ZantIQ Administration Onboarding Guide

Navigate to Settings → Security → SSO

Only Admins can access SSO configuration. You will need your IdP's metadata URL or XML, and ZantIQ's SP metadata to paste into your IdP.

Retrieve ZantIQ SP metadata

Copy the ACS URL and Entity ID from the SSO setup page. These are the values you paste into your identity provider to register ZantIQ as a service provider.

Configure attribute mapping

ZantIQ requires the following SAML attributes to be sent in the assertion:

Paste your IdP metadata and test

Paste your IdP's metadata URL or XML into the ZantIQ SSO form and click Test SSO. ZantIQ will perform a test login flow and confirm the attributes received. Once the test passes, click Enable SSO.

Enforce SSO (optional)

Toggle "Require SSO for all users" to block password-based login. Ensure at least one Admin account is confirmed working via SSO before enforcing.

Source of Truth

The system that holds the authoritative version of each contract field — typically your CRM (Salesforce or HubSpot). ZantIQ reads from this system and treats it as the "contract truth". Select the source using the Source of Truth dropdown on the Field Mapping page.

Target Systems

All other connected systems (Jira, Zendesk, Vitally, etc.) are potential targets. ZantIQ pushes — or validates — that the source field values match what is configured in each target.

Mapping Rules

Each rule maps one source field to one target field, with an optional Transform expression (JavaScript) that converts values between systems. Rules have a status of Active or Conflict — a conflict means the current target value does not match what the contract specifies.

Select Source of Truth and Target System

Use the two dropdowns below the pipeline header. The canvas updates immediately to show the Source Fields (left panel) and Target Fields (right panel) for the selected pair.

Click a source field to begin mapping

The field row highlights in indigo and the canvas shows the instruction "Now click a target field to map". Click a field in the Target Fields panel to complete the mapping. A bezier line is drawn between the two fields.

Review the Active Mapping Rules panel

The bottom-left panel lists all rules for the current source→target pair. Each row shows source field → target field → transform expression. Hover a row and click the trash icon to delete a rule.

Check the Propagation Status panel

The bottom-right panel shows the sync status and last-sync time for each target system, and flags how many rules have conflicts per system.

Enable Layer 3 for the tenant

Only admins see the Layer 3 Push Config panel at the top of Field Mapping. Toggle push_enabled, list the connectors that accept push in Allowed target connectors, and set a per-connector daily push cap. Layer 3 is off by default.

Request a push (Maker)

On a mismatched row, click Push. The 3-step wizard runs a live preview, surfaces server-side warnings (numeric-looking value being replaced, non-empty target about to be overwritten, day-cap approaching), and requires typing PUSH to confirm. Instead of pushing, choose I'll fix it manually + rescan to update the target directly in the vendor UI and then re-verify.

Approve (Checker, different admin)

A different admin opens the Push Request Queue. The row is chip-flagged if it is the current user's own request; the API refuses self-approval regardless. On Approve, the server re-reads the live source value and refuses with status stale_source if it has drifted since the request was created.

Revert if needed

The Push History timeline shows every push, rescan, and revert. Any successful push has a Request revert button — clicking it creates a new push request that writes the original target value back. A different admin still needs to approve it before the revert lands on the vendor.

Pick the trigger

Current triggers: breach.created, breach.status_changed, breach.resolved. commitment.mismatched and obligation.upcoming ship in v2. Every enabled rule for a matching trigger evaluates in parallel per event.

Build the condition tree

Nested AND / OR / NOT groups over leaf comparators (eq, gte, in, contains, is_null, and more). Fields are dot-paths on the resolved context — e.g. customer.plan, breach.type, customer.arr, commitment.contract_value. Missing fields resolve to None and never raise. Depth ≤ 8, ≤ 64 nodes.

Pick the actions

Ordered list. Four action types ship in v1: slack.notify (post to a channel), email.send (transactional email), ticket.open (Jira / Linear / Zendesk / Freshservice), health.set (Vitally / Gainsight health score). Templates use {customer.name}-style interpolation from context; missing values render <unknown>.

Dry-run before enabling

POST /rules/{id}/test with a real resource_id returns the full condition trace + the interpolated action payloads without dispatching. Verify the tree evaluates the way you expect before flipping enabled=true.

Review execution history

Every fire — matched, throttled, no-match, or error — records a RuleExecution row with the trace and per-action outcome. GET /rules/{id}/executions returns the last 50 by default; the audit log also captures every rule create / update / disable.

Declare an output schema

Each extractor declares 1–20 output fields. Each field has a name (lowercase snake_case), a type (string / integer / number / boolean), a description, and a landing rule: Enforce as Commitment creates a Commitment row on every contract (runs through breach detection + rules engine + health score), Reference on Contract lands as passive metadata in Contract.extracted_fields. Mix both in a single extractor.

