Basis · free textREC-2026-0419 · loop PSV-114
Disposition
Existing safeguards adequate.
Approved by
J. Smith
Basis not in file
PSM · PHA disposition — the decision-evidence layer for industrial maintenance
A declined safeguard recommendation needs a record. Not a memory.
Your systems record what happened. Sovel keeps the reasoning — and who signed it.
When a PHA recommendation is declined or modified, the reasoning is what an auditor asks for three years later — and it's almost never where they can find it. Sovel turns each disposition into a governed Decision Card: the grounds, the evidence, and a qualified reviewer's sign-off, in one tamper-evident record.
Built with input from process-safety practitioners. You keep your data; Sovel keeps the audit chain.
REC-2026-0419PHA disposition
Recommendation
Reconcile PSV-114 set pressure against the revalidated relief load.
Disposition & rationale
Accepted with revised set point. Basis recomputed from the 2026 relief study and captured against this recommendation.
Reviewed and signed by a qualified reviewerAPI 510 inspector · cred #IN-510-4471 · 2026-04-19 14:02 CT
Sealeded25519:9f3c1d…a71b · evidence-anchored · ALCOA+ retained · illustrative example
01 · The gap
‘Existing safeguards adequate’ is not a record.
Your PHA / HAZOP team generates recommendations. Most get accepted and tracked to closure. But the ones that get declined or modified are where the audit risk lives — because the rationale for not acting is the hardest thing to reconstruct later.
In most facilities that rationale lives in a free-text “Basis” field, a meeting minute, or a retired engineer's memory. None of those survive:
Basis decay
The conditions that justified declining a recommendation in 2021 — a since-removed interlock, a since-changed operating envelope — are no longer true, and nobody re-checked.
Turnover
The process-safety engineer who wrote the disposition rationale has moved on. The next revalidation team inherits a decision with no defensible reasoning attached.
Revalidation archaeology
Every five-year PHA revalidation turns into a forensic dig: why did we say no to this last cycle, and is that still valid? That dig is billable consultant-hours your team pays for repeatedly.
This is not a knowledge-management nicety. It is recurring spend — revalidation archaeology hours, audit-prep scramble, declined-recommendation re-litigation — plus a tail risk every PSM leader recognizes: the declined recommendation that an investigation later identifies as the one that mattered, with no documented, qualified basis behind the decision.
02 · How it works
One recommendation. One Decision Card. One sealed record.
A PHA recommendation surfaces — from a HAZOP, a LOPA, a revalidation, a what-if. Sovel walks the disposition through a structured Decision Card.
Step 01Signal
An overridden set point, a shift-handover gap, a PHA recommendation left open.
in: PSV-114 reset · basis not in file
Step 02Question
Sovel asks the qualified person the one question the record cannot answer on its own.
ask: why was the prior basis wrong?
Step 03Record
The human explanation is captured and bound to the evidence that supports it.
bind: rationale → 2026 relief study
Step 04Proof
Reviewer sign-off is sealed with cryptographic provenance: who signed what, when, with what evidence.
seal: ed25519 · API 510 · 14:02 CT
The division of labor
FDD watches the sensor. Sovel watches the reviewer.
03 · The grounds
Declining a recommendation, done right.
When a team declines or modifies a PHA recommendation, good practice — CCPS guidance, OSHA PSM enforcement practice, Primatech and peer methodology — recognizes four defensible grounds. Most PHA tools leave this as a free-text box. Sovel makes it a structured, evidence-backed, signed selection.
GroundPlain meaningWhat Sovel attaches
MFEMaterial factual errorThe analysis the recommendation rests on contains a factual error.The corrected fact + its source.
NNNot necessaryThe recommendation isn't necessary to protect employee safety and health.The basis for why the existing safeguards already cover it.
AMAlternative measureA different measure provides an equal or greater level of protection.The alternative, and the evidence it's at least as protective.
INFInfeasibleThe recommendation is not feasible as written.The feasibility constraint + what was done instead.
Defensible best-practice taxonomy — not a CFR-mandated schema. The value is audit-evidence quality and consistency.
The difference
PHA tools record the disposition as free text. Sovel records it as a structured ground, with the evidence attached and a qualified reviewer's sign-off — the difference between “Approved By: J. Smith” in a text box and a defensible, tamper-evident record.
04 · What's in the room
What your PHA tool doesn't do.
What's already in the room — and what Sovel adds.
In the roomWhat it doesWhat Sovel adds
PHA / HAZOP toolsPHA-Pro, SpheraCloud ARA, EnablonRun the study, list recommendations, store dispositions as free text.Enforces a structured ground + attached evidence + qualified sign-off on every disposition, sealed in a tamper-evident record.
CMMS AIMaintainX, UpKeep, eMaint, FiixRetrieves what's in the work order.Finds what was never documented — the disposition reasoning that never hit the CMMS.
FDD / MBCxSkySpark, Aspen Mtell, AuguryWatches the sensor; flags drift and overrides.Complement. Those findings generate Sovel dispositions — Sovel watches the reviewer.
05 · The seal
The record that survives a photocopier
Closure you can hand to an inspector.
“Who signed what, when, with what evidence” — the line every audit comes down to, answered before the auditor asks.
The messy live signal resolves into a single static record: ink on paper, qualified reviewer sign-off tied to the evidence, tamper-evident by design. You keep your data; Sovel keeps the audit chain.
Who signedA qualified reviewerAPI 510 · cred #IN-510-4471
On what evidence2026 relief studyevidence-anchored · ALCOA+
Live reviewer console
Reviewer readoutREC-2026-0419 · loop PSV-114
Why unwrittenRelief valve PSV-114 was reset, but the basis for the prior set point is not in the file.Gap
AI-suggested draft“Set point likely revised per 2026 relief study.” — unverified, awaiting a qualified reviewer.Suggested
Governed sign-offReviewed and signed by a qualified reviewer; basis recomputed and bound to the relief study.Governed
Disposition & rationale
Accepted with revised set point. The missing basis is now captured against this recommendation, not left to the next revalidation to rediscover.
SealedAPI 510 inspector · cred #IN-510-4471 · 2026-04-19 14:02 CT
ed25519:9f3c…a71b · evidence-anchored · ALCOA+ retained · illustrative fixture
06 · Work with Sovel
Have declined PHA recommendations you can't fully defend?
We usually start with a Decision-Card Sprint — a short working pass over the declined or deferred recommendations you already have, turned into signed Decision Cards a named reviewer stands behind. It fits naturally before a PHA revalidation or an audit. What we'd need, and the scope, we settle together on a call.