AlliCrew Get in touch
A worked example · fictional data

A service call comes in. Which unit is it — and what does the technician need to know before rolling?

Here's what you're looking at: on the left is a scattered service archive — a call, two installs, a stack of old tickets. In the middle is a hand-authored example of the brief a Service Context System produces from it. Every line in the brief shows its work — click one to see the files behind it on the right.

A customer calls a fictional shop, Riverbend Manifold Works, about a leaking coolant manifold. The catch: Riverbend built this customer two manifolds, and the caller doesn't know which one is leaking. This is a worked example of a Service Context System reading the service archive and returning a brief that figures out which unit — and what's open, what's overdue, and what parts to bring — without guessing past what the records support.

Everything below is fictional but realistic. The records are a representative service archive; the brief is a hand-authored example of what a Service Context System produces, built to the same sourcing discipline as our live Quote-to-Spec example; the evidence is the source text behind every claim. The real version runs this on a representative slice of your own service records.

Panel 1
The service archive
The incoming call, plus the records the system gathers to place it.
call SVC-CALL-2026-0420 · customer: Harborline · "the manifold you guys did"
The service call
The trigger. The caller can't identify the unit.
MD
call_intake.md
Coolant weeping on "the one you guys did." Caller said "the newer one," then corrected to "the one from a few years back." Dispatch note: pull install history first — don't assume it's the recent unit.
equipment not confirmed on the call
The Harborline record set — what the system gathers
One account. Two Riverbend manifolds. Records that disagree in two places.
MD
account_note.md
Four name variants — Harborline Marine Services / Marine Svcs / Marine / HMS — all one account, same billing and site address.
4 customer-name variants
MD
install_2019.md  ·  install_2022.md
Two units: a 2019 304L manifold on the port genset (serial RMW-CM-19-0114), a 2022 316L manifold on the starboard genset (serial RMW-CM-22-0233). Different outlet count, different tube size.
MD
equipment_2019.md  ·  equipment_2022.md
Equipment records for the two serials. The 2019 record carries an install date that doesn't match the 2019 install record — five weeks apart, unreconciled.
install-date conflict on the 2019 unit
MD
ticket_2020_0117 · 2021_0288 · 2023_0451 · 2024_0677
Four service tickets, mixed order, mixed completeness. 2020 routine. 2021 grommet repair. 2023 can't be tied to either unit. 2024 found an issue and was left open.
2021 is a scanned legacy record 2024 ticket: open, no resolution
MD
resolution_2020_0117 · 2021_0288 · 2023_0451
Three resolution records. The 2024 ticket has none — its recommended follow-up was never scheduled.
MD
parts_record.md
Parts history and grommet compatibility. The two units do not take the same grommet size — 1.5" for the 2019 unit, 1.25" for the 2022.
Standing documents
MD
service_interval_std.md  ·  warranty_policy.md  ·  lord_vendor_file.md
Service-interval standard, warranty policy with a reported-issue exception, and the LORD grommet vendor file with part numbers by tube size.
The distractors — gathered, then correctly left out
Different customers. Similar equipment. The system pulls them in to check — and must keep them out of Harborline's history.
MD
coastal_freight_ticket.md
Coastal Freight Carriers — a different account. Same weeping symptom, similar Riverbend manifold. Not Harborline equipment.
different customer · same symptom
MD
northwater_install.md
Northwater Pilots Association — a different account. A 2022 316L unit nearly identical in configuration to Harborline's. Same year, same config, wrong customer.
different customer · near-identical unit
Two Riverbend manifolds at this customer. The caller doesn't know which one is leaking. Two pairs of records disagree with each other. And two near-identical units belong to other customers entirely. Sorting that out — before assigning a technician — is the work.
Panel 2
The service-context brief
A hand-authored example of the brief. Click any underlined line to see the files behind it.
Service Context Brief — SVC-CALL-2026-0420
Hand-authored example · built from a 19-document service archive — 17 records cited below, 2 distractors gathered and correctly excluded · authored to the same sourcing discipline as the live Quote-to-Spec example
What the system caught

The leaking unit is probably the 2022 starboard 316L manifold — the open 2024 finding describes the same symptom on that unit — but the call did not confirm it, and the brief does not treat it as confirmed.not confirmed

Two Riverbend records disagree on the 2019 unit's install date — five weeks apart, unreconciled. The brief surfaces the conflict instead of picking the more convenient date.conflict

The 2024 service finding on the starboard unit is still open — no resolution record on file, and the recommended grommet replacement was never scheduled.open issue

The call, as received

Harborline reports intermittent coolant weeping at a branch outlet on a Riverbend-built manifold, started a few weeks ago and getting worse. The caller could not identify which hull or which genset, and had no serial or job number.sourced

Equipment identity

Riverbend has two manifolds at this customer: a 2019 304L unit on the Tidewatch port genset (serial RMW-CM-19-0114), and a 2022 316L unit on the starboard genset (serial RMW-CM-22-0233).sourced

The strongest available signal points to the 2022 starboard unit — see what the system caught, above. Identity is routed to the dispatch checklist for confirmation, not asserted here.

