Transforming unstructured data into AI-ready data slices
Unstructured data is easy to collect but hard to make usable. We built a context engine that continuously curates it so AI projects don’t stall or drift in production.
Make it easy for engineers and data scientists to turn data into a structured resource that AI models can use
Deasy Labs was founded by a team of data experts who previously built McKinsey & Company’s AI data quality platform. And throughout their careers, they kept running into the same problem: AI projects don’t fail because models are bad, they fail because finding, curating, and maintaining the right unstructured data is incredibly challenging in most organizations.
Our mission is to make it easy for engineers and data scientists to turn this data into a structured resource that AI models can use, saving a lot of steps and rework, and making projects safe enough to move into production.
So we built and optimized a toolchain that automates the entire process of transforming unstructured data into AI-ready slices. It’s a context engine for creating perfect knowledge products that keep sensitive data away from models, enhance RAG and prevent model drift after a project is in production.
We're proud to be backed by a group of visionary investors who believe in our mission.
-
Reece Griffiths
CEO, Co-founder
As a product manager within McKinsey’s data & AI practice, QuantumBlack, Reece previously led the build of McKinsey’s flagship AI data quality product. Prior to this, Reece spent several years serving McKinsey’s Fortune 500 clients on data transformation topics, and holds a masters in engineering from the University of Cambridge.
-
Leonard Platzer
CTO, Co-founder
Leo began his engineering career at a chatbot startup, before moving to Amazon and then eventually QuantumBlack - where he was the Machine Learning engineer on McKinsey’s award-winning data quality platform. Leo holds a degree from the National University of Singapore.
-
Mikko Peiponen
Chief Architect, Co-founder
Mikko spent more than 6 years serving global leaders on data science topics within McKinsey’s data and risk practices, before leading the build of McKinsey’s flagship ML-based data quality product. Mikko holds a graduate degree in quantitative finance from MIT.
What we do
Deasy Labs automates every step of turning unstructured data into structured, easy to find and manage data. From data discovery to tagging, filtering and AI enrichment.
Our context engine is designed to be the unstructured data backbone for organizations, allowing them to skip the traditional, manual and slow processes to gain access to clean, actionable data. And our platform does it all—30x faster and with more accuracy than open-source solutions.
Why Deasy Labs?
Most teams start by doing this themselves, inefficiently stitching together general-purpose LLMs or waiting on a domain expert to define taxonomy. It can work, but it’s expensive, slow and hard to keep consistent as data changes. The bigger problem shows up later.
-
Ungoverned knowledge decays
Setting up your knowledge base for RAG or agentic systems as a one-time exercise won't cut it here. It should be a living system. As unstructured content evolves, ungoverned knowledge begins to decay.
-
Data quality degrades
Duplication creeps in, metadata drifts, sensitive data slips through, and retrieval quality quietly degrades. The impact often isn’t immediate, but it compounds over time, driving down accuracy and pushing AI costs up month after month.
-
Deasy Labs removes that friction
Our automated tagging, filtering, and enrichment pipeline connects directly to your source systems. As new content is created, it’s automatically identified, classified, and routed to the right AI use cases, keeping your data slices continuously up to date over time.The result is faster and more accurate retrieval, less rework and fewer surprises in production.
Deasy Labs was acquired in 2025 by Collibra, the global leader in enterprise data and AI governance—bringing Deasy’s unstructured AI capabilities into a mature, trusted governance and catalog ecosystem. This means greater scale, long-term stability, and native integration into the broader enterprise data stack.
