Data warehouse
FrostLogic Explore's Data Warehouse module turns your manuals, procedures and compliance records into a searchable knowledge graph. Hybrid retrieval grounds every answer in your own documents, embedded on our own EU hardware, not a generic energy data management software guessing from the web.
Book a time directlyThe problem
Most energy data management system deployments stop at telemetry: meters, sensors, BMS points. The manuals, commissioning reports, compliance certificates and floor plans that explain what that telemetry means sit in a shared drive nobody searches, a filing cabinet nobody opens, or a folder structure only the person who built it understands. When a technician needs the fire-damper commissioning report at 2am, the ranked queue from the sensors doesn't help.
FrostLogic Explore's Data Warehouse module closes that gap. Upload the manuals, procedures, specs, compliance documents and photos for a site, and Explore turns them into a searchable knowledge graph: a structure that understands a document is about a specific piece of equipment, a specific floor, a specific compliance framework, not just a blob of text. A hybrid retrieval approach, combining semantic search, keyword search and graph traversal, means a question gets answered from the right passage in the right document, not the nearest-sounding sentence in the wrong one.
Every document is embedded on FrostLogic's own hardware in the EU using BGE-M3, not sent to a third-party US API, so the same data-residency posture that governs your sensor data covers your documents too. AI-assisted curation extracts entities automatically and links them into the sites, buildings, equipment and points Explore already models, and anything the model isn't confident about waits in a review queue for a person to confirm before it enters the graph.
Already reading your BMS or your energy meters through energy management software or building management system software? Data Warehouse sits beside that telemetry layer, not instead of it, and the same evidence that proves compliance now sits next to the readings it explains. The result: your building energy data and your building documents, finally in one place.
The difference
| Shared drive | Generic AI chatbot | Dedicated DMS | Data Warehouse | |
|---|---|---|---|---|
| Search | Filename and folder, if you know where to look | Free text, with no grounding in your own documents | Full text search within the document store | Hybrid semantic, keyword and graph search across every document |
| Grounding | None. You read the whole document yourself | Answers can be plausible and wrong | None. The DMS stores documents, it doesn't answer questions | Every answer traces back to a specific passage in a specific document |
| Connects to building data | No | No | No | Yes. Documents link into the same sites, buildings, equipment and points Explore already models |
| Curation | Manual, whenever someone has time | None | Manual tagging and metadata entry | AI-assisted, with a human review gate before anything enters the graph |
| Hosting | Wherever IT put it | Usually a third-party US API | Varies by vendor | EU, on FrostLogic's own hardware |
| Best for | Nothing you need to find quickly | General questions with no stakes if the answer is wrong | Heavy document workflows: versioning and approval chains | Finding the right building fact, fast, with the source attached |
At a glance
Data Warehouse is a module inside FrostLogic Explore, not a separate product. It shares the same tenancy, the same EU hosting and the same security posture as the rest of Explore, applied to your documents instead of your readings.
How it works
01
Add the manuals, procedures, specs, compliance documents and photos for a site, as-is. No reformatting and no manual tagging required to start.
02
Explore extracts entities, embeds every document on our own EU hardware, and links each one into the sites, buildings, equipment and points it already models. Anything the model isn't confident about waits for a human to confirm.
03
Ask a question in plain language. The AI agent answers from your own documents, hybrid search finds the right passage, and the answer comes with its source attached.
Before you ask
Data Warehouse is a module inside FrostLogic Explore. Upload the manuals, procedures, specs, compliance documents and photos for a site, and Explore turns them into a searchable knowledge graph an AI chat agent can query.
No. Data Warehouse isn't a document management system and it isn't a CMMS. It doesn't handle versioning, approvals or work orders. It's a search and retrieval layer over the documents you already keep, wherever they live today, so the right fact surfaces when someone asks for it.
A generic chatbot answers from whatever it was trained on, and it can sound confident while being wrong. Data Warehouse only answers from your own uploaded documents, using hybrid search to find the right passage, and every answer traces back to its source document.
In the EU, on FrostLogic's own hardware. Document embeddings run on-prem using BGE-M3, not a third-party US API, and are stored in PostgreSQL with pgvector. The same data-residency posture that covers your sensor data covers your documents.
Yes. Entity extraction is AI-assisted, but anything the model isn't confident about sits in a review queue until a person confirms it. Nothing enters the knowledge graph unchecked when confidence is low.
Both. As energy data management software it grounds every answer in your own manuals and specs rather than a generic model's guess. It also stores the compliance certificates and audit evidence that sit alongside your BREEAM, LEED or Nordic Swan tracking, so the proof and the readings it explains live in one place.
Tell us what's buried in your building's documents, a commissioning report you can't find or a compliance certificate nobody can locate. We'll show you what Data Warehouse finds on a sample of your own documents. No commitment either way.