An Aivell checks your key numbers against your own database every hour, compares them with the usual pace, and speaks up only when a threshold is crossed — with context: what moved, since when, and how unusual it is. No dashboard to open, no daily noise, nothing leaving your network.
starts when: Every hour, the schedule checks the numbers
Metric-watching fails in two directions. Either there’s a dashboard that nobody opens — the refund spike sat there for nine days, plotted beautifully, unseen — or there’s an alerting rule that cried wolf twice a day until everyone filtered it to a folder. Both end the same way: you learn about the problem from a customer, or from the month-end numbers.
The middle ground takes judgment. “Tuesday mornings are always slow” is knowledge a threshold doesn’t have; “under pace for a Tuesday” is the alert you actually want. Encoding that in a monitoring tool is a project. Saying it to a colleague takes one sentence.
The guardrails keep it honest: the database connection cannot write, and the mail cannot leave your domain. Your trading numbers are computed onboard and stay behind your firewall — no metrics platform holds a copy. A colleague who watches your numbers all day, inside your walls, and only taps your shoulder when it matters.
No flowcharts, no code — a chore is just a message to your Aivell, in your own words. It shapes it into a solid, guarded procedure and follows it to the letter. This one:
every hour between 8 and 20, check the day's order count and revenue
compare with the same weekday over the past four weeks
revenue tracking 20% under the usual pace? email Paolo with the numbers
a single order above €10,000 — flag it for a manual check, quietly
refund rate above 3% on any product — alert, with the top three offenders
stay silent when everything is normal; no daily all-fine mail
if the same alert fired in the last 24 hours, don't repeat it
friday evening, send a one-paragraph recap of the week's alerts
Connected with single-click passkeys — no copied tokens, no OAuth hell.
Every tool can be limited to exactly what this chore needs. For this one, you might set:
Two ways. The rules are sentences — "20% under the usual pace for this weekday" — not formula builders. And the alert arrives with context already attached — what moved, since when, against what baseline — so you don't open a dashboard to find out why it fired.
Silence is the default. The chore says when to speak — thresholds, deduplication within 24 hours, a weekly recap — and you tune those lines the way you'd brief a colleague who pinged you too often.
Entirely on the box. It queries your database across your own network and reasons onboard. No metrics platform, no analytics vendor, no outside copy of your trading numbers.
Postgres, MySQL, Oracle, SQL Server — anything reachable with a standard driver — plus Google Sheets, if that's where a number lives.
Your Aivell queries the database, builds the weekly report and mails it out on schedule — computed onboard, so your numbers stay behind your firewall.
Type a question in the chat, get the answer from your own SQL data — queried read-only, computed onboard, with the query shown for the curious.
Order status, returns and price checks handled by a private AI colleague — onboard inference, so customer data never leaves your shop.
Booking inquiries answered in the guest's language, reviews digested, arrivals prepped — an AI colleague that never sends guest data to the cloud.
An Aivell is an on-premise AI colleague: a small box that plugs into your network, runs its own AI onboard and takes the repetitive work off your desk. Prompts, documents, data — nothing ever leaves your office. Unplug the internet: it keeps working.
You describe each task as a chore, in plain language. Your Aivell turns it into a solid, guarded procedure and runs it in the background — with guardrails on every tool, a complete audit trail, and approvals in your hand when you want them. One-time setup, fixed monthly fee. No tokens, no overages.
onboard inference · air-gapped · fixed monthly fee