I could not read this worksheet{ws ? ws.name + ' · ' + ws.version : 'no worksheet'}
Unsupported layout. No coverage rows could be extracted, so I staged nothing. A partial parse never becomes a partial grid — I will not guess at a client's limit selections.
Your options. Re-request the selection on Madison's standard template, or key the selection manually so the shadow comparison still has a record for this renewal. Either way the client's returned document stays on file with its provenance.
Nothing about the renewal's position changes — this only decides how the selection gets read.
;
}
if (kind === 'verify') {
const c = r.close; const bad = c.rows.filter((x) => ['exact', 'corrected'].indexOf(x.st) < 0);
const mapped = window.MD.mappings.filter((m) => m.code);
return
{bad.length} unresolved line{bad.length === 1 ? '' : 's'}The AmRisk lock is authoritative — premium comparison is exact
{!bad.length ?
All lines now reconcile. The coverage grid staged automatically and is waiting for approval.
: null}
{bad.map((x) =>
{x.coverage}{x.st === 'map' ? 'No risk-code mapping' : 'Conflicts with the lock'}worksheet {money(x.ws)} · locked {money(x.lk)}
{x.st === 'map' ?
:
setVal(e.target.value)} />
A correction that still conflicts with the lock cannot be approved.
}
)}
;
}
return
Record the decision
setReason(e.target.value)} placeholder={d.id === 'DEC-5' ? 'Note for the operations lead' : 'Reason, recorded with your name and the time'} />
Human decisions stay visible in the renewal history and the Salesforce audit record.
{done ? <>{done}Recorded by {app.user.name} · {dt(new Date(MD.TODAY).toISOString())}
: <>
{r ? : null}
{d.autonomy === 'blocked' ? 'The agent stopped here deliberately — it will not guess between the options.' : 'The agent prepared this. Nothing sends until you approve.'}
}
{running ? Live · morning run : Morning run · complete 06:14}
Your renewal book is ready.
I reviewed all {c.evaluated} renewals overnight, read the latest correspondence, matched incoming documents, suppressed unnecessary follow-ups, and prepared the items that need your judgment. Most of the book is progressing without intervention.
{c.decisions} decisionsneed you today
{c.exceptions} exceptionsneed human help
{c.normal} renewalsI left alone deliberately
Evaluated{c.evaluated}
Progressing{c.normal}
Handled automatically{c.auto}
Communications prepared12
Need your decision{c.decisions}
Exceptions{c.exceptions}
{running || peek ?
{AG.RUN_STEPS.map((s, i) =>
{s.label}
{!running || i <= live ? s.at : ''}
)}
: null}
{!running ? : null}
:
Run complete 06:14
Evaluated{c.evaluated}
Progressing{c.normal}
Handled{c.auto}
Prepared12
Need you{c.decisions}
}
Your renewal worklist · Tue 17 Nov
I evaluated {c.evaluated} renewals this morning; {c.normal} are progressing without help. Below is the small part that needs you — each item is a decision I have already prepared and defended.
{decs.length} decisions prepared
Needs your decision{today.length}Act today
{today.map((d) => )}
Needs your decision{week.length}This week
{week.map((d) => )}
Prepared and readyAssembled by the agent · quick review, then it goes
{AG.prepared().map((p) =>
app.nav(p.route)}>
{p.label}
{p.note}
{p.n}
)}
Settled automaticallyNo human action was required for any of this
{AG.LEDGER.map((l) =>
setLed(led === l.id ? null : l.id)}>
{l.meta}
{l.count}
{led === l.id ?
{l.items.map((i, k) =>
{i}
)}
: null}
)}
Needs human help{help.length}The agent could not safely determine these
{help.map((x) => )}
MonitoringWatched deliberately · no action would add anything today
{AG.monitoring().map(({ r, why, meta }) =>
app.nav('#/renewal/' + r.id)}>
{r.account}{why}
{meta}
Monitoring
Next evaluation 14:00
)}
{c.normal} renewals were examined and deliberately left alone. The agent found nothing that needed a message, an escalation, or a person. That is the intended outcome for most of the book.