Make complexity queryable.

Research takes shape when it meets a real problem.

Ando Lab grows out of work on enterprise knowledge: distributed information, meanings to reconstruct and questions spanning several systems. Ando Engine takes shape through this work: our first product, now in private beta by invitation.

Engine is a concrete starting point. It lets us take an architecture out of the lab, confront it with use cases and build a broader research direction. We want every product to be an opportunity to understand which problems deserve a new solution.

UKS, TRAMA and Lakehouse are our foundation.

A business is made of entities, events, meanings, documents and operational data. Their value also emerges through connections: a clause changes the meaning of an amount, a relationship links a document to a process, a version changes what can be considered valid.

UKS, TRAMA and Lakehouse form the foundation on which we develop the engine: organizing knowledge, navigating evidence and querying data. From this base, we want to grow our own technology that connects structure, sources and computation.

Querying a business means crossing several worlds.

Enterprise systems accumulate large volumes of data, while the questions people ask cross boundaries that applications keep apart. Structured data, documents, the history of decisions and business relationships are needed together.

We see Big Data and Lakehouse architectures as a central research territory: making heterogeneous information queryable, connecting meaning and computation, bringing a person’s question closer to the complexity of the systems that can answer it. The scale of the problem includes volume, variety and the depth of relationships.

Between a question and an answer lies an entire system to develop.

Understand the question and what it leaves implicit. Read a schema and recognize relevant relationships. Choose where to look, which sources to open and which evidence to connect. Execute queries and calculations against the sources when the answer requires data.

Our research program addresses this path as a whole: from understanding to navigation, from execution to checking evidence and completeness. We want to develop proprietary technologies across the different stages of resolution, making them work as parts of a coherent system.

Technological independence grows through choosing where to do research.

Knowledge representations, navigation strategies, query planning, specialized models, adapters and embeddings are possible areas of work. We are interested in developing capabilities for specific enterprise problems and understanding where they could deliver a concrete advantage.

For specialized models, the direction is learning to navigate and solve: read structures, follow relationships and propose paths. Enterprise knowledge remains in the connected systems and sources, with its updates and permissions; we do not treat it as content to embed into model weights through fine-tuning. These possibilities remain research until the work supports their use in the product.

We want to make questions tractable that currently require too much fragmented work.

A useful answer must be traceable to its supporting evidence. Completeness makes sense relative to the question and available sources: conditions, missing information and unresolved questions are part of the picture.

Our ambition is to grow reliability and efficiency together. Cost per correct answer is a criterion connecting quality, latency and the work required to reach a result. In Research, we describe how we study these questions, with evaluations and limitations kept distinct from the Manifesto’s vision.

We want Ando Lab to become a place where enterprise research can move forward.

Our ambition is to build a research and excellence center dedicated to systems for querying and understanding enterprise knowledge. A place connecting research, engineering and industrial problems, giving continuity to ideas that deserve further development.

This is a horizon to build over time: through technical depth, recognizable contributions and products that put ideas to the test. We want to create the conditions for different skills and perspectives to meet around important problems.

Start with a concrete problem and work so the contribution can travel far.

The complexity of enterprise knowledge crosses industries and countries. We want our work to have that openness too: engage with international research, build a recognizable technical voice and create opportunities for collaboration grounded in real problems and results.

Ando Engine opens the first chapter. Ando Lab is the broader path we want to build: our own technology, new products and research capable of nourishing both. A concrete first step, looking toward what we can make possible tomorrow.