Service history
  • 2020 — routine first-year inspection on the 2019 port unit. No fault, no parts.sourced
  • 2021 — grommet replacement on the 2019 port unit. Identity inferred from the genset reference: the ticket is a scanned legacy record and its serial was not legible on the scan.inferred id
  • 2023 — a routine check on "the Tidewatch coolant manifold" that cannot be assigned to either unit; the contract technician recorded no serial and no genset side.unassignable
  • 2024 — an early-grommet-seep finding on the 2022 starboard unit, left open.open
Open issues

The 2024 finding — slight dampness at outlet 7 of the starboard unit, noted as likely early grommet seep — has no resolution on file. The 2026 call describing worsening weeping may be a continuation of it.likely linked

Service gaps

The 2019 port unit is overdue for routine inspection by more than four years — its last confirmed service was the 2021 grommet replacement, and the ambiguous 2023 ticket does not reset its interval clock.sourced

Warranty status

The 2019 port unit is clearly out of warranty — a 36-month window from early 2019 closed in 2022, and both disputed install dates land it outside regardless.sourced

The 2022 starboard unit's original window closed in November 2025 — but the 2024 issue was reported in writing while the unit was still under warranty, so under the reported-issue exception a repair traceable to that report may still be warranty-eligible. This is a "confirm," not a settled answer.conditional

Likely parts for this call

If the unit is confirmed as the 2022 starboard manifold, the repair needs 1.25" LORD grommets (LM-125-ISO) — not the 1.5" grommets used in the 2021 port-side repair. The two sizes are not interchangeable, and the 2024 technician recorded not having 316L-configuration grommet stock on the truck.sourced

Uncertainty flags
  • Equipment identity is not confirmed — strong signal toward the 2022 unit, but unconfirmed.
  • The 2019 unit's install date is in conflict between two Riverbend records.
  • The 2023 ticket cannot be assigned to either unit.
  • The 2021 ticket is a scanned legacy record; its serial was inferred, not read.
  • The 2022 unit's warranty status is conditional, pending confirmation of the reported-issue exception.
  • Grommet lead time for the 316L configuration should be confirmed before dispatch.
Dispatch-readiness checklist
  • Confirm the unit on site — serial and genset side — before the visit. The records point to the 2022 starboard unit but do not establish it.confirm first
  • Carry 1.25" LORD grommets for the likely 2022 unit, and confirm 316L-configuration stock before the truck rolls.sourced
Recommended dispatch context

The technician should arrive knowing this is most likely a continuation of a known 2024 open finding on the starboard unit — not a new failure — and that the warranty question turns on confirming the reported-issue exception, which the office should resolve before the visit.context

Sources

Every claim above resolves to documents in the Harborline service archive — click any underlined line to see them. The two distractor records (Coastal Freight, Northwater) were gathered and reviewed, and appear in none of the citations — that absence is deliberate.

