If you've ever walked a refinery control room, you've seen APC in action. Advanced Process Control. The system that keeps a continuous process unit running at its optimal operating point — balancing yield against safety against energy against throughput — in real time, thousands of times per minute. Nobody touches the setpoints manually. Nobody argues about priorities. The system does what humans can't: optimize a many-dimensional problem continuously.
Upstream and midstream operators don't use APC language. But the principle — balancing competing objectives continuously in real time — is exactly what's missing from field work management.
Your refinery runs APC. Your field runs on Excel. Why?
The Problem With How Work Gets Prioritized Today
Take a typical morning at a 2,000-well operator. The superintendent has:
- 200+ overnight SCADA alarms, most of them noise
- A maintenance backlog that includes 30 PMs due this week and 12 critical repairs deferred from last week
- 5 pumper routes that need to be assigned — and the routes haven't been updated since 2019
- 3 contractor crews on-site for workovers — one needs a JSA reviewed before they can start
- Weather — a storm front coming through by 2 PM that will delay 4 pad visits
- Commodity price just moved $3 overnight, making some previously uneconomic interventions now in-the-money
The superintendent has to rank all of this, assign crews, and get the plan to the field by 6 AM. Every morning. For the rest of their career.
How do they actually do it? They rank it badly. They rank by whatever alarm is loudest, whatever crew is closest, whatever repair is due "this week." They don't rank by dollar impact because they don't have time to calculate dollar impact. They don't rank by risk because risk scores are qualitative. They make the best decision they can with the information they can hold in their head.
That's not their fault. It's the nature of an N-dimensional problem solved by humans. The best superintendent in the basin is still making a 70% right decision on average — because nobody can balance 6 variables in real time.
What APC Does That Humans Can't
APC doesn't pick between "yield" and "safety." It optimizes for yield subject to safety as a hard constraint. It doesn't pick between "energy" and "throughput." It finds the Pareto-optimal point where energy is minimized subject to throughput being maximized.
This is the mathematical trick: constraints vs. objectives. Safety is never a weight. It's a hard constraint. The system can never violate safety limits, no matter how much yield you could gain. Inside the constraint envelope, the system optimizes whatever you're maximizing — yield, throughput, cash flow.
Applying APC to Field Work Management
At WorkSync we built an AI agent that does the same thing for field work. Every task gets three scores:
1. Economic Impact Score
Dollar impact per hour of inaction. Production at risk × commodity price × working interest × probability of intervention success.
A well that's down 50 BOE/day at $55 oil with a 90% chance of a successful repair on a 100% WI: $2,475/day of deferred revenue. A PM that's due but not critical: $0 of deferred revenue. Ranking by this score puts the real cash-flow work first.
2. Risk Score (Hard Constraint)
The Hazardous Task Index. Composite score of asset hazard profile × assigned crew competency × environment (H2S, weather, terrain) × historical incident frequency.
Any task above the risk threshold either requires a senior crew, enhanced monitoring, or a re-scheduled window. The agent cannot choose to override the risk score. Safety is a hard constraint.
3. Proactive / Reactive Balance
The operational ratio. Pure reactive operations burn out the team and miss scheduled maintenance, which creates bigger reactive events later. Pure proactive operations ignore the 200 alarms that came in this morning — which is expensive in the short term. The agent targets a mix (typically 60/40 reactive/proactive on a steady day, 80/20 after a storm, 40/60 during turnaround prep). The mix is configurable by leadership.
The Output
The agent takes SCADA + production + field data + weather + commodity price + crew availability + qualifications + tank levels + hazard profile + incident history — and produces, for every crew, a ranked list of work for the day. Delivered to the truck cab by 6 AM. Re-ranked every hour as conditions change.
The superintendent doesn't disappear. They review the plan, make the judgment calls, handle the edge cases, manage the relationships. What they don't do is rank 200 SCADA alarms by dollar impact at 5:30 AM, every morning, for 30 years.
This is the APC of work management. Cash flow as the objective. Safety as the constraint. Proactive-reactive balance as the control setting. The agent optimizes continuously.
The Results
4,000+ wells. 18 months in live production across three basins (Western Anadarko, Permian, Wyoming). 15%+ sustained free cash flow uplift. TRIR from 1.8 to 0.3. 35% fewer site visits. 40% OPEX reduction.
None of those numbers came from better technology. They came from continuously optimizing a problem that humans can't continuously optimize.
Want to see the APC of work management running on your data? Apply for the Free Pilot. We'll build a sample 6 AM plan from your SCADA — live, on the call. Integration takes under a week on today's platform.
