Overview
ProDex is an AI-native platform for manufacturing operations. It brings discrete event simulation, production planning, and BOMs into a single environment — and puts an operations-literate AI agent, Dexter, at the center of it. You work with Dexter in Chat. Describe what you need in plain language, and Dexter builds simulation models, profiles and transforms your uploaded data, runs analyses and scenario comparisons, generates reports, and captures recurring procedures as reusable workflows — grounded in your actual operation and auditable at every step. Everything you build lives in a Factory: the workspace container that holds your uploaded data and datasets, simulation models, BOMs, configuration templates, plans, Gantt timelines, apps, run history, and the knowledge base Dexter maintains about your operation. Each Factory belongs to the person who created it; sharing one gives a teammate their own copy. The core idea: instead of managing your operations across spreadsheets, point tools, and manual reports, ProDex gives you a live model of your factory — data, structure, and operational knowledge — that you and Dexter build and reason about together.New to the vocabulary? Key Concepts defines the handful of terms — Factory, model, run, artifact, knowledge base — that every other page assumes.
Who Is It For?
ProDex is built for the people who run manufacturing operations day-to-day:- Operations managers who need to understand throughput, bottlenecks, capacity, and sensitivity to disruption
- Production planners who turn demand into feasible plans that respect capacity, inventory, and BOM structure
- Schedulers who lay out jobs on machines and crews over time and keep that timeline current as work moves
- Analysts and engineers who model what-if scenarios: cycle time studies, layout changes, staffing changes
- Executives who want decision-ready reports grounded in the model, not in slideware
How You Work with Dexter
Dexter is not a chatbot bolted onto a tool. It is an operations partner with direct access to your Factory, your uploaded data, and a persistent memory of your operation:- Chat is the primary interface. Ask Dexter to profile a spreadsheet, build a simulation of a line, run a planning optimization, or draft a report — in plain language.
- Response modes: Turbo, Fast (the default), Max, and Ultra trade speed for thoroughness — how many questions Dexter asks and whether it uses sub-agents.
- Session-context awareness: Dexter knows which page you’re on and which objects you have selected (a model, a run, a pipeline), and works against those without you having to restate the context.
- Tasks: for multi-step work, Dexter maintains a visible task list so you can see exactly what’s in flight, what’s done, and what’s next.
- Knowledge Base: durable, per-factory memory of how your operation actually runs, what your data means, and what has been built in the platform. It survives across conversations; you can browse it, and Dexter updates it when you ask.
- Rules & Preferences: standing instructions you set for how Dexter should work — conventions, defaults, and guardrails it follows in every conversation.
- Workflows: once Dexter has learned to do a recurring task the way your team likes it, you can save that procedure as a workflow and reuse it in later conversations.
- Goals: for a long-running task with a clear finish line, set a goal. You approve it once, and Dexter keeps working until it has verified the goal is done.
Where to Start
Most new users follow this path:- Bring your data in: upload spreadsheets, MES exports, ERP reports, PDFs of work instructions, even screenshots or CAD files. Dexter profiles what you provide and confirms how to read it with you before building anything on it.
- Build a model: describe your operation to Dexter in Chat. Dexter constructs the model in the Modeler and marks each value it sets as Data, Stated, or Assumed, so every assumption surfaces for your review.
- Run simulations: see throughput, bottlenecks, utilization, and WIP, with full event traces.
- Compare scenarios: run each variant of a model or schedule as its own run and have Dexter lay the KPIs out side by side.
- Decide and share: generate production plans, Gantt timelines, and reports (PDF, PowerPoint, Word, Excel).
Key Capabilities
Modeler
Discrete event simulation, built as a flow graph — sources, processes, stations, buffers, routers, and sinks — with entities (parts, batches, work orders) flowing through. Shared resources (operators, machines, tools) are modeled explicitly, along with combiners, separators, and transformers for assembly, splitting, and conversion. Schedules anchor the run to real-world timing, and scheduled events change resource capacity for shifts, breakdowns, and staffing changes. For statistical confidence, replications run each case several times with different random seeds so you can see how much your KPIs vary.Data
How raw operational data becomes platform inputs. Upload an MES export, an ERP report, or a screenshot, or connect a system through an integration; Dexter profiles it with you and produces what you asked for — an answer, cycle time distributions or processing times for a model, a dashboard, or a report. For data that arrives again and again, Dexter builds a pipeline that prepares it the same way every time and publishes versioned tables to Datasets. See the Data overview for the full picture.Runs
Run variants of a model and schedule, then compare results side by side with KPIs and charts. This is how you do what-if work: different schedules, resource configurations, or parameter settings. Replications with several random seeds show how much each result varies from run to run.Custom modules
Interactive pages Dexter builds for your factory: what-if calculators, decision boards, and views for exploring your data. For numbers the team revisits — cycle-time trends, order backlog, quality rates — start from the Data insights template: once Dexter connects it to your data, it filters a dataset once and recomputes a set of linked charts from it. For a simulation run, use the Simulation results template or ask Dexter for a custom results view. Connect templates to your own data before using their sample values to make decisions.Production Planning
Combine demand orders, current inventory, and your BOMs; the planning optimizer produces a plan of what to make, when, and in what quantities, balancing on-time demand against inventory targets within the resource capacity you’ve modeled. Custom constraints and objectives — minimum lot sizes, minimum run lengths, smoother production — are supported.Gantt
Editable timelines of jobs on machines, crews, or project phases, which you and Dexter build together. Enforced dependencies keep work in order, exclusive rows flag double-bookings, and the critical path shows what sets the finish date. You or Dexter decide the job order; the Gantt doesn’t sequence jobs for you.BOMs and Configuration
BOMs capture product recipes and their component structure. Configuration templates capture configurable products as option classes and their nested decisions; in the Configurator, you or Dexter walk an order through one, decision by decision, to produce a configuration with a concrete parts list for that order.CAD
Open STEP and STL files in an interactive 3D viewer to rotate, section, and measure them; DXF files open there as a flat 2D drawing. Ask Dexter to analyze DWG, DXF, or STEP files — what the part is, its labels and part marks, the assembly’s structure, and its size — or to create a new part, modify an uploaded STEP model, or convert a file to another format.Reports
Downloadable, shareable deliverables — PDF, PowerPoint, Word, Excel — generated from your Factory’s models, runs, and data. When people need to work in a page rather than read a file, Dexter can build a module instead.Beta features are set up per organization, so some surfaces may not appear in your instance yet. Their pages in these docs carry a Beta badge. If you don’t see one you expect, ask your ProDex representative.