Panel 3
The evidence behind it
Every claim traces to documents in the service archive. Click a claim to see them.
Click any highlighted claim in the brief to see the source files behind it — the actual document text, and why it supports the claim.
"The leaking unit is probably the 2022 starboard 316L manifold — but the call did not confirm it."
Source documents
svc_call / call_intake.mdthe call
Caller was not certain which hull or which genset. Said "the newer one" then corrected to "the one from a few years back." Equipment identity is not confirmed from the call alone. Caller did not have a serial number or a Riverbend job number.
harborline_records / ticket_2024_0677.mdopen finding, same symptom
This is an open issue on the 2022 starboard unit (RMW-CM-22-0233). The 2026 service call describing weeping on a Riverbend manifold may be a continuation of this unresolved 2024 finding.
Why "probably," not "confirmed": The system connects the 2026 symptom to the 2024 open finding on the same unit — that's the strongest available signal. But it's an inference from symptom-match, not a confirmation. The brief says "probably" and routes identity confirmation to the dispatch checklist.
"Two Riverbend records disagree on the 2019 unit's install date — five weeks apart, unreconciled."
Source documents
harborline_records / equipment_2019.mdequipment record · 2019-03-18
Install date on this record: 2019-03-18. The install date on this equipment record does not match the install date on install record HMS-INST-2019-04 (2019-04-22). The discrepancy has not been reconciled. Do not treat either date as confirmed.
harborline_records / install_2019.mdinstall record · 2019-04-22
Install date: 2019-04-22. Riverbend job reference RMW-2019-0331. Equipment: coolant manifold assembly, 304L stainless, serial RMW-CM-19-0114.
Why this is the centerpiece: Two Riverbend documents, same unit, dates five weeks apart. The system surfaces the conflict rather than picking the more convenient date — and notes that warranty math depends on resolving it. In this case, both dates land the unit out of warranty anyway (see the warranty claim), so the conflict does not change the outcome here — but the brief still refuses to paper over it.
"The 2024 service finding on the starboard unit is still open — no resolution record on file."
Source documents
harborline_records / ticket_2024_0677.mdservice ticket · open
Resolution: OPEN — no resolution record on file. Technician recommended a follow-up grommet replacement visit. No follow-up ticket was created and no resolution record was filed. This service event was left open.
harborline_records / equipment_2022.mdequipment record
One open service issue: HMS-SVC-2024-0677, an unresolved early-grommet-seep finding at outlet 7. No resolution record on file. Recommended grommet replacement was never scheduled.
Why this is flagged open: Two documents independently record the 2024 finding as unresolved. The brief does not treat the issue as closed just because time has passed — an open finding with no resolution stays open until a resolution record says otherwise.
"Harborline reports intermittent coolant weeping at a branch outlet; the caller could not identify the unit."
Source documents
svc_call / call_intake.mdthe call
Caller reports intermittent coolant weeping at one of the branch outlet fittings on a Riverbend-built coolant manifold. Says it "started a few weeks ago, getting worse." Caller was not certain which hull or which genset. Caller did not have a serial number or a Riverbend job number.
Why this is the starting point: The brief records the call as it was received — symptom, timing, and the gap. The equipment-identity gap is not a flaw in the brief; it's a fact about the call, carried forward honestly rather than guessed away.
"Riverbend has two manifolds at this customer: a 2019 304L port unit and a 2022 316L starboard unit."
Source documents
harborline_records / install_2019.mdthe 2019 unit
Equipment: coolant manifold assembly, 304L stainless, serial RMW-CM-19-0114. Installed on Harborline support vessel "Tidewatch" — port main genset. 8 branch outlets, 1.5" OD branch tube.
harborline_records / install_2022.mdthe 2022 unit
Equipment: coolant manifold assembly, 316L stainless, serial RMW-CM-22-0233. Installed on Harborline support vessel "Tidewatch" — starboard main genset. 10 branch outlets, 1.25" OD branch tube.
harborline_records / account_note.mdaccount consolidation
Known Riverbend equipment at this account: at least two coolant manifold assemblies. See install records HMS-INST-2019-04 and HMS-INST-2022-11.
Why this matters: The two units differ in material, outlet count, and tube size. Establishing that there are two — and how they differ — is what makes the rest of the brief possible. The obvious move, assuming it's the newer unit, is exactly the one the brief is careful with.
"2020 — routine first-year inspection on the 2019 port unit. No fault, no parts."
Source documents
harborline_records / ticket_2020_0117.mdservice ticket
Equipment: coolant manifold, serial RMW-CM-19-0114 — Tidewatch port genset. Routine first-year service check. No fault reported. Grommets in good condition. No weeping, no corrosion noted.
harborline_records / resolution_2020_0117.mdresolution record
Routine service completed. No parts consumed. No follow-up required. Manifold and grommets within normal condition for service age. Ticket closed same day.
