Service Levels

What Sí Call does and does not commit to on availability, and what actually happens when something breaks.

We do not publish an availability figure, and this page used to. It carried monthly uptime targets until 2026-08-07. They were removed because Sí Call has never measured uptime — there is no production deployment, no external monitoring, and nothing that would notice an outage. A published percentage is read as a measurement, and we had no measurement behind it. The specific figures are deliberately not repeated here: quoting them would put the same numbers back on the same page.

We would rather tell you that than print a number we cannot defend. When the monitoring exists and has run long enough to mean something, this page will carry a figure and will say when it started counting.

1. What we commit to today

Nothing numerical, and no service credits. There is no availability commitment on this page, for standard or government tenants, because there is no instrument behind one. Anyone who needs a contractual availability term should ask for it in writing at legal@sicall.ai — it would be a negotiated term with a defined measurement method, not a figure lifted from a marketing page.

2. What actually happens when something breaks

The honest answer, which is the part most vendors leave out:

• Nothing automated notices. There are no external uptime checks and no alert routing to a phone. If the service stops answering calls at 2am, no alarm goes off. Building that is scheduled work, not a thing we are quietly assuming.
• You telling us is the detection path. Write to support@sicall.ai. That reaches people, not a queue, and it is currently the fastest way for us to learn something is wrong.
• We are a small team without a 24/7 rota, so we do not publish a response time. An hours-of-the-day answer we can keep is worth more than a minutes-to-respond number we cannot.
• Plan your line so we are not a single point of failure. Call forwarding is reversible from your carrier in minutes — that is your fastest fallback and it does not depend on us being awake.

3. Incident severity

How we classify what has gone wrong. This is a shared vocabulary for talking about an incident, not a response-time commitment — see section 2 for what actually happens.

For anything we would class P1 or P2, we contact affected customers directly and write up what happened.

4. Questions

Related Documents