About

Built for the people who answer the calls.

MSP technicians carry context for dozens of clients in their heads. TierSignals is a signal channel that keeps the right context next to them, in real time, while the call is live.

Why we built this

Tier 1 technicians spend a meaningful share of every call paused — pulling up the PSA, searching a runbook, remembering which client has Duo and which uses Microsoft Authenticator, and trying to recall whether a permission change needs manager approval.

Those pauses show up in average handle time, ticket quality, and CSAT. They also show up in technician fatigue. TierSignals removes them by listening to the call, applying the client's PSA context, and surfacing short, operational guidance the technician can act on or ignore.

We won't touch a client system. We won't open or close a ticket. We won't autopilot a password reset. That's not a hedge — it's the product.

Advisory only

Suggest, don't act. Every recommendation is a card the technician reads, edits, or ignores.

Technician controlled

No autonomous changes on client systems. Ever. The technician owns the keyboard.

Operational, not theatrical

Short, useful cards beat long monologues. The feed is a tool, not a chat.

Built with service desks

Designed with MSP Tier 1 and Tier 2 leads who know the friction by heart.

A short history

From shadowing technicians to shipping a copilot.

We didn't start in a research lab. We started with a notepad next to a service desk. The product is the residue of those calls.

  1. 2024
    First conversations
    Spent four months sitting next to Tier 1 technicians at three different MSPs.
  2. 2025
    Private prototype
    Built a rules-only signal engine to validate that in-call guidance changes outcomes.
  3. 2026
    Production hardening
    Added Recall ingestion, workspace RLS, role-based access, audit logs, and a pilot smoke test framework.
  4. Now
    Pilot cohort
    Running structured 30-day pilots with a small group of MSPs in the US and EU.
What we won't do

A short list, on purpose.

  • We won't autopilot actions on client systems.
  • We won't sell your transcripts or train shared models on them.
  • We won't add a feature that the service desk hasn't asked for twice.
  • We won't price by tokens, calls, or other proxies for usefulness.