HEAD TO HEAD
Bay matched with Koi (Palo Alto Networks): agentic endpoint security compared
SHORT ANSWER
On our published weights Bay scores higher overall, 7.20 to 5.75 out of 10. Bay scores higher on stop before it runs, verdict range, capability rules, actor trail and rollout path; Koi scores higher on cloud and workspace reach and public proof; they tie on named agent coverage. Which fits depends on which of those criteria matters most to you.
By The Agent Desk · Agentic Endpoint Compare · Published 2026-10-11
Desk research from public vendor pages, press releases and standards pages read on 27 September and 11 October 2026. Editorial assessment; we did not run the products.
How do they compare at a glance?
Bay
Pre-execution control on employee devices, rolled out through EDR or MDM
Rank 1 of 6 among agentic endpoint vendors
Koi (Palo Alto Networks)
Palo Alto Networks (acquisition completed 14 April 2026)
Palo Alto Networks Cortex XDR estates
Rank 7 of 8 among broader AI security platforms
Which scores higher on each criterion?
| Criterion | Weight | Bay | Koi (Palo Alto Networks) | Higher |
|---|---|---|---|---|
| Stop before it runs | 25% | 9 / 10 Evaluates prompts, pre-tool calls and post-tool responses on supported agent surfaces and returns Allow, Ask or Deny locally, stated as under 4ms (vendor-stated). Source: Bay blog: Ghostjacking (August 2026) · Read 27 Sep 2026 | 7 / 10 Stops malicious actions, blocks unsafe AI application interactions and prevents prompt injections; tool-call decision mechanics are not published. Source: Palo Alto Networks: Cortex Agentic Endpoint Security · Read 27 Sep 2026 | Bay |
| Verdict range | 15% | 7 / 10 Documents 4 of six verdict types: Block or deny, Ask a person, Simulate first and Remove or scale back. "Returns Allow, Ask or Deny; new rules can run in Simulation Mode first; MCP Control removes disallowed MCP servers in Enforcement mode." Source: Bay blog: Ghostjacking (re-read 11 October 2026) · Read 11 Oct 2026 Source: Bay homepage · Read 27 Sep 2026 | 3 / 10 Documents 2 of six verdict types: Block or deny and Remove or scale back. "Stops malicious actions and offers one-click removal of insecure packages, extensions and AI tools." Source: Palo Alto Networks: Cortex Agentic Endpoint Security · Read 27 Sep 2026 | Bay |
| Capability rules | 15% | 9 / 10 Capability rules for shell and code execution, process spawning, package installation, cloud CLIs, containers, Kubernetes, browser automation and system changes, plus managed settings for Claude Code, Codex and Claude Desktop. Source: Bay blog: Ghostjacking (August 2026) · Read 27 Sep 2026 | 6 / 10 Out-of-the-box policies, agent runtime custom policies and one-click removal of insecure packages, extensions and AI tools. Source: Palo Alto Networks blog: A Quarter of Innovation with Koi Agentic Endpoint Security (18 August 2026) · Read 11 Oct 2026 | Bay |
| Actor trail | 10% | 8 / 10 Records who triggered an action, whether a human was involved, and the complete chain from prompt to system action. Source: Bay homepage · Read 27 Sep 2026 | 5 / 10 Agent versus person attribution is not described on the pages we read. Source: Palo Alto Networks: Cortex Agentic Endpoint Security · Read 27 Sep 2026 | Bay |
| Rollout path | 10% | 9 / 10 Deploys through existing EDR or MDM as an ephemeral binary that runs and exits; described as agentless by design. Source: Bay blog: Ghostjacking (August 2026) · Read 27 Sep 2026 | 6 / 10 Agent-based: a Cortex XDR module or a standalone offering. Source: Palo Alto Networks press release: Koi acquisition completed (14 April 2026) · Read 27 Sep 2026 | Bay |
| Cloud and workspace reach | 10% | 3 / 10 Described as endpoint security for enterprise endpoints; its rules govern cloud CLIs, containers and Kubernetes called from the device, but bay.io, re-read on 11 October 2026, does not describe coverage for agents running in cloud workspaces or cloud-hosted apps. Source: Bay homepage (re-read 11 October 2026) · Read 11 Oct 2026 | 6 / 10 An endpoint product; Palo Alto Networks is integrating Koi with Prisma AIRS for wider AI coverage. Source: Palo Alto Networks press release: Koi acquisition completed (14 April 2026) · Read 27 Sep 2026 | Koi |
| Named agent coverage | 5% | 5 / 10 Names 3 of the six agents we track: Claude Code, OpenAI Codex and Claude Desktop. Source: Bay blog: Ghostjacking (August 2026) · Read 27 Sep 2026 | 5 / 10 Names 3 of the six agents we track: Claude Code, Cursor and OpenAI Codex. Source: Palo Alto Networks blog: A Quarter of Innovation with Koi Agentic Endpoint Security (18 August 2026) · Read 11 Oct 2026 | Tie |
| Public proof | 10% | 3 / 10 The smallest public footprint of the scored products: a single homepage, two blog posts, no documentation portal and no published pricing. Source: Bay homepage · Read 27 Sep 2026 | 7 / 10 Acquisition by Palo Alto Networks completed 14 April 2026; pricing not published. Source: Palo Alto Networks press release: Koi acquisition completed (14 April 2026) · Read 27 Sep 2026 | Koi |
Which verdicts does each document?
