What to actually evaluate when choosing an RMM + PSA platform

Every RMM/PSA vendor, us included, will hand you a feature-comparison table where their product wins. Feature tables are a real input, but they're the easiest thing to game and the least predictive of whether the platform will actually work for your team a year in. Here is the shortlist of questions we think matters more.

Architecture, not just feature checkboxes

• Does the monitoring agent require any inbound access at any client site? If yes, who owns maintaining and auditing that access over time — you or the vendor?
• Is your data properly tenant-isolated per client, or does the platform rely on application-level filtering that a bug could leak past?
• Are integration secrets (API keys, service credentials) encrypted at rest, or stored in a config table in plaintext?
• What is the platform's posture on outbound network calls it makes on your behalf — does it validate destinations, or will it happily fetch whatever URL an admin types into a settings field?

Data ownership and portability

• Can you export your full ticket history, device inventory, and client data in a usable format, on demand, without opening a support ticket?
• If you left tomorrow, what would you actually be able to take with you?

What happens on your worst day

• If the vendor has an outage, what visibility do you have into your clients' networks in the meantime?
• If an AI feature is part of the platform, does it take autonomous action, or does a human stay in the loop before anything executes? Ask this specifically — "AI-powered" is not a specification.
• How is a security incident on the vendor's own infrastructure disclosed to you, and on what timeline?

We're building Nexus to answer every one of these questions well — outbound-only agents, per-tenant isolation, encrypted secrets, SSRF-guarded outbound calls, and an AI layer that drafts but never auto-executes. We'd rather you ask us these questions directly than take a feature table at face value, from us or anyone else.