Webex Contact CenterCX Hive

Webex Contact Center, November 17: One Desktop URL and AI Agent Oversight That Follows the Customer

Cisco is retiring data-center Desktop URLs in favour of wxcc.cisco.com and giving AI Agent Studio an org-wide observability dashboard. Here is the migration and the reporting rebuild.

Two changes landed for Webex Contact Center on November 17. The first is a hard deadline with a network change attached: every agent Desktop moves to a single URL, wxcc.cisco.com, and the data-center URLs your agents use today stop working in February 2027. The second is a reporting change with a quieter sting: AI Agent Studio gains an organization-wide observability dashboard and a cross-agent Sessions view, and the per-agent Analytics page it replaces is already gone for Common Identity tenants. One is a migration you schedule; the other is a reporting habit you rebuild.

What changed

Cisco is moving Webex Contact Center Desktop away from data-center-specific URLs to a single common URL, wxcc.cisco.com. Cisco says the new URL is usable from November 16, 2026, with expected General Availability on November 30, 2026. The older data-center URLs keep working until February 8, 2027; after that they redirect to the new Desktop for two weeks and are then taken down. Cisco asks administrators to allow the new URL on firewalls and to update custom widgets that reference Desktop URLs, and says the common URL brings faster logins and lets client changes roll out progressively.

Separately, Cisco moved the Observability Dashboard for AI Agents to What's New. It gives an organization-level view of performance metrics across scripted and autonomous AI agents on voice and digital channels, with filters by agent, agent type, digital channel and language, trend comparison, a personal reporting time zone and per-metric warning thresholds. The individual-agent Analytics page is no longer available for Common Identity (CI) enabled AI Agent Studio tenants; non-CI legacy tenants keep it. Cisco also published a top-level Sessions experience that groups interactions involving multiple AI agents under one Interaction ID, so administrators can search by that identifier and supervisors can review timelines, handoffs, timestamps and channel context. Per-agent Sessions views remain.

Full records: Webex Contact Center release intelligence and the official Cisco What's new for administrators page (opens in a new tab). The updates behind this article are Webex Desktop moves to one URL, wxcc.cisco.com; data-center URLs stop after February 2027 and AI Agent Studio adds an observability dashboard and cross-agent sessions; Analytics page retired.

Pattern 1: Run the URL migration like a network change, not a bookmark edit

The tempting way to handle this is to send agents a new bookmark and move on. The failure mode Cisco is warning about says otherwise: an agent who cannot reach wxcc.cisco.com through a proxy or firewall cannot log in at all once the old URLs are retired, and every bookmark, SSO tile, kiosk shortcut and widget that hardcodes a data-center URL breaks on the same day. Treat it as a network change with a checklist, and run it while the old URLs still work.

Start with the allowlist. Confirm wxcc.cisco.com is reachable through every path an agent might take: the corporate proxy, cloud security gateway, VPN split tunnel, and content filter. Then test an agent login on the new URL from each path. A URL that resolves is not a URL an agent can sign in through.

Then hunt the hardcoded references. Search Desktop layouts and custom widget code for data-center-specific Desktop URLs. SSO tiles in your identity provider are the ones most often forgotten, because they live outside the contact-center admin's view. So are shared-desktop shortcuts, intranet links, and deep links inside a CRM or ITSM record. Each one gets the new URL, or it gets a redirect it cannot rely on after the two-week grace period ends.

Use the window Cisco gave you. The new URL is usable now and the old ones work until February 8, 2027, with a two-week redirect after that. Run the migration in November and December, when a missed reference surfaces as a ticket you can fix calmly, not in February, when it surfaces as an agent who cannot start a shift.

Pattern 2: Rebuild your AI agent reporting on the org-wide dashboard

If your tenant is Common Identity enabled, the per-agent Analytics page is already gone, so this is not a someday task. The rebuild starts with a list of what that page was used for: the daily containment check, the weekly review, the export someone pasted into a slide. Each check has to be rebuilt in the Observability Dashboard — a chance to make them better, because the new dashboard sees across agents instead of one at a time.

Set the dashboard up around the questions you actually ask. Filters by agent, agent type, digital channel and language let you split a blended number into its parts: a scripted agent and an autonomous agent can share a queue and still deserve separate scrutiny. Trend comparison is the feature to lean on hardest — a containment rate means little as a single number, and a lot as a line that bent the week you shipped a new flow. Pick the metrics you track, set a warning threshold on each, and agree in advance who acts when one trips. A threshold nobody owns is decoration.

Then connect the dashboard to the rest of your measurement. The dashboard tells you what the AI agents did; it does not tell you whether the customer left happy. Pair the containment view with your survey results and with the CRM record, so a contained interaction that produced a complaint is visible as a failure, not a success. If escalations land in ServiceNow or Zendesk, the handoff rate should reconcile with the ticket volume there; a gap between the two is where misrouted customers hide. Start with a weekly review where containment, escalation rate, and surveyed satisfaction for AI-handled contacts sit side by side.

Pattern 3: Use cross-agent Sessions to debug the journeys, not just the agents

The Sessions view fixes a real debugging problem. When a customer talks to a scripted agent, gets handed to an autonomous agent, and then reaches a human, the old per-agent views showed three separate fragments. The top-level Sessions experience groups those interactions under one Interaction ID, with timelines, handoffs, timestamps and channel context in one place. That turns a common complaint — "the bot kept me in a loop" — from an anecdote into something you can replay.

Build a testing habit around it. When you ship a new multi-agent flow, run test conversations through the full journey and pull them up by Interaction ID. Watch the handoffs: did the second agent know what the first one had already asked, or did it repeat a question? Did the customer repeat themselves to the human agent? Each repeated question is lost context, and each one is a design fix — usually a shared variable or a handoff summary — rather than a model problem.

The same view helps the human side of the handoff. A supervisor reviewing a session can see exactly what the AI agents tried before the customer reached an agent, so coaching gets specific: the agent can see the transcript of the automated attempts and pick up where they left off. Over time, the sessions that escalate most often are your roadmap for the next flow to build. If the same question keeps reaching a human after two AI attempts, that question is the next automation candidate — or the next knowledge article to write.

Summary

November 17 gives Webex teams two jobs with different clocks. Allow wxcc.cisco.com everywhere your agents log in, update every hardcoded reference, and test the new URL now, while the old ones still work until February 8, 2027. And rebuild your AI agent reporting on the Observability Dashboard, with thresholds someone owns and a weekly review pairing containment with satisfaction — then use Sessions to find the handoffs where customers repeat themselves. The URL change is a deadline; the dashboard change is a better way to see your automation.