Eight models.
One difficult decision.

Can AI turn imperfect surveillance data, conflicting instruments, barrier history and production pressure into a defensible well integrity decision?

MODELS08
WELLHALCYON P-17
OUTPUTFIRST PASS
MANUAL FIXESZERO

Decision quality,
not dashboard theatre.

Each model received the same self-contained evidence pack for HALCYON P-17, a fictional gas-lifted North Sea producer. The well remains inside absolute pressure limits, but annulus behaviour, integrity-test history and operating changes point towards possible developing barrier impairment.

At the same time, production is seeking another 400–600 barrels per day. The model must decide whether to continue, restrict, gather evidence or escalate—while being explicit about uncertainty.

CONTEXTVALIDATEDETECTCOMPAREINTERPRETHYPOTHESISEASSESS BARRIERSASSESS RISKDECIDEACTVERIFYLEARN
01 / CONFLICT

Instruments disagree

The evidence preserves a 27.1–31.0 barg uncertainty range rather than hiding it.

02 / HISTORY

The trend matters

Operating changes, overdue surveillance and a worsening passing-valve leak test must be read together.

03 / CONTROL

A restriction is active

No production increase without well-integrity approval; gas lift remains capped at 4.3 MMscf/d.

Inspect all eight.

These are the working applications as produced. No design harmonisation, no manual correction and no declared winner yet.

IMPORTANT

HALCYON P-17 and its data are fictional. These applications are benchmark demonstrators, not operational well-integrity advice. They do not replace an approved well integrity management system, competent well-integrity authority, formal risk assessment, barrier standards or company procedures.

Compare with Benchmark 01: HSG253