New Finds
Rovo: The Teamwork Graph-Powered AI Layer
Atlassian's pivot from "Atlassian Intelligence" to Rovo as the flagship AI platform. Built on the Teamwork Graph — a unified context layer spanning Jira, Confluence, and assets. Rovo delivers search, chat, and autonomous agents in one coherent stack.
- Teamwork Graph: The moat. Only delivers full value if org is already deep in Jira/Confluence ecosystem.
- Rovo Search: Cross-platform AI search with natural language understanding.
- Rovo Chat: Conversational interface to query and act on work data.
- Rovo Agents: "Your workflows, now on autopilot" — autonomous agents that keep work moving across Jira, Confluence, and Assets.
- MCP Support: Model Context Protocol integration for custom tool connections.
JSM Virtual Agent + Rovo Customer Service Agent
Two-tiered AI support: Virtual Agent for basic triage, Rovo Customer Service Agent for full resolution. The Virtual Agent precedes Rovo but remains relevant for simpler workflows.
- Virtual Agent: Native to JSM, handles ticket triage, 1,000 assisted conversations/month free.
- Rovo Customer Service Agent: 24/7 support powered by Confluence/JSM knowledge base, $1.00 per resolution.
- Rovo Dev: CLI-first interface for developers with long session memory for complex tasks.
- Assets Integration: CMDB-aware automations — 5k/50k/500k objects free depending on tier, then $0.02/object.
How Atlassian's AI Stack Layers
Teamwork Graph (Context Layer — unified across Jira/Confluence/Assets)
↓
Rovo AI Platform (Search + Chat + Agents)
↓
┌──────────────────┬──────────────────┐
│ Jira SM │ Confluence │
│ (Tickets/SLA) │ (Knowledge Base) │
└──────────────────┴──────────────────┘
↓
Assets (CMDB)
↓
MCP Connectors (Datadog, Slack, Custom)
Key insight: Rovo's value is captive to the Teamwork Graph. No Jira/Confluence investment = degraded AI experience. This is Atlassian's strategic moat against horizontal AI tools.
Pricing Intelligence
Consumption Model Breakdown
Atlassian's hybrid pricing: per-agent base + consumption overages. Unlike competitors' bundled models, this aligns incentives (pay per resolution, not per seat) but complicates MSP forecasting.
Service Collection Tiers (per agent/month):
| Tier | Price | Assets Objects | Virtual Agent |
|---|---|---|---|
| Free | $0 | — | — |
| Standard | ~$17-21 | 5,000 | 1,000/mo |
| Premium | ~$35-41 | 50,000 | 1,000/mo |
| Enterprise | Custom | 500,000 | 1,000/mo |
Consumption Overages:
- Rovo resolutions: $1.00/resolution (pay-per-outcome)
- Virtual Agent conversations: $0.30/conversation above 1,000/mo free
- Assets objects: $0.02/object/month above tier limits
MSP Challenge: Per-resolution pricing doesn't map cleanly to per-agent/per-endpoint MSP billing models. Requires custom cost allocation.
Worth Watching
Where Atlassian Fits — The Full Picture
| Capability | Atlassian | ServiceNow | Atera | Freshworks | HaloPSA |
|---|---|---|---|---|---|
| Autonomous resolution | Rovo $1/res | ITSM Agent (GA) | AI Copilot (assist) | Freddy Agent (80%) | No |
| Knowledge base AI | ✓ Confluence-native | KB Generation | Auto KB from tickets | Limited | AI Report Builder |
| PSA integration | None | FW Integrations | ✓ Native | ✓ Native | ✓ Native |
| RMM/endpoint | None | ITOM hooks | ✓ Built-in | Limited | 50+ integrations |
| Pricing model | Per-agent + consumption | Enterprise+ | PSA + add-on | CSM-tier bundled | Per-agent PSA |
| MSP-specific features | None | Limited | ✓ Core | ✓ Core | ✓ Core |
Atlassian's gaps: No native PSA, no MSP billing workflows, no RMM, consumption pricing misaligned with MSP unit economics. Their play is enterprise IT teams already in Atlassian — not MSPs.
The MSP-Unfriendly Pricing is the Opening
Atlassian's consumption model ($1/resolution) is innovation-friendly but MSP-hostile. MSPs need predictable per-seat or per-endpoint pricing to bill clients. Atlassian's answer: "custom integration."
- SuperOps opportunity: Offer autonomous resolution + flat per-agent pricing that MSPs can actually resell.
- Atlassian's weakness: No PSA layer, no MSP workflows, no client billing integration.
- Teamwork Graph lock-in: Only works if customer is already Jira-deep. Greenfield MSPs have no sunk cost to trap them.
If SuperOps builds agentic resolution at MSP-friendly unit economics, Atlassian's $1/resolution becomes a competitive disadvantage, not a feature.
Action Items
- Microsoft Copilot for Service: Dynamics + Power Platform angle for MSPs
- Zendesk AI: How their $1.10/agent/day pricing compares
- PagerDuty + AIOps: Incident response automation landscape
- SuperOps AI roadmap: Internal competitive positioning audit