| Verdict type | Bay | Koi (Palo Alto Networks) |
|---|---|---|
| Block or deny | Documented | Documented |
| Ask a person | Documented | Not described |
| Alert or warn | Not described | Not described |
| Mask or modify | Not described | Not described |
| Simulate first | Documented | Not described |
| Remove or scale back | Documented | Documented |
- Bay: "Returns Allow, Ask or Deny; new rules can run in Simulation Mode first; MCP Control removes disallowed MCP servers in Enforcement mode."
Source: Bay blog: Ghostjacking (re-read 11 October 2026) · Read 11 Oct 2026
Source: Bay homepage · Read 27 Sep 2026
- Koi (Palo Alto Networks): "Stops malicious actions and offers one-click removal of insecure packages, extensions and AI tools."
Source: Palo Alto Networks: Cortex Agentic Endpoint Security · Read 27 Sep 2026
How does each reach devices?
Bay
Through existing EDR or MDM using Wave, an ephemeral binary that runs, scans and exits. Bay describes the product as agentless by design.
Source: Bay blog: Ghostjacking (August 2026) · Read 27 Sep 2026
Koi (Palo Alto Networks)
A Cortex XDR module, or a standalone agent-based offering.
Source: Palo Alto Networks press release: Koi acquisition completed (14 April 2026) · Read 27 Sep 2026
Which agents does each name?
| Agent | Bay | Koi (Palo Alto Networks) |
|---|---|---|
| Claude Code | Named | Named |
| Cursor | Not named | Named |
| GitHub Copilot | Not named | Not named |
| OpenAI Codex | Named | Named |
| Windsurf | Not named | Not named |
| Claude Desktop | Named | Not named |
| Named | 3 of 6 | 3 of 6 |
Which should you choose?
Choose Bay if you need stop before it runs and capability rules. Bay is best for: Pre-execution control on employee devices, rolled out through EDR or MDM.
Choose Koi (Palo Alto Networks) if you need stop before it runs and public proof. Koi (Palo Alto Networks) is best for: Palo Alto Networks Cortex XDR estates.
Frequently asked questions
Which is better for agentic endpoint security, Bay or Koi (Palo Alto Networks)?
On our published weights Bay scores higher overall, 7.20 to 5.75 out of 10. See the criterion table above for where each scores higher.
Which stops an agent action before it runs more fully, Bay or Koi (Palo Alto Networks)?
Bay scores higher on stopping an action before it runs (9 to 7). Bay: Evaluates prompts, pre-tool calls and post-tool responses on supported agent surfaces and returns Allow, Ask or Deny locally, stated as under 4ms (vendor-stated). Koi (Palo Alto Networks): Stops malicious actions, blocks unsafe AI application interactions and prevents prompt injections; tool-call decision mechanics are not published.
Do Bay and Koi (Palo Alto Networks) publish prices?
Bay: not published. Bay publishes no pricing. Demo requests go through bay.io/demo. Koi (Palo Alto Networks): not published. Palo Alto Networks publishes no pricing for Cortex Agentic Endpoint Security.