Why this is cleanly assigned: Both the ticket and its resolution name the serial directly — RMW-CM-19-0114. This is what a confidently-assigned service event looks like, and it's the contrast that makes the 2023 ticket's ambiguity visible.
"2021 — grommet replacement on the 2019 port unit. Identity inferred from the genset reference; the ticket is a scanned legacy record."
Source documents
harborline_records / ticket_2021_0288.mdscanned legacy · medium OCR
Document origin: scanned legacy PDF, OCR confidence medium. Serial number field on this scanned ticket is not fully legible. The ticket says "port genset" — and the only Riverbend manifold on the port genset is the 2019 unit, serial RMW-CM-19-0114. Equipment identity inferred from the genset reference, not read directly from the scan.
harborline_records / resolution_2021_0288.mdresolution record
Equipment: Tidewatch port genset coolant manifold (2019 unit, RMW-CM-19-0114). Grommet replacement completed. Two LORD isolation grommets consumed (VND-LORD-01).
Why "inferred," not "read": The scan's serial field isn't legible. The brief is explicit that the identity comes from the genset reference plus the fact that only one unit sits on the port genset — a sound inference, but labelled as an inference, not presented as if the serial were read off the page.
"2023 — a routine check on 'the Tidewatch coolant manifold' that cannot be assigned to either unit."
Source documents
harborline_records / ticket_2023_0451.mdservice ticket · ambiguous
This ticket cannot be confidently assigned to either the 2019 or the 2022 manifold. It says "Tidewatch" and "coolant manifold" but not which genset, and the contract technician did not record a serial. Treat this service event as ambiguous.
harborline_records / resolution_2023_0451.mdresolution record
Resolution record carries the same gap as the ticket: it does not record which manifold was inspected. Cannot be used to establish service date for a specific serial.
Why the brief does not guess: Both the ticket and its resolution name "Tidewatch" and "coolant manifold" but never the genset side or a serial. The system does not silently attach the event to one unit. An unassignable service event does not reset either unit's service clock — which is why the 2019 unit still reads as overdue.
"The 2024 finding — dampness at outlet 7 of the starboard unit — has no resolution; the 2026 call may be a continuation of it."
Source documents
harborline_records / ticket_2024_0677.mdservice ticket · open
Confirmed slight dampness at the outlet the customer described — outlet position 7 of 10. Technician's note: "Looks like early grommet seep. Recommend grommet replacement at next opportunity." The 2026 service call may be a continuation of this unresolved 2024 finding.
Why "may be," not "is": The symptom matches and the unit matches, which is a strong link — but the brief does not collapse a likely connection into a certain one. It surfaces the probable continuity and lets the technician confirm it on site.
"The 2019 port unit is overdue for routine inspection by more than four years."
Source documents
standing_docs / service_interval_std.mdinterval standard
A manifold with no confirmed service in more than 30 months should be treated as overdue. 2019 unit: last confirmed service 2021-09-30. As of the 2026 call, that is more than 4 years — well past the 30-month overdue threshold. An ambiguous ticket that cannot be assigned to a serial does not reset the interval clock.
harborline_records / equipment_2019.mdequipment record
Last confirmed service: grommet replacement in 2021 (HMS-SVC-2021-0288). No confirmed service since 2021 that can be definitively tied to this serial.
Why the 2023 ticket doesn't count: The interval standard is explicit — confirmed service means a ticket and resolution tied to a specific serial. The 2023 ticket can't be tied to this serial, so it doesn't reset the clock. Last confirmed service stays at 2021, and the unit reads as 4+ years overdue.
"The 2019 port unit is clearly out of warranty — both disputed install dates land it outside."
Source documents
standing_docs / warranty_policy.mdwarranty policy
2019 unit: install date is in conflict between two Riverbend records (2019-03-18 vs 2019-04-22). Either way, a 36-month window from early 2019 closed in 2022 — this unit is well out of warranty regardless of which date is correct.
Why the conflict doesn't change this: The install-date conflict is real and the brief surfaces it — but here it happens not to be decisive. A 36-month window from either March or April 2019 closed in 2022. The brief is honest about both facts at once: the conflict exists, and in this instance it doesn't move the warranty answer.
"The 2022 starboard unit's window closed in 2025 — but the 2024 reported issue may keep a repair warranty-eligible."
Source documents
standing_docs / warranty_policy.mdreported-issue exception
An issue reported in writing while the unit is inside the warranty window remains eligible for warranty consideration even if the repair visit happens after the window closes — provided the original report is on file as a service ticket.
harborline_records / equipment_2022.mdequipment record
Warranty window: 2022-11-08 through 2025-11-08. As of the 2026 call, this unit is outside the original warranty window — though the 2024 open issue was reported while the unit was still under warranty.
Why this stays conditional: The build guide's hardest must-not-claim — the brief must not say the 2022 unit is "in warranty" or "out of warranty" flatly. Its status genuinely depends on whether a 2026 repair can be traced to the 2024 written report under the exception. The brief flags it as "confirm," and routes the question to the office.
