Industries / Industrial equipment and services / Industrial services
Process safety consulting
Process safety consultants help plants that handle flammable, explosive or toxic materials work out how those materials could escape, and check that the safeguards against fires, explosions and toxic releases will work.
1963
At ICI, the British chemical company, production manager Ken Gee applied the company's "critical examination" method to the design of a new phenol plant, in what Trevor Kletz later called the first recognisable hazard and operability study (HAZOP).
35
Milestones achieved across 7 domains.
10
To watch for: announced, not yet confirmed outside.
14
Open deliverables. Nobody has shipped them.
In one minute
Where it started
In 1963 an ICI production manager spent four months questioning every part of a new phenol plant's design, and found hazards and operating problems nobody had foreseen. That review became the hazard and operability study (HAZOP). Risk calculations (Canvey Island, 1978), bow-tie diagrams (ICI, 1979) and rules for reviewing changes and auditing safety programs (API RP 750, 1990; UK offshore safety cases, 1993) followed.
Where it stands
Hazard studies are now a legal duty in much of the world. OSHA's process safety standard requires the hazard analysis of each covered US process to be redone at least every five years by a team that includes someone who knows the method, and Europe's Seveso III Directive covered 11,059 sites in 2022, by the European Commission's count. The work still rests on judgment: in one EU benchmark, seven teams' estimates of how often the same pipelines would fail differed up to 10,000-fold, and a 2007 review found that BP's audits checked that safety systems existed, not that they worked.
What is open
Language models now draft HAZOP worksheets, bow-ties and summaries of accident reports, but the published tests score them against one expert's work or a vendor's own case study. Most of the open jobs are tests and checkers that would show whether such tools catch what people miss: scenarios past studies overlooked, safeguards that share parts, gaps in safety specifications, dead legs left on drawings and changes nobody reviewed.
Domains
AchievedWatch forOpen
Domain
Progress
Last achieved
Watch for
Open
Hazard studies
Hazard studies are structured team reviews of a design or a running plant that ask how each part could go wrong and what would stop it.
4 · 2 · 2
Last achievedAn international HAZOP guide2001
Watch forWorksheets drawn from one drawing
2
openConsequence and risk modeling
Consequence and risk modeling works out how far a release, fire or explosion could reach and how often it might happen, so that plants, buildings and towns can be kept a safe distance apart.
6 · 1 · 2
Last achievedTwenty-one teams model ammonia2024
Watch forFast stand-ins for CFD
2
openLayers of protection
Layers of protection work decides how many independent safeguards a hazard needs, and shows that each one really is independent.
5 · 2 · 2
Last achievedLOPA studies checked by regulators2009
Watch forBow-ties drafted from tables
2
openSafety systems
Safety systems are the automatic trips, relief valves and emergency shutdown valves that act when people and ordinary controls fail. Consultants specify them, check their reliability figures and plan how they are tested.
5 · 1 · 2
Last achievedSecurity enters the safety lifecycle2016
Watch forAgents that act before trips
2
openManaging changes
Managing change means reviewing a change to equipment, chemicals, procedures or staffing for new hazards before the change is made.
5 · 1 · 2
Last achievedInspectors list missed changes2017
Watch forAI cross-links plant records
2
openIncident investigation
Incident investigation finds out why an accident happened, beyond the last error made, so that it does not happen again.
5 · 2 · 2
Last achievedReporting rule by court order2020
Watch forBoard's funding in doubt
2
openAudits
Auditing checks that a plant's safety program works in practice, not only that it exists on paper.
5 · 1 · 2
Last achievedOutside auditors after an accident2024
Watch forOutside audits may be cut
2
open
Advances to build with
3 research notes

Claude Fable 5.2
Three working devices
Anthropic says its AI model Claude Fable 5.2 designed three prototype devices that worked without electrical modifications, with no outside test reported.

Mistral Large 4
A trillion open parameters
Mistral published its Large 4 AI model, a trillion numbers by its count, free to download and sell, though firms that rent out models need a separate license.

Cursor Agent SDK
Findings that cite notes
Cursor released a kit for building AI agents, and its example, a code reviewer that cites the note behind each finding, is known only from Cursor's account.
Open in process safety consulting
3 of 14 open deliverables
Open · Hazard studies
Scenarios real studies missed
A public test set of 25 accidents in which the investigators found that a hazard study had missed or misjudged the scenario that happened. Each case holds only what a study team had before the accident: a redrawn P&ID, the operating conditions and the materials. The missed scenario is kept sealed as the answer. AI hazard-study tools, and one human team for comparison, are scored on whether they raise it.
Built with what exists
CSB investigation reports; pyDEXPI, an open-source Python library for DEXPI drawing files, to redraw each P&ID; current models as the tools under test, such as Claude Fable 5.2 (Note: "Three working devices") and the open-weight Mistral Large 4 (Note: "A trillion open parameters").
Open · Consequence and risk modeling
An assumption ledger for QRAs
A tool that reads a quantitative risk assessment (QRA) report and lists every number it rests on, with its page and the data source the report names: failure frequencies, hole sizes, pipe lengths, part-time factors, ignition probabilities and weather cases. It also lines up two reports side by side, so the assumptions behind a gap in their answers can be seen.
Built with what exists
QRA reports that are public in planning and permit files; an AI reviewer of long documents that cites the passage behind each item, such as the Claude-based reviewer in the Note "Seven more flagged" or an agent built with Cursor's Agent SDK (Note: "Findings that cite notes"); pdfplumber to pull tables from PDFs.
Open · Layers of protection
Flag safeguards that share parts
A checker that reads LOPA worksheets and flags every credited protection layer that is not independent of the initiating cause or of another credited layer: an alarm on the same controller as the control loop that failed, a trip that uses the same transmitter as the alarm, or one operator response counted twice.
Built with what exists
Kenexis Open-PHA Desktop, free on Windows, Mac and Linux, which stores studies as JSON; Python rules for shared sensors, logic solvers and final elements; a model such as Claude or Gemini to read the free-text safeguard descriptions, citing the worksheet row and rule behind each flag (Note: "Findings that cite notes").
Where it sits
Broader