Nestack Agent Care
Industries / Public Relations / Inquiry triage agent

Public Relations AI agent · Inquiry triage

Media Inquiry Triage AI Agent

Log the inbound press inquiry the hour it lands, resolve who the sender actually is, route it to a named duty officer with the clock it started, and hold the answer for them.

4–6 weeksTypical delivery
Your stackDeployment
Intake firstNamed officer
Agent CareAfter launch

What this agent does

Routes the inquiry, never answers it

In
01

Take in the message from the press inbox, the switchboard note or the media form, and stamp the hour.

02

Resolve who the sender is, not what the byline claims: 243.100(b)(1) lists recipients, not trades.

Reason
03

Classify by subject matter, because a question about an unapproved use has to leave the press office.

04

Hold the stated deadline apart from the statutory clocks; EU and UK MAR now diverge on Article 17.

05

Hand the inquiry to a named duty officer, with the sender, the hour and the subject travelling with it.

Decide
06

Escalate an inquiry that evidences a leak, because MAR Art. 17(7) can end a lawful delay.

07

Acknowledge receipt to the sender — who now holds it, and by when — and stop at the receipt.

Out
08

Retain the message, the routing, the escalation and the words that went back, filed under the sender.

09

Write back into case records only inside the approval boundaries agreed during implementation.

Product statement

Logging, sender resolution, routing and escalation belong to the agent. The answer belongs to a named officer, who approves the words before they go back.

Example workflow

One inquiry, inbox to answer

AgentHuman
1Inquiry receivedPress inbox, switchboard note, media form or a message sent straight to a spokesperson
2Sender and subject resolvedThe address, the outlet, the prior contacts, the subject matter and the hour it arrived
3Routing proposedThe desk, the duty officer, the clocks in play and the routing confidence
4Controls appliedRecipient-category checks, subject-matter checks, delay-register checks and routing confidence
No human action required

Stages 1 to 4 run unaided, and nothing goes back to the sender at any of them — the agent is routing, and the officer lane opens at the confidence gate.

5DecisionSplits at the confidence gate
Confident routing

Goes to the named duty officer.

Anything low-margin

Abstains to the press-office lead first.

Officer review

The inquiry is held with its sender record, the clocks it touches and the routing confidence.

Answer · Reassign · Escalate
Answered — by the named officer
6Case and contact records updatedOnly where write access and the records policy permit it
7Outcome evaluatedRouting correctness, reassignments, queue dwell and what the officer had to add
Reassignments

Each reassignment made in review is counted in the evaluation.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required8 items
Answering a reporter on the record.
Deciding whether a delay is still lawful.
Judging who the sender is under the rule.
Confirming, denying or declining to comment.
Automation boundaryAgent acts unaided
Log the inquiry and stamp the hour it arrived.
Resolve the sender against the covered-recipient list before routing.
Hand it to the named duty officer and record the hand-off hour.
Acknowledge receipt, and name who now holds it.
Nothing goes back to a reporter except from a named officer, inside the agreed boundaries.
Deciding an inquiry is a records request.
Telling a reporter what the company will say.
Reviving an approved line for a new inquiry.
Changes to routing, retention or the rota.

Example output

One inquiry, annotated

This serves a press office that may have to reconstruct who was asked what and when, years later; in In re E.ON AG (Rel. 34-43372, 28 September 2000) the inquiry volume and the approval chain both sat in the SEC order.

Inquiry record · single messageIllustrative example
Sender
Asked
Arrived
Logged from
Confidence
Held for
Freelance contributor, trade weekly
Whether a supplier contract was cancelled
Thursday, 16:41
Press inbox header
Unanswered, routed
The duty press officer, by name
As receivedTaken from the message and the header beneath it, and nothing on this side is the wording of the agent.
What the intake holds Message header Switchboard note Prior contact log
Why it routed hereSubject matter decides the desk, not the trade the sender follows.
ActionRouteReassignEscalate to the officer
What the score decidesBelow the configured threshold the inquiry abstains and reaches the press-office lead first.

Value

Where AI adds value

The same four claims, placed at the point in the workflow where each one applies.

Where the value landsValue 01 – 04
Each inbound inquiryFrom the inbox that received it
03Routing

Resolve the sender

The agent does not settle who a sender is under the securities rules; it records the address, the outlet and the prior contacts, and hands that to the officer who must settle it.

01Approved path

A stranger started a clock

No statute started it. Since 5 June 2026 MAR Art. 17(4)(b) measures a lawful delay against the latest communication — and a holding line is a communication.

