Technology Integration for Petrochemical & Manufacturing Operators in Irving, TX
Irving is 240,000 people inside the city, anchoring the western edge of the Dallas-Fort Worth metroplex with a metro reach of 7.9 million. Las Colinas, master-planned in the 1970s, became one of the largest mixed-use developments in the country and pulled corporate headquarters from across the energy and industrial economy. ExxonMobil's relocation from Irving's Las Colinas to Spring, TX in 2023 reshaped the local executive footprint, but Pioneer (now part of ExxonMobil), Kimberly-Clark, Fluor, Vistra, and dozens of mid-cap industrial companies still anchor the corridor. DFW Airport on the city's western boundary makes Irving one of the most-flown-from corporate addresses in the country.
Irving sits in a different operational position than the corridor towns to the south. The plants and processors aren't on the city's doorstep — Las Colinas is corporate territory, the home of ExxonMobil's headquarters at Hidden Ridge, Pioneer Natural Resources' offices, Kimberly-Clark, Fluor, McKesson, and a dense cluster of decision-making for assets that physically run somewhere else. That changes what technology integration looks like for an Irving-based operator. The conversation here usually isn't about a control system on a plant floor down the street. It's about the gap between Las Colinas spreadsheets and Mont Belvieu, Baytown, Beaumont, and Baton Rouge plant data — between what a corporate planner sees in SAP and what a unit operator sees in DeltaV at 3 a.m. on a Tuesday. MSG closes that gap. We build the integration layer between corporate ERP, MES, plant historians, and the dozen specialty systems that actually run a chemical or manufacturing operation, so the numbers an Irving executive presents on Thursday match the numbers the plant produced on Monday.
The operator pattern in Irving is almost always headquarters-and-remote-asset. The plants aren't here. They're in Mont Belvieu (Enterprise, Targa, Lone Star NGL), Baytown (ExxonMobil's massive complex), Beaumont-Port Arthur (Motiva, Valero, ExxonMobil), Lake Charles (Westlake, Sasol, Phillips 66), and Baton Rouge (ExxonMobil, Dow, Shell). An Irving-based VP of operations is making decisions about plants four to six hours away, and the quality of that decision-making depends entirely on the integration between what corporate sees and what the plant produces. When that integration breaks — when the historian extract is two days old, when SAP PM is out of sync with the maintenance reality, when MES batch records don't match production accounting — the executive ends up flying down to figure out what's actually happening.
MSG is 285 miles southeast of Irving on I-45. We work both ends of the typical Irving engagement: the Las Colinas corporate stack and the actual plant where the data originates. Our team has done integration work in Beaumont-Port Arthur, Baytown, and across the chemical corridor, and we know what an OSI PI AF structure actually looks like at a real refinery — not what a vendor slide says it should look like. That dual perspective is what makes Irving engagements productive instead of frustrating.
Most integration consulting firms that pitch Irving corporate offices are either big-firm advisors who don't write code or vendor-aligned implementers who'll sell you whatever platform pays the highest margin. MSG is neither. We're builders. We've shipped production software for ten years — ServiceStorm, MFGBase, LocalAISource — and we bring that engineering discipline into integration work. When we say a system is going to talk to OSI PI AF, we know what that means at the API level, not just at the slide level.
We also take the Gulf Coast seriously as a single operating environment. An Irving operator with assets in Mont Belvieu, Beaumont, and Baton Rouge isn't running three separate problems — they're running one integration problem distributed across three plant cultures. We've worked across all three corridors. Beaumont to Baton Rouge is a regular drive for our team. We don't fly in for kickoff and Zoom the rest of the engagement.
And we refuse to leave engagements at the dashboard. Integration projects that end with a Power BI report and no operational change are vanity projects. We scope every engagement to include the operational handoff — runbooks, training, observability, and the first 90 days of post-go-live support — so the system is still producing value at month 18 without us.
How the work unfolds
We start by mapping the actual data flow between corporate and plant. That means a week onsite in Las Colinas with the planning, finance, and ops teams, then a week (or several) at the asset itself with the unit operators, MES administrators, and historian admins. We document where data originates, how it's transformed, who touches it, and what gets lost in translation. Most operators are surprised by the answer — there are typically four to seven manual reconciliation steps no one has audited in a decade.
From there we build the integration architecture. SAP PM and PP modules tied to plant-side maintenance reality (often Maximo or a CMMS the plant standardized on independently). MES platforms — AspenTech, Rockwell FactoryTalk, Honeywell Experion PKS, Siemens Opcenter — connected to OSI PI or AVEVA PI System for time-series ground truth. Production accounting (Quorum, Merrick, Allegro for energy-adjacent ops) reconciled against MES batch records. ERP-to-MES integration through the standard ISA-95 levels, but built for what your data actually looks like, not for a textbook reference architecture.
We also build the analytics layer on top — the part most integration projects skip. Power BI or Tableau dashboards that pull from a unified semantic layer, not from seventeen different ODS extracts that disagree with each other. Self-service reporting that doesn't require a four-person BI team to maintain. And alerting that flags the divergence between corporate forecast and plant reality before the monthly close meeting, not during it. Handoff includes runbooks, observability, and a training pass so your team owns the system at month 18.
What's specific to Petrochem & Mfg
Petrochemical and manufacturing integration has three patterns that consistently break Irving-based corporate operators trying to manage remote assets.
