{{ screenTitle }}
{{ screenSub }}
⌕
🔔
{{ notifCount }}
Notifications
{{ nf.title }}
{{ nf.detail }}
{{ nf.time }}
✦
VI
Vektra Instruments
{{ greeting }}
Agents handled 31 items overnight - {{ needsYouCount }} need you now.
{{ a.pill }} {{ a.title }} {{ a.meta }} Open →
{{ k.label }}
{{ k.value }}
{{ k.delta }}
Operations at SLA riskSLA monitoring →
{{ r.id }} {{ r.title }} {{ r.sla }}
Agent network - todayActivity →
{{ g.name }}
{{ g.autonomy }}
{{ g.done }}
actions
{{ g.waiting }}
waiting
Est. employee time saved this week: 142 h · human interventions: 9 roadmap metric · directional
{{ f.label }}
{{ caseCount }} cases · sorted by SLA urgency
OperationObjectiveCustomer · ProductSevStatusOwnerSLA
{{ c.id }} {{ c.issue }} {{ c.customer }} · {{ c.product }} {{ c.severity }} {{ c.status }} {{ c.owner }} {{ c.sla }}
← Live operations OP-2026-0418 Deuterium lamp fails to ignite - error E-217 {{ heroStatus }} S2 · core function impaired Warranty active · 24 mo
Created Jul 16, 09:12 · updated {{ heroUpdated }}
Operation summary
{{ s.k }}{{ s.v }}
SLA & risk{{ slaChip }}
Resolution plan due{{ slaRemaining }}
{{ s.k }}{{ s.v }}
Recommended intervention: {{ slaIntervention }}
{{ t.label }}
AIMost likely cause - lamp power-supply board (PSU-D2) failure87% confidence
Deuterium lamp fails to strike with error E-217. Replacement lamp shows the same failure, ruling out the lamp itself. Instrument log shows the D2 anode rail collapsing to 41 V during ignition (spec ≥ 85 V) - consistent with a degraded PSU-D2 board. Not remotely repairable.
Recommended:Open RMA - depot repairalternatives considered: remote fix (ruled out) · escalate to engineering (not needed)
Possible causes
{{ c.name }}{{ c.p }}
Diagnostic steps performed
{{ d.mark }}{{ d.text }}
Supporting evidence
{{ e.src }}{{ e.text }}
Knowledge sources · similar cases
📄 SM-VX4200 §6.3 - D2 ignition circuit diagnostics 📄 KB-1182 - E-217 fault tree (updated May 2026) ⟲ CS-2025-2291 - same fault · PSU-D2 replaced · 9 days ⟲ OP-2026-0102 - same fault · depot repair · 8 days
Information collected: serial ✓ · firmware v4.2.1 ✓ · instrument log ✓ · error photo ✓ · usage hours ✓  Missing: none
AI summary: Customer reported total D2 lamp failure during method validation. Guided diagnostics ruled out the lamp; log analysis points to the PSU-D2 board. Customer has a regulatory deadline Aug 3 - repair turnaround matters.
{{ m.channel }} {{ m.who }} AI {{ m.time }}
👍👎💬
{{ m.body }}
{{ transcriptLabel }}
recording ref REC-8837 · 07:42 duration · consent logged
{{ t.who }}{{ t.text }}
Every action is logged with actor, permission and timestamp. Open in audit trail →
{{ e.actor }} {{ e.title }} {{ e.tag }} {{ e.time }}
{{ e.detail }}
✓ chain verifiedHash-chained ledger · {{ caseAuditCount }} entries for this operation · tamper-evident
{{ a.actor }} {{ a.action }} {{ a.ts }}
Permission {{ a.permission }} · policy {{ a.policy }} · accessed {{ a.accessed }} · outcome {{ a.outcome }}
Approval requiredAPR-0142 · due 17:00
Create RMA & send customer return instructions
Agent recommendation: depot repair under warranty. Confirmed hardware fault (PSU-D2), not remotely repairable, entitlement verified.
Why approvalPolicy RMA-01: RMA creation requires human sign-off
RiskLow - standard covered defect
Business impact~€1,240 vs on-site visit · protects Aug 3 deadline
Evidencelog excerpt · KB-1182 · warranty record
ApproverM. Lindqvist (you)
Your role ({{ roleLabel }}) cannot approve RMA creation. Ask an Approver - M. Lindqvist or S. Weber.
{{ approvalResultTitle }}
{{ approvalResultBody }}
RMA-2026-0097Depot repair · stage 4 of 17Open →
Next actions
{{ n.group }}
·{{ i }}
RMACustomer · ReasonTypeStageCase
{{ r.id }} {{ r.customer }} · {{ r.reason }} {{ r.type }} {{ r.stageLabel }} {{ r.caseId }}
The AI agent coordinates, monitors, reminds, documents and escalates - receiving, repair, inspection and shipping are performed by people. Physical steps below always name a human owner.