02Human review

What was searched for, and absent

No United States rule found turns an inbound press inquiry into a response deadline, and no sector rule found requires a press inquiry to be logged as such. Reg FD reaches securities professionals and holders, not reporters, and Gallagher v. Abbott (7th Cir. 2001) holds a firm may keep silent.

04Build an evidence trail

The inquiry, the hour it arrived and the spokesperson it reached stay together.

Integrations

Typical integrations

Five system groups connect to the same agent. Which of them are in scope is decided in discovery.

Press inbox and switchboardShared press inbox · press line
Web media form · voicemail log
Reporter and outlet recordsCRM · prior contact history
Outlet, byline and past asks
Disclosure and legal recordsDelay registers · insider lists
Prior outbound lines and their dates

Agent

Media inquiry triage

Reads the inquiry
Routes and escalates
Holds for the officer

Rota and case handlingDuty rota · press-office tickets
Escalation paths and hand-off logs
Observability & evaluationOpenTelemetry · Langfuse
Supported monitoring/evaluation sources

Integration availability depends on the client's existing systems and API access.

Agent controls

Six sorts between the inbox and the officer

Six sorts down one chute, the last the tightest. What lands is drawn in the map below.

L6 · Outermost — last line of defenceInward → L1 · closest to the model
L6Rollback / safe modeNarrow the agent to logging and acknowledgement when evaluation or production signals degrade.Roll back
L5Version monitoringTrack model, prompt and routing rules, and the rule text under them: 243.100(b)(2) was renumbered in 2010.Track
L4TraceabilityRecord the message, the hour, the routing, the escalation and each read of the file.Record
L3Officer releaseHold the answer for a named officer; the hold governs release, not whether the answer is right.Gate
L2Routing guardrailsTest each routing against the configured rules, and return anything whose subject falls outside them.Restrict
L1Confidence thresholdsSend a low-margin assignment to the press-office lead before it reaches a desk.Require review
Model coreInquiry logged — the sender, the hour, the subject, the stated deadline and the routing confidence
L1 – L2Test whether a routing may stand
L3Leaves the answer to a named officer
L4 – L5Keep the inquiry and the routing behind it
L6Routes to the duty officer when signals degrade

How Nestack evaluates it

Evaluate the whole intake — not only the desk the inquiry reached.

Coverage runs the whole depth of the workflow, and every layer is cut by slice.

Surface — the routing the press office works
Depth of coverage ▼
E1Final-output evaluationDid the inquiry reach a desk that could actually answer it?
E2Step-level evaluationDid the agent read the right sender record, rota and delay register?
E3Tool evaluationDid it write the correct case and the correct hand-off?
E4Confidence calibrationDo low-confidence routings actually attract more reassignment?
E5Slice evaluationHow does performance change across specific inquiry classes?
E6Business outcomeHow many inquiries were reassigned before the officer answered?
Floor — the record the officer answers from

Failure modes

Where each failure originates in the agent

Seven failure modes, each set at the stage where it first becomes visible.

Agent lifecycleDirection of processing →
01 · Retrieval1 mode
SZ-03

Sender typed from the byline

A covered recipient is filed as press.

Stage gathersThe message, the sender, the hour and the desk
02 · Reasoning2 modes
SZ-04

Out-of-scope inquiry

The one message unlike any other lands in the largest bucket.

SZ-06

One due field, four clocks

Differently anchored deadlines share a single field.

Stage proposesSender, subject, stated clock and confidence
03 · Tool / write2 modes
SZ-02

Unread queue dwell

An inquiry ages unread while a delay is running.

SZ-05

Approved line reissued

The same holding line goes out once more.

Stage writesOnly where write access and approval policy allow it
04 · Output1 mode
SZ-01

Hand-off without the log

The delay call is taken without last week line.

Stage returnsThe desk that holds it, and the hour it arrived
05 · Change / Version1 mode
SZ-07

Free-text notes on a reporter

A note is personal data, and Article 15 reaches it.

Stage tracksModel, prompt, routing rules and note fields
Sev-1 · answered without a named officer Sev-2 · a live delay dies in the queue Sev-3 · signals degrade, inquiry held back

Affected slices

Unfamiliar senders absorb the reassignments

A press-office reassignment figure can read clean while unfamiliar outlets and freelance senders carry most of the rework behind it. Nestack reports the rate by inquiry class, not only in total.