Write the prompt template

Free-form text; must contain the literal {text} placeholder — that's where the contract text is substituted at run time. The service wraps the tenant prompt with a system-authored preamble that forces the LLM to return JSON matching the declared schema. Extras the LLM produces are dropped; missing fields are logged as validation errors.

Test before activating

POST /extractors/{id}/test with a contract_id runs the extractor against that contract's stored text and returns the extracted fields plus a landing-rule preview. No Commitments are created and Contract.extracted_fields isn't written. Use this to iterate on the prompt before flipping status to active.

Activate + Pause lifecycle

Draft extractors never run. Active extractors fan out during every customer contract ingest, vendor contract ingest, and background ingestion. To edit an active extractor, click Pause (moves to draft) first — this prevents in-flight ingestions from seeing a schema that no longer matches the prompt. After 5 consecutive failures, the extractor is auto-moved to draft with the last error visible on the row.

Provenance on every Commitment

Commitments carry an extracted_by_extractor_slug column — system for platform prompts, the tenant's slug for custom extractors, NULL for manually-created rows. Exposed on the API today; UI chip on Commitment rows ships as a follow-up polish.

Go to Settings → Team

The Team page shows all active users, pending invitations, and a list of recently deactivated accounts.

Click "Invite User"

Enter the invitee's work email address. Select a role. Optionally set a CSM assignment so this user is pre-linked to specific customer accounts.

Bulk invite via CSV

For onboarding a team, click "Import CSV". Format: email,first_name,last_name,role. Each person receives an individual invite email.

Deactivate users who leave

When a team member leaves, click the three-dot menu on their row → Deactivate. This immediately revokes their access. Their historical activity (audit logs, comments) is preserved.

Filtering & search

Use the search bar to filter by customer name. Use the tier buttons — All, Compliant, Review, At Risk — to focus on customers that need attention.

Customer detail view

Click any customer row to open their detail page. The detail view contains tabbed sections: Overview, Contracts, Compliance, and Connectors.

Health tiers

Compliant health score ≥ 90% · Review 70–89% · At Risk <70%. These tiers drive alert priority and QBR preparation workflows.

Upload the document

Click Ingest Contract on the Contracts page. Drag and drop or browse for a PDF or DOCX file. Maximum 50 MB.

Automatic extraction runs

ZantIQ's AI engine parses the document (15–60 seconds). It extracts: contract type, customer name, contract value, start and end dates, SLA tiers, response and resolution time commitments, seat counts, uptime guarantees, penalty caps, and escalation paths.

Review confidence score

The Extraction Confidence column shows how confident ZantIQ is in its extraction (0–100%).

• 85%+ — Ready to use. Commitments are auto-activated.
• 70–84% — Review recommended before activating enforcement.
• <70% — Manual review required.

Commitments are auto-created

Navigate to Commitments to see every obligation extracted from the contract as a separate row — type, contract value, actual value (pulled from connected systems), and compliance status.

Customer filter

Use the Customer dropdown in the top-right to focus on a single customer. Selecting "All Customers" shows aggregate compliance across your entire book of business.

Compliance Score donut chart

Shows the percentage of commitments that are compliant for the selected customer. The centre number is the headline compliance rate.

Status Breakdown bar chart

Shows commitment count by status (Compliant, Breach, Mismatch, Missing, etc.) to identify which problem types are most prevalent.

Non-Compliant Commitments table

A sortable table of every commitment that is not Compliant. Columns: commitment name, customer, type, contract value, actual value, status, and source system. Use this as your weekly action list.

Customer dropdown

Select a specific customer or view all customers in aggregate. The stat cards, charts, and source breakdown table all update to reflect the selection.

SLA Attainment chart

A grouped bar chart showing Met % vs. Breached % for each priority tier (P1, P2, P3). P1 issues are your most urgent — a Breach here typically triggers financial penalties per contract terms.

Response Time chart

A side-by-side bar chart comparing Actual response times versus the Contracted target for each priority tier. Bars above the target line indicate underperformance.

Overall SLA Health stat cards

Four headline metrics at the top: Met, Breached, At Risk, and Mean Response Time. These update live as ticket data flows from Jira, Zendesk, or Pylon.

Connect the Slack integration

Go to Connectors → click Connect on the Slack tile. Authorise the ZantIQ Slack App to access your workspace. ZantIQ will request channels:read and chat:write scopes.

Set notification channels

Go to Settings → Notifications → Slack. Configure a default channel for breach alerts (e.g., #cs-alerts). Optionally, set per-customer channels so that TechCorp alerts go to #cs-techcorp.

Set email notification recipients

Go to Settings → Notifications → Email. Add email addresses to receive daily digest reports and real-time breach alerts. The assigned CSM for a customer is automatically copied on breach alerts for their accounts.

Notice period configuration

Go to Settings → Renewals to set your default notice periods (e.g., 90 days, 60 days, 30 days). Contracts entering each threshold receive a notification.

Renewal status tracking

Mark renewals as In Progress, Won, or Churned to maintain your renewal pipeline within ZantIQ alongside your CRM.