First, the OT/IT boundary is where most projects die. The plant historian, DCS, and MES live on a process control network with strict security boundaries. The corporate ERP, BI, and planning systems live on the IT network. Integration across that boundary requires a data diode or unidirectional gateway, a DMZ historian, or some equivalent architecture that the plant's controls engineers will actually approve. Vendors who promise 'real-time integration' without a credible OT/IT story get rejected by the plant's cyber team and the project dies in change control. We've worked through these reviews with controls teams across the Gulf Coast and we design every integration to clear them on the first pass.
Second, batch reconciliation is harder than continuous-process reconciliation, and most petrochemical operators run a mix. Polymer plants, specialty chemical lines, and pharma-adjacent manufacturing run batch logic that has to reconcile against continuous feedstock and utility consumption. The MES batch record, the historian time series, the ERP production confirmation, and the lab QA result all need to converge on a single truth. They usually don't, and the gaps are where margin leaks.
Third, maintenance integration is where the ROI lives but where most projects under-deliver. Tying actual asset condition (vibration, temperature, pressure deltas, performance degradation) into PM scheduling and ERP work order generation is the difference between calendar-based maintenance and condition-based maintenance — and the savings are real, often 15-25% on PM spend. Most operators have the data; they don't have the integration to act on it.
Twelve to eighteen months in, an Irving corporate operator has a single source of truth that connects Las Colinas to the plant floor. SAP, MES, historian, and production accounting reconcile automatically, not through a four-person manual close process. Maintenance is condition-based on the assets where it matters and calendar-based where it doesn't, with a defensible decision behind each. The monthly business review starts with numbers everyone agrees on, not with three teams arguing about whose extract is right.
Things operators ask
Our headquarters is in Las Colinas but our plants are across the Gulf Coast. Does MSG work both ends?
Yes — that's actually our strongest engagement profile. The corporate-to-plant integration gap is where the highest-leverage work lives, and you can't solve it from one end. We spend the first phase in Las Colinas understanding the corporate data needs, planning rhythm, and reporting requirements. Then we spend equivalent time at the actual plant — typically multiple plants if your assets are distributed — understanding the OT environment, the historian and MES setup, and the operational reality. Beaumont to Irving is 285 miles, Baton Rouge to Irving is 450 miles, Mont Belvieu to Irving is 250 miles. We treat all of them as part of the engagement, not as travel exceptions.
We have OSI PI on the plant side and SAP on the corporate side. How do you bridge those without a year-long project?
Through a properly designed integration layer that respects the OT/IT boundary. Our standard pattern is a DMZ historian or unidirectional gateway feeding a corporate-side data warehouse or operational data store, with the SAP integration happening on the IT side against that ODS — not against PI directly. That keeps the controls team comfortable with the OT security posture and lets the IT team manage the corporate-side integration on standard SAP integration patterns. Timeline for a first production integration is typically 12-16 weeks from kickoff, not 12 months. Longer programs come from doing too many plants or too many systems at once, not from the integration itself being slow.
How do you handle the OT/IT security boundary? Our plant cyber team is strict.
We design for that team's approval as a first-class requirement, not as a check at the end. Standard pattern: data diode or unidirectional gateway out of the process control network, DMZ historian or staging environment in between, no inbound connections from corporate to the PCN under any circumstances. We work through the change control review with the controls team and cyber team together, document the data flows, and provide the architecture diagrams the cyber team needs for their own audit. We've cleared these reviews at multiple Gulf Coast operators. The work is more about discipline than novelty.
We've already invested in AspenTech, AVEVA, or a similar MES platform. Are we starting over?
No — we build on what you have. The expensive mistake is ripping out a working MES because a new vendor made a better demo. Our integration approach is platform-agnostic: we design around your existing stack and add the layers that are missing. AspenTech mtell or aspenONE for production execution, AVEVA System Platform or Wonderware, Honeywell Experion, Rockwell FactoryTalk — we've worked across all of them. The gaps are usually in the integration to ERP and the analytics layer, not in the MES itself.
What does ROI look like for a Gulf Coast petrochemical or manufacturing integration?
Three areas where we typically see measurable returns. Maintenance optimization through condition-based PM is the largest single bucket — typically 15-25% reduction in PM spend on assets where the data justifies the change, plus reduction in unplanned downtime that's often a multiple of the direct PM savings. Production accounting and reconciliation cycle time reduction — closing the books faster, fewer manual adjustments, less time arguing about whose number is right. And executive decision quality, which is harder to measure but real: when corporate and plant agree on the numbers, the decisions made off those numbers are better. We scope each engagement against specific metrics up front so the ROI conversation is concrete.
How often will MSG be onsite given our plants are scattered across the Gulf Coast?
Depends on the phase. During discovery and architecture, we're at each plant for at least a week, often more for complex assets. During build, we're in a hybrid pattern — Las Colinas corporate sessions weekly or biweekly, plant-side visits tied to integration milestones and testing cycles. During go-live and stabilization, we're at the plant continuously for the first two weeks and on call through the first end-of-month close. Total onsite days for a typical Irving engagement with three Gulf Coast plants runs 60-90 days over 12 months. We're 285 miles from Irving and we've been to every plant town on the corridor recently. Travel isn't a constraint.
Other Industries in Irving
Tech Integration in Other Cities
Other MSG Services
Running plants from Las Colinas?
Let's close the gap between corporate dashboards and plant-floor reality.