01Import project sources
P&IDs, registers, datasheets, vendor documents and lessons learnt come in as they are. Nothing is re-typed.
For process & safety engineering teams
Every value.
In one connected graph.
Plexus extracts values from your registers, datasheets and P&IDs and traces each one to its source. Calculations, revision checks and safety work run on the same data. Nothing enters the project record until an engineer approves it.
The cost of consistency
Every revision brings
the same work back.
A source changes. Formats drift. One discipline moves ahead while another works from stale values. Engineers rebuild the comparison by hand. Plexus turns that repeated rebuild into an evidence-backed review.
In June 2026 Schneider Electric agreed to pay $3.1 billion, about 18 times its 2025 revenue, for Cognite: software that resolves which records in different systems describe the same equipment. Those mismatches are created during FEED and detailed engineering. Plexus keeps the register consistent while the project runs.
Inside Plexus
The project state,
at a glance.
Open conflicts by severity, registered objects, documents and import batches. Each conflict links to the object it affects and the review that settles it. Behind every number sits the record and the evidence that produced it.
Sample with pre-entered data. The production interface evolves and may differ in detail.
Different tasks,
the same project graph.
Each module does one engineering job. Use one, combine several or run them all on the same connected project graph.
P&IDs reviewed like a senior engineer would
Equipment choices, protections and connections reviewed at the level of design decisions.
- Checked against typicals and your project knowledge base.
- Every finding pinned to its drawing location.
- Tags like P-101 and lines like 40-P-1204 matched to their registers.
Run the savings math on your own numbers.
Defaults describe a typical FEED scope. Every figure recomputes from the inputs you set.
490 hsaved per review package
€41,650estimated cost savings
- Baseline effort
- 1,400 manhours
- Reduction rate
- 35%
- Hours saved
- 490 manhours
- At €85/hr
- €41,650
88 hsaved per HAZID study
€8,379estimated cost savings
- Baseline effort
- 252 manhours
- Reduction rate
- 35%
- Hours saved
- 88 manhours
- At €95/hr
- €8,379
110 hsaved per HAZOP study
€10,474estimated cost savings
- Baseline effort
- 315 manhours
- Reduction rate
- 35%
- Hours saved
- 110 manhours
- At €95/hr
- €10,474
385 hsaved per project
€34,650estimated cost savings
- Baseline effort
- 700 manhours
- Reduction rate
- 55%
- Hours saved
- 385 manhours
- At €90/hr
- €34,650
Assumption-based estimates; the P&ID review reduction rate reflects early pilot experience. Actual values are determined per project.
How Plexus works
From scattered sources to
verified engineering data.
Six steps take a project from scattered sources to approved, traceable data.
02Connect to the graph
Tags and units are normalized and matched to the objects already in the graph.
03Surface conflicts
Disagreements between sources are flagged with severity and the evidence behind them.
04Run engineering calculations
Sizing, comparisons and datasheets draw their inputs from the graph.
05Engineer review gate
Candidate values stay candidates until an engineer approves them.
06Versioned and traceable
Approved values keep their evidence and full history.
Trust and engineering authority
Automation that stops at
the engineering decision.
Plexus can extract, compare, flag, draft and recommend. It does not make engineering decisions on its own.
Engineer approval
Extracted values stay marked as candidates until an engineer with review authority approves them.
Source evidence and history
Every proposed value links back to the exact place in its source document. Changes and review decisions are versioned.
Data isolation
Each client's project data is isolated at the database layer with Row Level Security.
EU data residency
European data centers by default. Before a deployment processes project data, we document which providers and models it uses.
Questions engineering teams ask
Clear boundaries build
useful engineering software.
Honest constraints up front. Easier than discovering them three months in.
Can Plexus reliably read every drawing?
No. Extraction produces candidates, each carrying the drawing evidence behind it; a value without evidence is not accepted. Engineers review every candidate against its source. Where the source gives no defensible value, the field stays empty and flagged.
Is project information used to train models?
No. Plexus does not train models on your project data. The data-processing terms of the providers enabled for a deployment are documented before project information is processed.
Which project sources are supported?
Structured registers and tag lists, P&IDs (PDF or DWG), datasheets, specifications, calculations and vendor documents. Values extracted from drawings wait for engineer review regardless of format.
Does Plexus replace our existing engineering tools?
No. Your team keeps producing registers, datasheets and drawings in the tools they already use. Plexus works on those outputs and keeps them consistent, reviewed and traceable.
Does Plexus depend on one AI provider?
No. Provider selection is abstracted, so each deployment can choose the AI models it uses without tying the project to one vendor.
Who builds Plexus
Plexus is built by two process engineers, close to 30 years of project work between us. Every workflow in the product is something we have done by hand, revision after revision. We build the software we wished we had on those projects.
Getting started
A pilot on your project,
then you decide.
Plexus starts as a bounded pilot on your own project data, reviewed by your own engineers. Four steps from first call to a rollout decision.
- Step 1
Intro call
A short call on your situation and which of your workflows Plexus can shorten. We agree the pilot scope: one project, a bounded set of documents.
- Step 2
Pilot run
We load the agreed documents into an isolated workspace and generate the deliverables.
- Step 3
Review with your engineers
Your engineers review what came out: flagged conflicts, the evidence behind each value, calculation results. They judge it against the original documents.
- Step 4
Rollout decision
You decide whether to expand to more documents and projects or stop there. The pilot findings are yours either way.
It starts with a short call.