{{ rmaHead }}↳ {{ rmaLink }}
{{ m.k }}{{ m.v }}
Next required action: {{ rmaNext }}
⚠ Exception: {{ rmaException }}
Process stages
{{ s.mark }} {{ s.name }}
IDRMADirCarrierRouteStatusETA
{{ s.id }} {{ s.rma }} {{ s.dir }} {{ s.carrier }} {{ s.from }} → {{ s.to }} {{ s.status }} {{ s.eta }}
Logistics & Customs Agent tracks progress, detects delays and holds, prepares documentation, and escalates - it never books or changes carriers itself.
{{ shipHead }}
{{ d.k }}{{ d.v }}
{{ a.id }}
{{ a.action }}
Requested by {{ a.agent }} · {{ a.caseId }} · {{ a.reason }}
✓ if approved: {{ a.afterA }}  ·  ✕ if rejected: {{ a.afterR }}
Due {{ a.due }}
Approver {{ a.approver }}
{{ a.status }}
Review →
Approvals expire per policy if not decided; expired requests block the invocation and return to the agent. All decisions are written to the audit ledger.
{{ g.name }}AI
{{ g.type }} · {{ g.dept }}
Now: {{ g.task }}
Supervisor {{ g.sup }} · {{ g.deleg }}
{{ g.cases }}
active
{{ g.done }}
done 30d
{{ g.waiting }}
waiting
{{ g.esc }}
escal.
SLA {{ g.slaPerf }} · AHT {{ g.aht }} · est. saved {{ g.saved }}
Recent agent actions
TimeAgentActionRefPermissionResult
{{ f.t }} {{ f.agent }} {{ f.action }} via {{ f.sys }} {{ f.caseId }} {{ f.perm }} {{ f.result }}
RefCommitmentResponsibleStageTime remaining
{{ r.id }} {{ r.what }} {{ r.owner }} {{ r.stage }}
{{ r.left }}
Breaches fire the escalation ladder automatically (Agent → Supervisor +10m → Manager +20m → VP +40m for S1). Escalations are logged and visible on the case timeline.
{{ f.label }}
✓ hash chain verified · 4,812 entries Export bundle (ISO audit) →
TimestampActorAction · data accessedPermissionPolicy · refOutcomeCost · durHashRate
{{ a.ts }} {{ a.actor }} {{ a.action }} · {{ a.accessed }} {{ a.permission }} {{ a.polRef }} {{ a.outcome }} {{ a.cost }} {{ a.hash }}
👍👎💬
{{ g.name }}
{{ g.type }}
{{ g.autonomy }}
{{ agName }}{{ agType }}Read-onlyOwner: M. Lindqvist
Authority per capability
{{ c.cap }}{{ c.authority }}
Authority is granted per capability and per action - no blanket autonomy. Your role has view access only; changes are made by the accountable owner and written to the activity ledger.
{{ k.k }}
{{ k.v }}
Allowed actions
✓{{ x }}
Restricted actions
⛔{{ x }}
Scope & limits
{{ c.k }}{{ c.v }}
{{ x.type }}
{{ x.desc }}
{{ x.ref }}
Owner: {{ x.owner }} {{ x.goLabel }}
Exceptions are raised by the platform runtime - permission denials, low-confidence recommendations, SLA risk, failed system actions and cost anomalies all land here with an accountable owner.
Prototype previewAgent creation is unavailable in this prototype - templates shown for orientation.
Diagnosis agent
Triage, guided diagnostics, evidence-backed recommendations. Starts at Recommend authority.
Coordination agent
Runs playbooks across departments - reminders, documentation, hand-offs, approvals.
Monitoring agent
Watches external systems, detects delays and anomalies, escalates per ladder.
Technical diagnosis to product returnActive · 2 live runsView live run →
Agents coordinate, monitor, remind, document and escalate - receiving, repair and shipping are performed by named humans at each physical stage.
{{ s.n }}{{ s.name }}
Remote resolution
Intake → identify → guided diagnostics → remote fix → verification → close & survey. Fully agent-run below S2; human gate on config changes.
Escalation handling
Detect breach risk → notify accountable owner → escalate per ladder → track to resolution. Eligibility rules and ladders are maintained by the operations team.
SystemGranted scopesAgents usingStatusLast event
{{ c.sys }} {{ c.scopes }} {{ c.agents }} {{ c.status }} {{ c.last }}
Agents reach systems only through governed tools with the scopes above - every invocation is metered and written to the activity ledger.
