Everything a support engineer needs to close the case

Arcus is a desktop investigation desk for technical and customer support engineers. It drafts the first response, remembers past fixes and hands engineering a clean escalation. You approve anything customers will see.

Root cause with the evidence

Paste a ticket, a log line or a symptom. Arcus reads the ticket and its files, your code and your team’s past cases, and answers with the root cause, the file and line, what it ruled out and a workaround for today.

  • Confidence is shown with every root cause, and citations Arcus cannot verify are flagged.
  • Cross-module code reasoning: when Arcus opens a file it also pulls related definitions so answers cite the right symbols across packages. Cost-efficient by design: reused prompts get a discount, repeated lookups are cached and the code index persists between runs.
  • HubSpot or Salesforce CRM: plan, contract value, renewal and owner sit next to the case when connected.
  • Your org chooses the AI provider, or brings its own key.
How investigations work →
Recorded in the app with the demo org (sample data)

First response on every new ticket

When a Zendesk or ServiceNow ticket is created, Arcus reads it with its attachments and posts an internal note within seconds. It also drafts a public reply for you to review.

  • Logs, archives, HAR files, PDFs and screenshots are read. Every finding cites its source, like halyard.log:8812.
  • Confidence is deterministic and shown on the note, so you can see why a cause was suggested.
  • Public replies stay drafts unless your org turns on auto-posting and the evidence clears your threshold.
Set up first response →
First response panel with likely cause, attachment findings and confidence
Internal note with likely cause and attachment evidence (sample data)

Internal vs public guard

Every customer-facing draft is checked against the ticket’s internal comments before it can be posted. If a reply repeats internal text, posting is blocked and the reason is shown.

  • Internal notes stay internal. Public replies need an explicit approval.
  • Fixes that need your team’s own action (restarting a pod, a patch from engineering) are never offered to customers.
  • Urgent and critical tickets never get an automatic public reply.
Suggested public reply blocked by the public reply guard
A draft blocked for copying internal text (sample data)

Déjà vu resolution memory

Arcus matches the symptom against your team’s resolved cases and shows the root cause, the fix that worked and who fixed it. Matches are tagged public or internal, so you know what is safe to share.

  • Matches on error signatures, not only words.
  • Results are scoped to your org.
  • Confirmed matches feed the déjà vu hit rate in Insights.
  • Case Memory brings a confirmed past case into the investigation, cited as “seen before”.
Investigate a ticket →
Recorded in the app with the demo org (sample data)

Slack and Jira linking, escalation with approval

Link the Slack thread and the Jira issue to the case. When it needs engineering, Arcus drafts the bug with the repro, evidence and impact. Nothing is filed or posted until you approve it.

  • Linked tickets, Slack threads and Jira issues show up in one panel.
  • Possible duplicates are flagged before you file.
  • Approve, edit or skip each escalation.
Connect Jira →
Recorded in the app with the demo org (sample data)

Call transcripts with speaker identification Beta

Live calls are transcribed with speakers labeled, so notes, follow-ups and déjà vu searches come straight from what the customer said.

  • Speakers are matched to the agent and the customer.
  • The summary, next steps and linked ticket are saved to History.
  • Calendar meetings carry their ticket and prep notes.
Speaker identification →
Post-call summary with transcript
Post-call summary and transcript (sample data)

Cases sidebar: every open ticket

Your open queue from Jira, Zendesk and ServiceNow stays in the left sidebar with a live count. Assign a ticket to yourself in the helpdesk and it shows up after refresh — no restart.

  • Same open-ticket scope as Tickets → Active, not a short capped list.
  • Refresh is always visible; the list also updates about every minute and when you return to Arcus.
  • Closed tickets drop off the sidebar and stay findable through search.
Tickets and cases queue with open count
Open cases across connected helpdesks (sample data)

SLA tracking

SLA targets from Zendesk and ServiceNow are read into one view. Tickets are sorted by time to breach and flagged when they get close.

  • States: on track, at risk (60 minutes to breach by default), breached or paused.
  • At-risk tickets appear on Home next to your queue.
Recorded in the app with the demo org (sample data)

Insights and CSAT

Team and personal support metrics with honest math: CSAT among answered surveys, response rate, first response and resolution time (median and p90), FCR, reopens, escalations, SLA breaches and measured Arcus impact. Zendesk, ServiceNow or Jira can be the metrics source.

  • Every number shows its sample size and a confidence label.
  • CSAT drivers group the reasons behind bad ratings.
  • Personal view shows your numbers next to the team median, without rankings.
  • Jira breakdowns by customer, component and version when Jira is the source.
Metric definitions →
Team insights dashboard with CSAT, response times and trends
Team insights (synthetic data)

Zendesk and ServiceNow

Zendesk is fully supported. ServiceNow incidents work for first response, SLA and work notes, and are in beta.

  • Webhooks start a first response the moment a ticket is created.
  • Credentials are scoped to the permissions you grant.
  • Knowledge guides can publish to Zendesk Guide, a GitHub docs PR, ServiceNow, Salesforce Knowledge, Notion or Confluence. Each team picks a default destination.
Connect ServiceNow →
Tickets view with Zendesk queue
Tickets (sample data)