Slice performance — reported separately, not only in aggregateIllustrative example
SliceFailure rateLift Lift vs. thresholdStatus
Unfamiliar outlets and freelancers12.4%3.7× Review
Out-of-hours arrivals8.8%2.6× Review
Multi-jurisdiction subjects5.5%1.6× Watch
Named-desk regulars2.5%0.7× Normal
Bar: reassignment-rate lift vs. named-desk baseline · scale 0–4.0× · tick marks the 2.0× review threshold 2 of 4 slices over threshold

Evidence-linked improvement

What a reused line costs

A loop closes when the question that reached nobody before deadline is a standing case. That suite is what the next queue worked is measured against.

Improvement cycle · five stagesSwitchback — the path turns at Improve and returns at Learn
01Detect

Reassignment rate rises on unfamiliar senders.

02Diagnose

The inquiry that reached nobody until the reporter published is worked backwards until one cause is left standing.

03Improve

Number the acknowledgement; the inquiries that drove it are filed beneath it.

04Verify

A single unrouted inquiry case stops the whole queue closing.

05Learn

One case joins the suite, and one line joins the routing rules.

Learn → DetectThe return edge. The next queue is worked against a suite one case longer.

Typical build scope

Twelve workstreams across six weeks

The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, routing workflow, evaluation, integration, then production validation and handover.

Workstream Week 1Week 2Week 3Week 4Week 5Week 6
01Inquiry-queue discovery and automation-boundary work.
02Inbox, switchboard and CRM assessment.
03Sender-resolution and subject-classification rules.
04Inquiry intake and logging.
05Sender resolution and routing logic.
06Confidence scoring and abstain routing.
07Duty-officer approval workflow.
08Inbox and rota integration.
09Routing and deadline cases.
10Guardrails and acknowledgement controls.
11Inquiry-trail instrumentation.
12Deployment, documentation and Agent Care handover.
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel Final scope and sequence confirmed in discovery

Engagement tiers

What each tier includes

Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.

Capability✓ in scope · — not at this tier PilotOne inbox, one desk ProductionProduction press-office workflow AdvancedMultiple desks / jurisdictions
Introduced at Pilot
Routing to your rota and rules
Named officer approval
Inbound-queue baseline
Introduced at Production
Reporting by inquiry class
Officer review workflow in your systems
Approved case write-back
Inbox-and-rota integration
Introduced at Advanced
Cross-border deadline rules
Multi-desk escalation chains
High inquiry volume
Multi-desk escalation controls
Build price From $5,000 From $8,000 Custom quote
Final build priceConfirmed after discovery based on integrations, workflow complexity, inquiry volume, approval controls and deployment requirements.
Separate from buildBuild pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.

What we need from you

What you bring, and what we build with it

Each input maps to a piece of build scope and a week in the delivery timeline.

You bringWe build with it
01Your press inbox and the desks behind it Inquiry ingestion and hour stampingWeek 1
02Representative inbound inquiries, routine and awkward Sender resolution, routing logic and the intake baselineWeek 2
03Your duty rota and the officers it names Routing-rule mapping, rota binding and the automation boundaryWeek 1
04Access to relevant inboxes, APIs or exports Inbox, switchboard and CRM assessment, then integration setupWeek 2
05Questions you would not want unanswered Routing cases and the failure-mode roundWeek 4
06What no acknowledgement may concede Confidence scoring, abstain routing, guardrails and release controlsWeek 3
07A named officer who answers the reporter Duty-officer approval workflow, then pilot and production validationWeeks 5–6
Nothing else is required Deployment, documentation and Agent Care handover are ours.

Delivery timeline

Four phases across six weeks

Two of these phases do share week five, and no band was widened to make the column look tidier.

Phase W1W2W3W4W5W6
Discovery W1
Build W2 – W3
Evaluate W4 – W5
Pilot & Launch W5 – W6
Week focus W1Queue discovery, routing-rule mapping and the automation boundary W2Inbox and rota integration and the intake baseline W3Routing workflow, confidence logic and escalation controls W4Evaluation suite, routing cases and failure-mode testing W5Rota integration, pilot inquiries and targeted corrections W6One news cycle run under the press-office lead, then Agent Care handover
Reading the bandA bar sits on the weeks its own work is named in, and the week five overlap is real.
At the end of W6When the inquiry record validates, Agent Care picks the agent up.
DurationSix-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Public Relations AI agent

Build an inquiry triage agent around the hour a stranger question lands.

Show us one week of your press inbox and the rota behind it. Veba had no comment available on day one and reached for it on day thirty-two; the days in between were a queue that worked. If nobody can say which desk held Thursday, the record is gone.

Nestack Agents · Media inquiry triageAGT-COM-03 · Agent Care available after launch