Decision intelligence for oil and gas connects production forecasts, asset-integrity findings, field logistics and cash flow so each intervention — pulling a failing ESP, restricting a tank’s fill height, re-routing sand trucks or deferring a pad — is chosen on its combined operational and financial impact and approved by the accountable engineer or executive.
The decisions
| Decision | Signal | Context it needs | Authority | Outcome measured |
|---|---|---|---|---|
| Well intervention | ESP amps up, intake pressure down | Production forecast, rig availability, AFE, deferred oil | Augmented — production engineer | |
| Choke optimization | Gas interference, slugging | Reservoir constraints, facility limits | Automated in bounds | |
| Tank fill restriction / repair | Wall loss at a weld | Thickness history, fill plan, neighboring capacity | Human — integrity engineer | |
| Capital program timing | Price or production change | Price deck, well economics, liquidity | Human — CFO and VP Operations | |
| Field fleet re-routing | Road closure, crew demand | Truck positions, frac schedule, SWD capacity | On-the-loop — dispatcher |
Production → finance → fleet
On the homepage demo, the production forecast drives the 12-month free-cash-flow forecast, and the price slider and pad toggles rewrite the recommended capital decision in real time. Pulling a failing pump early protects roughly 100,000 barrels over 90 days in the illustrative scenario — and that volume flows straight into the cash forecast. Open the production, finance and fleet demo →
Asset integrity decisions on the tank twin
- DetectA magnetic crawler maps shell thickness; loss clusters at weld 7.
- AssessRemaining thickness is compared with minimum required thickness at each fill height.
- OptionsContinue, restrict fill height and confirm with UT, or drain and repair.
- DecideThe integrity engineer approves a 70% fill restriction and a UT confirmation before Friday.
- LearnThe confirmed reading recalibrates the corrosion-rate model for the tank farm.
Production forecasting that changes decisions
A useful production forecast is probabilistic (P10/P50/P90), well-level and scenario-aware: it shows what happens if a failing ESP is pulled now versus left to fail, or if chokes are optimized. On the homepage demo the ‘do nothing’ path loses about 1,450 bbl/d for three months when the ESP on 14H fails. See scenario intelligence.
What to measure
| KPI | Why |
|---|---|
| Deferred production | Barrels lost to downtime |
| Forecast accuracy (P50) | Planning confidence |
| Integrity findings closed | Risk reduced |
| Free cash flow vs plan | Financial outcome |
| Non-productive time | Fleet and crew efficiency |
- Production, integrity, fleet and finance decisions are coupled.
- Twins make integrity trade-offs visible to operations and finance.
- Capital decisions should move with the forecast, not the budget calendar.