# Aegis: Simio builds simulations from FactoryLogix MES data

By MarketScale Newsroom · Published 2026-11-19 · Industrial IoT on MarketScale
Canonical: https://www.marketscale.com/industries/industrial-iot/aegis-builds-simio-factory-simulations-from-live-mes-data-instead-of-by-hand

> Simio's four new tools build schedules from live FactoryLogix data and give each order an on-time probability. Here is what planners should test.

## Key points

- For plants already on FactoryLogix, simulation-based scheduling can start from existing MES data—so routing, bill-of-material and WIP record quality strongly affects the schedule.
- Working in ranges lets Simio APS put an on-time probability on each order before a date is promised, a different input for customer commitments than a single capacity number.
- The fair test for neural-network routing is the plant's best existing dispatch rule, not random assignment; if the network does not clearly win, the simpler rule stays.

In many plants, late material or a machine that goes down still ends the same way: someone reopens the spreadsheet and rebuilds the production schedule by hand. Aegis Software's answer, released November 18, is a schedule produced by a simulation that nobody on the plant staff has to build.

Simio, the simulation company Aegis acquired and now operates as a subsidiary, has introduced four products: Simio APS, Simio Essentials, Simio Essentials Plus and Simio Virtual Factory Designer. According to Assembly Magazine, each one combines Simio's discrete-event simulation engine with adaptive AI. A planner can preview how the factory is likely to perform before committing to a plan, get a plain-language answer on why a given order is going to run late, and replan in near real time.

Simulation-based scheduling is not new. Dennis Pegden, Aegis's chief product officer for APS and simulation, said neural networks have been making decisions inside Simio for years. The change is where the model comes from. For customers of FactoryLogix, Aegis's manufacturing execution system (MES), Simio generates the simulation from live production data and refreshes it on demand, which the company says replaces the weeks of manual setup simulation has traditionally needed.

If that claim holds on the floor, the weeks of hand-built modeling that kept simulation-based scheduling in a specialist's corner become a refresh button for any plant already running FactoryLogix.

## A twin generated from MES records

The inputs are records an MES already keeps: orders, routings, bills of material, resources, materials and work in progress. Aegis's Business Wire announcement describes a company founded in 1996 that sells MES software to discrete manufacturers in aerospace, defense, electronics, medical and automotive, and supports more than 2,200 factories across North America, Europe, Asia and Latin America. That installed base is the audience for the two no-configuration tiers, which generate a schedule in a click.

The design has a practical consequence. A simulation built from MES records can only be as faithful as those records. A routing with the wrong cycle time, or a bill of material nobody updated after an engineering change, would flow straight into the twin. Aegis appears to anticipate this: Simio Essentials checks for the data it needs before it runs.

For a plant weighing the tools, the first pilot task is an audit of routings and bills of material on the lines in scope. CEO Jason Spera said customers can grow from Essentials to full APS and factory design without revisiting the foundational data, so cleanup done for the entry tier would carry forward to the next one instead of being redone.

## Event by event, with odds attached

Aegis positions the launch against the way most scheduling software works. In its description, traditional APS tools, including the scheduling add-ons bundled with many MES and ERP systems, run on rules and historical averages. That holds while the plant stays close to normal. Rules cannot capture how constraints interact, though, so when something changes the plan breaks, regenerating it can take hours, and planners patch it by hand. Hand-built simulations kept apart from live data drift out of date, and newer lightweight tools tend to answer one question, such as whether a new line will hit its target output, rather than schedule production every day.

Simio steps through the operation one event at a time against equipment, labor, materials, changeovers and business rules. Because it works with ranges rather than single values, Simio APS reports how likely each order is to ship on time. Pegden tied that to the moment a customer gets a date: the point is to see the risk of a late order before the promise is made.

That probability may be the most useful output for people outside the planning office. A team quoting delivery from a single capacity figure gets a yes or a no. A likelihood per order could let the plant separate safe commitments from the ones that need an extra shift or a different line to hold.

Aegis goes further, saying Simio schedules can go to production with zero intervention. That is a vendor claim, and a testable one. Running the Simio schedule alongside the current process for a few weeks and counting the hand edits would show how close a given plant gets.