{{ k.k }}
{{ k.v }}
{{ k.s }}
ROI · 30 days {{ roiNet }} net value {{ roiMult }} return on Fronto cost {{ roiHours }} human hours saved Open ROI →
Operations closed per dayJul 6-19
{{ b.d }}
Agent-completedHuman-completed
SLA attainment by weektarget 95%
{{ b.v }}
{{ b.w }}
Dip in W3 traces to the customs-hold cluster - see Activity ledger.
Messages by channel30 days
{{ c.name }}
{{ c.n }}
Full breakdown per agent in Comms & calls.
Agent runtime cost per dayJul 6-19 · €
{{ b.d }}
Metered per action from the activity ledger · ≈ €0.16 per action · connected to outcomes in Value & ROI.
{{ t.ch }}view →
{{ t.today }} today
{{ t.sub }}
Sent by agentstoday
⌕
{{ e.from }}{{ e.time }}
{{ e.subj }}
{{ e.preview }}
Every outbound email passes tone, claim and disclosure checks before send.
{{ emailSubj }}
{{ emailFrom }} · AIto {{ emailTo }}{{ emailTime }}
⌕
{{ emailSearchNote }}
{{ p }}
📎 {{ a }}
{{ emailOp }}👍👎💬
IS
Dr. Ingrid Solheim
NordBio Labs · opt-in confirmed · OP-2026-0418
Agent identified as AI
⌕
{{ m.text }}
{{ m.time }}✓✓
👍👎💬
Customer-opt-in only · policy Comms-01 · every message is written to the activity ledger.
# logistics-emea · # repair-centerinternal coordination · 3 channels
⌕
{{ m.init }}
{{ m.who }}AI{{ m.time }}{{ m.ch }}
{{ m.text }}
{{ m.op }}
👍👎💬
Slack is internal only - agents never contact customers here. Mentions notify the responsible employee.
⌕
CallTimeContactHandled byDurSentiment
{{ c.id }} {{ c.time }} {{ c.contact }} {{ c.by }} {{ c.dur }} {{ c.sent }}
Every call is recorded with consent, transcribed and attributed - agent-handled calls carry the policies under which the agent spoke.
{{ callId }}{{ callDir }}{{ callBy }}
{{ callOutcome }}
{{ m.k }}{{ m.v }}
{{ callDurSel }}
Consent captured at call start · retention 90 days · every playback is logged to the activity ledger
Transcript
⌕
{{ l.who }}{{ l.text }}
Rate how the agent handled this call👍 Good👎 Needs review💬 Comment
Hands-on activity by agent - today
AgentEmailWhatsAppSlackVoiceTotalMedian reply
{{ r.name }} {{ r.email }} {{ r.wa }} {{ r.slack }} {{ r.voice }} {{ r.total }} {{ r.reply }}
Counts include only messages the agent authored or answered itself - notifications and system events are excluded.
Channel policy
Outbound messages pass tone, claim and disclosure checks before send - agents always identify as AI on customer channels. WhatsApp is customer-opt-in only; Slack is internal coordination. Per-channel authority is set in Policies & autonomy.
Net value delivered · 30 days
{{ roiNet }}
Cumulative since go-live (Feb): {{ roiCum }}
ROI multiple
{{ roiMult }}
net value ÷ Fronto cost
Payback
{{ roiPayback }}
cumulative value overtakes cost
{{ roiTierLabel }}
Headline = measured value only.
Modelled and estimated tiers are opt-in.
{{ t.tier }} {{ t.name }}
{{ t.desc }}
{{ t.amount }}
{{ k.k }}
{{ k.v }}
{{ k.s }}
Cumulative value vs costsince go-live · {{ roiTierLabel }}
{{ b.m }}
Value deliveredFronto cost{{ roiCrossNote }}
Value by department / workflow30 days
{{ d.name }} {{ d.tier }}
{{ d.v }}
Modelled workflow rows appear when Tier B is switched on - they never blend into the measured headline.
Assumptionsset v3 · locked Jul 1 · M. Lindqvistversioned & timestamped - board numbers stay reproducible
{{ a.k }}{{ a.v }}
Default conservative: attribution below 1, under-claim rather than over-claim. Every rate is customer-agreed at onboarding - a number the customer set themselves, they cannot later dispute. Value and cost both derive from the same case object, so the two sides are consistent by construction.
Runtime cost & operational value
{{ t.k }}
{{ t.v }}
Cost & time saved by agent (30d)value estimates · directional
{{ c.name }}
{{ c.cost }} {{ c.saved }}
Cost is metered per action from the ledger; value connects completed operations, saved human time and SLA performance to each agent.
✦ Assistant
✦
Console assistant
manage the platform by chat · actions are logged
✕
{{ asstGreeting }}
{{ m.text }}
✓ {{ m.tag }}
{{ c.label }}
Feedback to agent team
{{ fbcContext }}
Sent to the agent's supervisor and training queue · logged in the activity ledger with your name.
{{ t }}