KC-01Foundational
Technical Knowledge Assistant (RAG)
Ask a question about your own field in Persian, Arabic, English or Spanish and get an answer with the report and page it came from.
Knowledge & CopilotAI
KC-03FrontierWave 3AI
Multimodal Subsurface Assistant
In one sentence
Understands log images, maps and cross-sections alongside text, so you can ask a question about a figure rather than only about a paragraph.
The problem
A large share of subsurface knowledge lives in figures: log displays, structure maps, seismic sections, core photographs. Text-only retrieval is blind to all of it, and the figure caption is rarely enough.
How it works
Vision-language models index figures alongside text, with domain-specific handling for log tracks, map annotations and section displays. Answers reference the specific figure and region, and the system is deliberately conservative about reading values off a plotted curve.
Data required
- Figure-rich document archive with adequate scan resolution
- Annotated figure examples for domain tuning
- On-premise GPU capacity for vision model inference
- Expert evaluation set covering figure-based questions
What you get
- Answers referencing specific figures with the relevant region highlighted
- Figure search across the archive by content rather than caption
- Cross-references between figures and related text sections
What it cannot do
Reading precise numeric values off a plotted curve is unreliable and we do not present it as a feature — the model locates and describes figures rather than digitising them. Scan quality is the binding constraint: a low-resolution scan of a log display is unreadable to the model for the same reason it is unreadable to a person.
Every capability page states a real limitation. If a vendor cannot tell you where their product stops, you are not being told the truth about where it starts.
Related capabilities
- SR-06Frontier
Deep Learning Seismic Interpretation
Picks faults and horizons across a full seismic volume in hours, leaving the interpreter to judge rather than to click.
Subsurface & Reservoir IntelligenceAI - SR-04Intermediate
Digital Rock & Core Image Analysis
Measures porosity, grain size and fracture density from core photographs and CT scans instead of from a description written twenty years ago.
Subsurface & Reservoir IntelligenceAI - KC-02Frontier
Domain-Specific Language Model
A model further trained on petroleum engineering text, so it uses the field’s vocabulary correctly in Persian and Arabic as well as English.
Knowledge & CopilotAI