolajky 1097/3, 150 00 Prague 5, Czech Republic
- [Privacy policy](https://www.dnai.ai/privacy-policy)
- [Tekies (DNAi Group)](https://tekies.eu/)
- [Beyond AI podcast](https://beyond-ai.simplecast.com/)
- [LinkedIn](https://www.linkedin.com/company/dnaigroup/)