AI-First Intake
Portal, email and widget requests are classified and routed by AI, with a deterministic fallback when the model is unavailable or busy. Emails become cases with their attachments, replies by email join the right case, and junk-mail rules keep auto-replies and newsletters out. Catalog requests keep the routing their form and entitlement rules decided.
A Simple Requester Experience
People describe the problem, attach screenshots and pick the affected device. Matching help articles appear as they type; a progress tracker, service announcements and a one-click satisfaction rating — with up to five questions of your own — keep them informed.
Indicative Dashboards
Agents land on what needs them: assigned work, unassigned cases, breaches, first replies due, approvals and major incidents — every tile and bar drills into the matching list.
Reports and Live Gauges
A library of standard reports — backlog, SLA, first response, workload, time and cost, satisfaction, sentiment, changes, problems, contracts — each with a chart, a CSV download and a click-through to the cases, and emailed to the team on a schedule. Need something else? Build your own report: choose what to count, how to group it and which cases to include. Gauges show what is due within the hour.
Routing and Auto-Assign
Ordered routing rules send new cases to the right group or person by what they say and who raised them — including the requester's department. Groups share work round robin or to whoever has the fewest open cases, skipping people who are away and handing their work to a backup.
Tasks, Time and Closing Rules
Split a case into tasks with owners, due dates and "waits for" order, from reusable checklists; log time. A case resolves only when its tasks are done; the requester confirms the fix or reopens, and unanswered cases close on their own.
Requester Sentiment
AI reads how the requester feels — positive, neutral, frustrated or angry — and flags unhappy people in the queue so the team can step in before it escalates. Shown to staff only.
AI Case Assistant
On any case, agents can ask for a cited summary, likely causes, diagnostic steps, suggested actions and a draft reply — built from the case and its linked records. Advisory only: it never acts or sends on its own.
Policy-Gated Automation
Built-in runbooks triage incidents, fulfil request types with a prepared answer and tasks, review changes and summarise problems — each leaving a note of exactly what it did; connected tools do outside work. Everything runs within your autonomy policy, with approvals, kill switches and a full audit trail.
Security-Aware Operations
Security alerts, infrastructure state and monitoring alerts appear inside the case an agent is already working — explicit links with the reason for each, not guesses.
Infrastructure Context
A read-only connection to your virtualisation platform shows the state of the machines, nodes and storage named in a case, and raises incidents when a running guest stops or capacity crosses a threshold.
Connector Profiles & Secrets
Integrations are configured as encrypted connector profiles with readiness testing before activation. Private destinations are reachable only when an operator allowlists them, with certificate pinning.
Full Audit Trail
State changes, approvals, AI requests and outcomes, and automation executions are recorded with actor and timestamp, and are exportable for compliance reviews.
Tenant Isolation & Secure Sign-in
Every tenant table is protected by database row-level security enforced under a non-privileged application role, with automated checks on every release. Passwordless sign-in links are single-use, so email link scanners can't use them.
Privacy & Data Retention
Tenant-scoped retention policies, data export and deletion-request workflows, with requesters able to see and withdraw their own privacy requests.