## Four tiers, split by data source

All four products sit on one simulation foundation, so a manufacturer can start at one level and move up without switching platforms. Here is what each includes, per the announcement:

- Simio Essentials: ready-to-use scheduling for FactoryLogix customers who have no modeling or simulation expertise. The twin is built automatically, and every run comes with six preconfigured dashboards: bottleneck analysis, schedule summary, schedule detail, order performance, throughput and work in progress, and a capacity heatmap.
- Simio Essentials Plus: builds on Essentials with a longer scheduling horizon, what-if analysis, comparison between plans, and support for additional or expected demand. It still requires no simulation expertise.
- Simio APS: complete planning and scheduling for complex operations. It handles labor, tooling and material handling constraints, schedules primary resources automatically, embeds AI and gives each order its own on-time probability.
- Simio Virtual Factory Designer: a digital twin used to design, optimize and validate layouts ahead of commissioning, so the plant ends up running the design that was validated.

Read closely, the line between tiers follows where the data comes from as much as how complex the plant is. The two Essentials tiers exist only for FactoryLogix customers. A manufacturer on another MES can buy Simio APS or Virtual Factory Designer standalone, but it starts at the full-depth end, and the announcement describes automatic twin generation only for FactoryLogix plants.

Beyond the listed features, the materials do not spell out what a line supervisor would notice differently between tiers in the first week, so a demo on the plant's own data is the way to find out. Virtual Factory Designer closes a different loop: the layout a team validates before commissioning becomes the model that later schedules the floor. Aegis's own shorthand is that the same simulation designs the system and schedules it.

## Neural networks inside, chat assistants outside

The AI operates in two places. First, neural networks embedded in the simulation learn from shop floor context and from data the simulation itself produces, and Simio APS calls on them while a plan is running, for example to choose the best line for a job given the factory's current state. Second, built-in Model Context Protocol (MCP) support lets a company's existing assistants, such as Claude Code, Codex and GitHub Copilot CLI, connect straight to the simulation. Planners can ask which orders will be late and why, or launch a new scenario in plain language, and the responses are based on the factory's real orders and constraints.

A November 5 post on Simio's blog explains the mechanics. MCP is an open standard, so the software is not tied to one AI provider, and Simio ships two MCP servers: one for its documentation and one for the model that is open. The case for embedded networks is that lead time often depends on the mix of products already on a line, which a rule like shortest queue cannot see. A simulation can also generate as much training data as needed, including scenarios that have not happened yet, and networks trained elsewhere can come in through the ONNX format.

Block Field

Those figures come from demonstrations on sample models, not customer projects, and Simio's own guidance is blunt about the routing result. Random assignment is an easy baseline to beat. The right comparison is the plant's best existing dispatch rule, and if the network does not clearly win, Simio advises keeping the rule because it is simpler to explain and maintain. The same post calls AI answers a strong first draft that can vary between runs and should be reviewed before anyone acts on them.

A January 2026 preprint on arXiv from a multi-university research team offers a useful yardstick. It sorts AI in digital twins into four stages: modeling the physical system, mirroring it with real-time synchronization, intervening through prediction and optimization, and autonomous management by language models and agents. By that map, the Simio launch covers mirroring through the MES feed and intervention through embedded networks, with language models acting as a question-and-answer layer for a planner who still makes the call. Asking a vendor which stage its AI actually operates in is a quick way to read any scheduling demo.

## Testing it against last month

For production planning managers on FactoryLogix, Essentials Plus maps most directly to daily decisions: adding a resource, running an extra shift or accepting a rush order, each tested before the commitment. High-mix plants where orders shift and machines go down often stand to gain the most. A plant whose schedule rarely changes may find the spreadsheet is still good enough.

Block Field

One item belongs on the IT list too. MCP connects coding assistants to live order and constraint data, and Simio notes that MCP support varies by AI product and plan. The CIO will want to confirm which assistant accounts can connect, and who gets that access, before planners start asking it which orders will be late.

## Sources

- [Aegis Now Offers AI, Digital Twin Simulation, Advanced Planning](https://www.assemblymag.com/articles/100608-aegis-now-offers-ai-digital-twin-simulation-advanced-planning) (Assembly Magazine)
- [Aegis Software Brings AI-Driven Digital Twin Simulation, and ...](https://finance.yahoo.com/technology/ai/articles/aegis-software-brings-ai-driven-133000667.html)
- [5 Ways to Use AI in Simulation Modeling](https://www.simio.com/blog/5-ways-to-use-ai-in-simulation-modeling)
- [Digital Twin AI: Opportunities and Challenges from Large Language ...](https://arxiv.org/html/2601.01321v1) (arXiv)

Tags: Aegis Software, Simio, FactoryLogix, Simio APS, Model Context Protocol, production scheduling

---
Source: MarketScale, https://www.marketscale.com/industries/industrial-iot/aegis-builds-simio-factory-simulations-from-live-mes-data-instead-of-by-hand. Published for AI indexing and citation; cite the canonical URL. Site guide for agents: https://www.marketscale.com/llms.txt
