Understand the operation
Trace ticket lifecycles, agent actions, repeat contacts, and data quality. Find what creates the work before automating it.
Audits & systems architectureAmanda James · CX systems architect
The workflow. The data. The tools that don’t talk.
jrnyCX connects the
pieces so your team can do the work.
The customer asked once. The answer lives in three systems.
CONNECTED CONTEXT
History + commitments
Order + delivery status
People + relationships
Replacement dispatched. Tracking is available. The agent can review the details and reply without searching three tabs.
Explore the custom app demo →Illustrative workflow · synthetic data · no customer systems connected
01 / The approach
Better support starts with understanding what happens between the ticket, the people, and the systems behind it.
Trace ticket lifecycles, agent actions, repeat contacts, and data quality. Find what creates the work before automating it.
Audits & systems architecturePut backend data in the agent’s workspace. Design the relationships, integrations, and custom apps the process actually needs.
Custom apps & integrationsBuild intake and knowledge around reliable context. Define what AI can handle, what needs confirmation, and when a person takes over.
AI agents & human handoffs03 / Completed work
Real integrations, sidebar apps, and AI configurations — delivered to production for clients across telehealth, field services, live streaming, and edtech.
Agents were manually checking three external systems mid-ticket to piece together patient history, order status, and billing. A ZAF sidebar app now surfaces the patient journey, open Rx orders, billing status, and escalation routing in one panel — pulling live data through a Cloudflare Workers enrichment layer.
ZAFCloudflare WorkersREST API View app screen →Every inbound call opened with a manual identity lookup, even when the caller's number was already in the system. A ZAF Caller ID app matches the inbound number, resolves the caller's role (customer or contractor), and writes the account data to the ticket before the agent answers. Routing was redrawn with a safety-tier gate for urgent calls.
ZAFZendesk TalkZIS View app screen →Inbound support during events arrived in volume — payout disputes, account questions, and technical failures all in the same stream. A Messaging AI agent was designed with explicit context parameters for session state, payout history, and case evidence. AI-resolvable cases resolve; the rest escalate with the full context already attached.
Zendesk MessagingAI agentZISAutomations had accumulated over several platform iterations — inactive rules, conflicting tags, and custom actions mixed with native ones, with no record of what was still active or why it existed. A full audit mapped every trigger, identified what could be removed, and staged migration in verified sequence without disrupting live tickets.
Zendesk GuidePythoncontent pipeline04 / The build studio
Explore a few ways custom development changes the work.
Demo screens now.
Live integrations to follow.
Already have a backend? The right app brings it into the conversation.
Discuss a custom app ↗05 / Beyond the standard Help Center
A custom Help Center page can be a client project desk: requests, milestones, and shared deliverables alongside the support conversation.
Proposed custom-page experience. Try the preview with a fictional project.
Northstar Co. / Example project
06 / Start with the actual problem
Describe the work you want to change, or book time to talk it through.
Choose a time from Amanda’s Google booking page. The calendar handles availability and your timezone.
If the calendar doesn’t load here, use “Open booking calendar”.