Well intelligence for drilling engineers

Understand a well in its geological and historical context.

Explore nearby wells, compare geological formations, and trace historical drilling events back to their source records — all in one engineering workspace.

Open KINNECTSee how it works
  • Explain, don't assert
  • Geology over absolute depth
  • Human-in-the-loop
Well · formation · offset contextschematic
OVERBURDENSHALECLAYSANDSTONERESERVOIR123
Illustrative — generic formation types only, not operational well data.

How it works

From an active well to a decision you can trace back.

01 · Understand the context

See the active well among its neighbours and its geology.

Start from the well you are drilling. KINNECT places it against the formations it passes through and the nearby wells that share that ground, so a reading is never viewed in isolation.

02 · Compare relevant wells

Line offset wells up by formation, not just by raw depth.

Explore offset wells and compare them in formation-relative terms where the data supports it. When a formation top is missing, that gap is shown as missing — never filled with a guess.

Well A
Shale
Clay
Sandstone
Reservoir
Well B
Shale
not on file
Sandstone
Reservoir

03 · Trace historical evidence

Follow past drilling events back to the record behind them.

Inspect historical events and incidents with their depth, well, and source shown when available. Every item links toward the record that supports it — or an honest empty state when none exists.

  1. 1
    Formation top from source logDocumented
  2. 2
    Event aligned from a comparable wellHistorical
  3. 3
    Model-derived reading at this positionAI
  4. –✕ no record on file

04 · Review before deciding

Inspect the evidence and validate the interpretation.

Model-derived readings are marked as inference, never as fact. An engineer confirms, rejects, or partially accepts each one, so the interpretation is reviewed before it informs a decision.

Suggested interpretationAI

Model-derived — presented as a lead for review, not as established fact.

ConfirmPartially acceptReject
After reviewEngineer

Evidence · provenance · uncertainty

Every figure carries its certainty, and traces to its source.

Every reading in KINNECT carries how far it can be trusted. A figure shows what it claims, where it sits relative to the geology, and which record supports it — so an interpretation can always be checked against its source.

Certainty tiers are distinguished by shape, pattern, and label — never colour alone. AI inference is never presented as established operational fact.

Certainty taxonomy
  • Documented FactDocumented FactPresent in the source dataset record.
  • Historical CorrelationHistorical CorrelationAligned from a comparable well's history.
  • AI InferenceAI InferenceModel-derived; not source-verified.
  • Low-Confidence InferenceLow-Confidence InferenceWeak / sparse support — treat as a lead only.
  • Engineer ConfirmedEngineer ConfirmedReviewed and accepted by an engineer.

Product principles

Three commitments that shape every screen.

Geological context, not absolute depth alone

Wells are read against shared formation tops where the data supports it, so a comparison holds by geology rather than by raw measured depth.

Traceable evidence, not unexplained conclusions

Every reading points back toward the record behind it. Where a record is missing, KINNECT says so instead of presenting a gap as a finding.

Engineer validation, not unquestioned AI output

Model-derived interpretations are marked as inference and left for an engineer to confirm, reject, or partially accept before they inform a decision.

See the ground before you drill it.

Turn scattered drilling history into evidence you can act on — with every finding traced back to where it came from.

Open KINNECTSee how it works