The card for Agent-first
A team that wants agents to run investigations and uses playbooks only where a fixed, audited step is needed. It is testing whether an agent-run platform can take whole alert categories.
AI carries 24 and approvals 18, because the questions are what the agent can do, what it cannot, and who signs off an action. Ownership carries 16, since vendors in this group are young and funded by recent rounds: Torq announced a USD 140M Series D on 2026-01-12 and Mate Security a Series A on 2026-07-28. Migration drops to 4 and hosting to 4. Accuracy and closure percentages never lift the AI score above 8.0 without a published method, so a vendor cannot buy a place here with a number. On these weights Mate Security rises to third behind Cyware and Torq, on three documented autonomy modes and a Series A announced on 2026-07-28. Its limits on this card are the same as elsewhere: no deployment model, no integration count and no published method for its accuracy figures.
The covered platforms, in this order
- 1 Cyware 7.65 Opens a band.
- 2 Torq 7.40 Band with Cyware: within 0.25 of the top.
- 3 Mate Security 6.91 Opens a band.
- 4 Tines 6.76 In the same band as the row above.
- 5 Splunk SOAR (Cisco) 6.70 In the same band as the row above.
- 6 Swimlane (Turbine) 6.67 In the same band as the row above.
- 7 D3 Security 6.62 Opens a band.
- 8 Google Security Operations 6.42 In the same band as the row above.
- 9 BlinkOps 6.04 Opens a band.
- 10 Microsoft Sentinel 5.94 In the same band as the row above.
- 11 Palo Alto Cortex XSOAR 5.92 In the same band as the row above.
- 12 ServiceNow Security Operations 5.44 Opens a band.
The weights used
| Row | Weight on this card | Weight on the main card |
|---|---|---|
| Playbook authoring model | 12.0 | 14 |
| Integration count and depth documented | 10.0 | 14 |
| Case management and collaboration | 6.0 | 8 |
| Approvals and human-in-the-loop controls | 18.0 | 12 |
| Agentic or AI features that are documented | 24.0 | 12 |
| Hosting and data residency options | 4.0 | 8 |
| Pricing model and transparency | 6.0 | 8 |
| Migration tooling and import or export formats | 4.0 | 14 |
| Ownership stability | 16.0 | 10 |
Totals are computed from the same nine scores as everywhere on the site, reweighted. Totals within 0.25 are a band, not a ladder. How the scoring works.