What changes when operations teams use Offset for 30 days.
Early-access results from process operations and emergency coordination pilots. Small cohort, honest data.
Continuous process unit. UK chemical manufacturing site.
Continuous-process site with multiple parallel distillation columns. Offset connected to 4 data sources: DCS historian, shift log system, maintenance work order queue, and environmental monitoring feed. Pilot ran 38 days.
Before Offset
The team held 3-hour handover briefings because verbal transfer was the only way to pass along significant events from a shift. Incoming operators often used the first 20 to 30 minutes of each shift to reconstruct context instead of monitoring the process. Post-incident reviews linked three near-miss events in 18 months to context lost during shift handover.
Results after 38 days
- Average handover time fell from 3.1 hours to under 45 minutes. The recommendation log took over most of the verbal context handoff.
- Operators followed or marked 91% of recommendations as accurate in post-shift assessments. They overrode 9%; each override included a written rationale, which the operations manager saw as a process improvement on its own.
- Offset identified one potential deviation before the alarm threshold. Its recommendation flagged a reflux ratio adjustment 14 minutes before the first DCS alarm would have fired, using a feed composition trend correlated with a historical incident signature.
"The incoming operator now walks in, opens the recommendation log, and has context in under 5 minutes. That was not possible before." Operations Manager, UK chemical manufacturing, site anonymized by agreement
Multi-incident emergency response. Regional coordination exercise.
A regional emergency coordination team tested Offset against 60 days of historical incident records in tabletop format, then used it live in two regional exercises with multiple simultaneous simulated incidents.
The synthesis gap
In multi-incident scenarios, the team's problem was not access to information. Its ICS platform already surfaced most needed data. The difficulty was synthesis. With three or more situations unfolding at once, the interval from "situation is escalating" to "resource allocated" was too long because one or two senior commanders carried the full prioritization burden.
Evaluation results
- In exercises with Offset recommendations, average time from multi-incident onset to the first resource allocation decision was 2.1x faster than for the same team in equivalent exercises without Offset.
- The team's after-action coordinator reviewed the recommendation log from the historical tabletop evaluation. Their assessment was that it could cut post-incident reporting time by an estimated 60 to 80% across the 60 incidents in the historical set.
- The team most consistently valued Offset's cross-reference of wind direction with resource positioning. It exposed a dependency commanders called "obvious in hindsight, harder to maintain in real time."
"We already had the data. The problem was always the synthesis. Offset gives us one recommended starting point to argue with, which is faster than building the picture from scratch." Coordination team lead, regional emergency services, anonymized by agreement
Pilot cohort results
N=4 sites, Q4 2024 to Q3 2025. Small sample, not a controlled study.
All metrics are from early-access pilots. Site names and specific operational details are withheld by agreement. We do not publish identifiable operational data.
Evaluate Offset for 30 days.
We can scope a pilot in your operational environment: one process unit, 3 to 5 data sources, and 30 days. You decide whether the recommendation quality is useful for your context.