Top parcel delivery company: Scaling support with JSM and Atlassian Intelligence
Company size
Enterprise
Industry
Parcel / Delivery
Technologies
Jira Service Management
Technology
Generative AI
Executive summary
- A leading Polish parcel delivery company modernized its customer support by replacing fragmented shared mailboxes with JSM and Gen-AI.
Client profile
The client is a large parcel delivery and pickup network operating across Poland. Its business depends on reliable last-mile support, strong coordination with partner locations, and fast handling of issues related to delayed, lost, or damaged parcels.
Because the company works with a broad network of couriers, pickup points, and service hubs, support quality has a direct impact on customer satisfaction and brand trust. As shipment volumes increased, the support model needed to become more scalable, measurable, and resilient.
The challenge
The support team managed requests through a shared Outlook inbox fed by multiple public e-mail aliases. While simple to launch, that setup became a serious bottleneck as volumes grew.
- Limited visibility: Managers had no reliable real-time view of workload, backlogs, or response risk.
- Manual processing: Agents spent approx. 2 minutes time rewriting or softening customer messages before forwarding them to external partner locations.
- No trustworthy SLA reporting: Tracking relied on manual Excel logs, which made performance measurement incomplete and difficult to audit.
- Poor scalability: Increasing volume (15% YoY) could only be absorbed by adding more people.
- Fragmented communication: Conversations with internal teams and external partners were spread across e-mail threads instead of being tied to a single service record.
The solution
To address these issues, the company implemented Jira Service Management Cloud and redesigned the support workflow around structured intake, ticket-based communication, and automation.
Core changes introduced
- E-mail-to-ticket conversion: Multiple support addresses were connected to JSM so incoming messages automatically created categorized requests.
- Structured request handling: Predefined request types reduced manual triage and helped agents process issues faster.
- Centralized communication: All interactions related to a case were stored within the originating ticket, improving traceability.
- SLA management: The team introduced response and resolution targets, along with queue visibility and live monitoring.
- AI-assisted drafting: An Atlassian Intelligence automation rule generated a more neutral and polite version of customer messages for agent review before sending.
Implementation approach
| Phase | Key actions | Outcome |
|---|---|---|
| Foundation (3 weeks) | • Set up JSM Cloud Premium • Connected 3 e-mail channels to matching request types | One inbox → one queue; historical context preserved |
| Smart triage (1 week) | • Introduced an automation: “Communicate with pickup-point” • SLA goals: Time-to-First-Response 30 min, Time-to-Resolution 2 h | Agents see priority at a glance; managers get live SLA health |
| Gen-AI pilot (1 week) | • Atlassian Intelligence automation rule drafts a courteous version of the customer’s e-mail • Agent approves / edits via one button • Accuracy scoreboard to unlock full auto-send when > 98 % | 90s reclaimed per ticket; future full automation clearly gated |
Results
The transformation produced operational improvements in both speed and control. Agents spent less time on repetitive edits, requests were easier to track, and managers could monitor service levels during the day instead of relying on delayed reporting.
| Area | Before | After |
|---|---|---|
| Request intake | Shared mailbox with fragmented visibility | Centralized JSM queues with categorized requests |
| Message handling | Manual rewriting and forwarding | AI-assisted drafting with agent review |
| Tracking | Inbox history and spreadsheet logs | Full communication trace inside tickets |
| Service control | Limited KPI credibility | SLA-based operational monitoring |
| Scalability | Growth mainly required more headcount | Higher volume handled with better process efficiency |
Key success metrics
| Metric | Before (shared mailbox) | After (JSM + Gen-AI) | Improvement |
|---|---|---|---|
| Avg. Ticket Handling Time | 17 minutes | 10 minutes | 41% faster |
| SLA Achievement | Not measurable | 96% on-time | Full visibility |
| Agent Touches per Ticket | 7.1 | 4.2 | 40% reduction |
| Peak Season Savings | €0 | ~€46,000 | Cost avoidance |
| CSAT Score | 3.7 / 5.0 | 4.4 / 5.0 | +19% satisfaction |
The support team became better equipped to deal with daily volume and seasonal peaks. New agents could onboard faster, partner communication became more consistent, and operational leaders gained better visibility into bottlenecks.
The road ahead: AI-powered service at parcel-speed
The company plans to graduate the AI tone-polish from “suggest” to “auto-send” once accuracy hits 98 % over a rolling 30-day window.
Next on the roadmap:
- Asset sync – linking pickup-point IDs from the master CRM so agents can raise incidents straight from the asset record.
- Quarterly process reviews – ensuring SLAs and automations adapt with seasonal spikes.
With mailbox chaos behind them and a data-steady service desk in place, the company can scale support volu