"A repair on the 2022 unit needs 1.25" LORD grommets, not the 1.5" grommets used in the 2021 repair."
Source documents
standing_docs / lord_vendor_file.mdvendor file · part numbers
1.5" tube → LM-150-ISO, used on 2019-era 304L manifolds. 1.25" tube → LM-125-ISO, used on 2022-era 316L manifolds. The two grommet sizes are not interchangeable — they look similar but will not seal correctly if swapped.
harborline_records / parts_record.mdparts history
The 2021 repair consumed 1.5" grommets — but that was the port-side 2019 unit. A repair visit for either unit needs the correct grommet for that unit's tube size — confirmed against the serial, not assumed from the 2021 replacement.
harborline_records / ticket_2024_0677.mdtechnician note
Technician's note: "did not have 316L-compatible grommet stock on the truck for this configuration."
Why the part is contingent: The correct grommet depends on which unit it is — and identity isn't confirmed yet. The brief carries the right part forward for the likely unit, explicitly conditioned on confirming the serial. The 2024 technician's stock note is why "confirm before the truck rolls" is in the checklist.
"Grommet lead time for the 316L configuration should be confirmed before dispatch."
Source documents
standing_docs / lord_vendor_file.mdlead time
LM-150-ISO: typically in stock. LM-125-ISO: typically in stock; confirm before dispatch as 316L-config demand has been higher.
Why flagged: The vendor file itself singles out the 1.25" grommet as the one to check. The brief carries that instruction through rather than assuming stock — particularly since the likely unit is the 2022 316L one, which takes exactly that grommet.
"Confirm the unit on site — serial and genset side — before the visit."
Source documents
svc_call / call_intake.mddispatch note
Two Riverbend coolant manifolds may be in service at Harborline. Pull the install history before assigning a technician — do not assume this is the most recent unit.
Why this is the first checklist item: The dispatch note itself sets the rule — don't assume the recent unit. The brief's identity work points strongly to the 2022 unit but does not establish it, so confirming the serial on site is the gate before any parts or warranty decision is final.
"Carry 1.25" LORD grommets for the likely 2022 unit, and confirm 316L-configuration stock before the truck rolls."
Source documents
harborline_records / parts_record.mdfor the current call
If the equipment turns out to be the 2022 starboard unit — which the open 2024 issue suggests — the visit needs 1.25"-configuration LORD grommets, not the 1.5" grommets used in the 2021 port-side repair.
standing_docs / lord_vendor_file.mdservice note
A repair visit must confirm the unit's tube OD — by serial, against the equipment record — before pulling grommet stock.
Why "likely," carried into the checklist: The brief doesn't make the technician arrive empty-handed waiting for confirmation, nor does it pretend the unit is settled. It says: bring the part the records point to, and confirm both the unit and the stock before relying on either.
"This is most likely a continuation of a known 2024 open finding — not a new failure — and the warranty question turns on the reported-issue exception."
Source documents
harborline_records / ticket_2024_0677.mdthe open finding
The 2026 service call describing weeping on a Riverbend manifold may be a continuation of this unresolved 2024 finding. Recommended grommet replacement was never scheduled.
standing_docs / warranty_policy.mdreported-issue exception
An issue reported in writing while the unit is inside the warranty window remains eligible for warranty consideration even if the repair visit happens after the window closes.
Why this is the closing context: The brief hands the technician a frame, not just facts — most likely a known issue resurfacing, with a live warranty question the office should settle before the visit. Every piece of it is sourced; none of it is asserted past what the records support.
This brief was authored to a standard — the same one our live system enforces.

This brief is hand-authored to represent Service Context output. The Quote-to-Spec system it's modeled on ships with two layers of automated checks — every claim resolved to a source, every prohibited overclaim blocked. The Service Context system applies the same discipline: it must surface the open issue, the install-date conflict, and the equipment-identity gap — and it must not claim the leaking unit is identified, or the 2022 unit's warranty is settled, when the records don't support it. Same checks, same refusal to overclaim — on a service-context workload instead of a quote-to-spec one.

17
claims, all sourced
0
distractors cited
3
conflicts & ambiguities surfaced, not resolved away
On your data

This one was fictional. Yours wouldn't be.

The call, the records, and the brief above were built from invented data. The real version runs this on a representative slice of your own service archive — and the brief comes back built against your records, with the same discipline: which unit, what history, what's open, what's overdue, what to bring — every line sourced, every conflict surfaced, nothing claimed past what the records support.

Request a working sample

One slice of your service records. A brief back within a week. No charge for the first one.