What happens when something fails, and when we are away.

What we monitor

Every flow and every agent we run.

• Every run is logged: when it ran, with what input and with what result.
• Automatic alerts fire when a step fails, when something that should have run did not, and when a connection stops responding (for example, an expired credential).
• Monitoring is automatic and runs around the clock. Human support does not: it follows the hours below.
• In the client portal you can see what is active and what was approved.

Support hours

What we can actually deliver, in writing.

• Monday to Friday, 9 am to 6 pm Uruguay time (UTC−3), on business days.
• Outside those hours monitoring keeps running and alerts are logged, but human attention starts on the next business day.
• We do not offer 24/7 on-call. If your operation needs someone answering at 3 am, we would rather tell you before you sign.

When something fails

What happens, in order.

• Critical failure (a flow is down, or an error reaches your customers): first response on the same business day.
• Non-critical failure (it works, but worse than it should): first response within 2 business days.
• First response is not resolution: it means someone is on it and tells you what happened and what comes next. How long the fix takes depends on the cause.
• We design every flow so that if a step fails, it stops instead of carrying on with wrong data. Whatever was pending is logged so it can be resumed.
• Each recurring service contract can set specific times per flow, depending on how critical it is to your operation.

What support includes

And what it does not.

• Everything we deliver comes with a 30-day warranty from handover: if it does not do what the proposal says, we fix it at no cost.
• With a recurring service (AI Partner or Managed Operations) that coverage continues while the service is active: fixing failures in what we built and adjustments within the contracted scope.
• Anything that changes the scope (a new flow, another tool, another process) is quoted separately, with a written proposal.
• The exact detail for each service is in the Annex signed with the contract.

What depends on third parties

And we cannot fix ourselves.

• Your automations rely on other people's services: WhatsApp and Meta, Google, your CRM, AI model providers, hosting.
• If one of them goes down or changes its rules, we cannot bring it back sooner. What we do: detect it, stop the flow safely, let you know and resume what was pending once the service is back.
• When a provider changes its API, adapting to it is covered by support if it is within the contracted scope.

If the technical lead is away

The question every boutique has to answer.

• We are a boutique team: the person who builds your solution is the same person who maintains it. We will not tell you there is an on-call team that does not exist.
• If that person is unavailable, monitoring keeps working: a failing flow stops safely and you get the alert.
• Planned absences are announced in advance, and before each one we check that anything critical is stable.
• Every solution is documented — how it works and how to pause it — so your team or another engineer can step in if needed.

Documentation and control

What is yours stays yours.

• Everything runs on accounts in your name. If we stop working together tomorrow, nothing switches off.
• You receive the documentation of what was built: what it does, how it is configured, how to pause it and how to switch it off.
• You can revoke our access whenever you want, without depending on us.
• An honest caveat: owning the accounts and the documentation makes continuity possible, but it does not replace maintenance. If nobody maintains it, an automation can stop working when a provider changes something.

How we protect access, credentials and data is on the security page.