Three quarters,
one through-line.
The year energy stopped buying AI and started building it — read across our three most recent issues.
Subsurface Intelligence — Q3 2026
Most enterprise AI initiatives do not fail on the model. They fizzle on the data layer, in a repeatable loop: a tactical fix, then a specialist hire that confirms the fix, then a data strategy that produces a deck instead of access, then a fall back to another tactical fix, with data debt accumulating each turn. This whitepaper names that pathology from a real intervention on an Oman carbonate engagement, sets it against the three questions a stalled programme never answers (I can't access the data, I don't trust the data, it takes too long to get to it), and lays out the fix we ran: a practice-and-platform operating model delivered as three ordered four-day workshops on a three-horizon climb, not another model bolted onto the same broken estate.
From AI Customer to AI Builder
Operators start building what they used to buy: domain-native wellbore LLMs, basin-trained seismic foundation models, VeerNet log digitisation, agentic MLOps and drilling agents in the field, carbonate vug and fracture vision, and a first move into lignin.
One Log, Three Tracks
No single method reads carbonate rock, so this issue routes one image log into three tracks: a supervised fracture and bedding detector, classical-CV vug counts, and kriged well-to-well ties. The programme behind it lands in SPE Journal and at EAGE.