Longstop

What it does

• ✓The alert reaches you on WhatsApp, not by email. Every other tool tells you about your inbox inside the inbox you are already drowning in
• ✓Catches the bank-details scam: a supplier or client emailing that their account has changed is flagged at the highest urgency, left in your inbox and never filed. It is the most expensive email a small business receives
• ✓Knows the difference between something with a clock on it and something that merely arrived — a customer waiting on an order, a payment failure, a deadline, a complaint escalating
• ✓Drafts the reply you write twenty times a week, in your own wording, grounded only in your price list and policies. It never invents a date, a price or a stock level
• ✓Nothing is ever deleted. Marketing and app noise is archived and labelled, and one command puts an entire day back exactly as it was
• ✓We never hold your password. Google and Microsoft stopped allowing that in 2025 — you approve access on their own screen and revoke it from your own admin console in one click

What does it actually catch that we would miss?

Whatever has a clock on it and is sitting unread. A customer whose order is twelve days late and is now asking for a refund. A supplier saying their bank details have changed. A payment that failed and has stopped your account working. A regulator or a department writing with a response window. Those arrive in the same inbox as the newsletters and the platform digests, and by sheer volume the noise wins. The assistant separates them and puts the first group on your phone.

How is this different from a tool we could install ourselves?

Tools like SaneBox or a Gmail add-on hand you a dashboard and leave the tuning to you. This is operated. Paizah connects it inside your own account, writes the material the drafts are built from, sets what counts as urgent for your business, and reviews it monthly. You get a quieter inbox and a WhatsApp when it matters, not another tool to learn.

What stops it filing something that mattered?

Several things, enforced in the software rather than promised. Mail from your own domain is read-only. Anything read as a customer, a supplier, an invoice or a legal matter is never filed out of the inbox, whatever the AI concludes. When it is unsure it labels instead of filing, so an uncertain answer becomes more visible, not less. And every change is logged and reversible: the daily report lists exactly what was touched, with the command to undo all of it.

Will it send email as us?

No. It writes a draft and stops. Sending is a separate permission we do not hold at all, so a fault in the software cannot put a message on the wire over your name. If you later want a narrow set of routine replies sent automatically, that is unlocked deliberately, one category at a time, after your drafts have gone out unedited for a month.

We already have rules and folders. Why is this different?

A rule matches a sender or a word. It cannot tell that the customer writing "quick question about my order" is the one who has been waiting twelve days, and it cannot tell a routine acknowledgement from a demand. It also cannot put anything on your phone. Rules handle the easy part; this is for the part that needs judgement, where getting it wrong costs money.

How long before it is useful?

It runs in observe-only mode for the first fortnight, changing nothing, while you read the daily report and tell us what it got wrong. That tuning is included in the setup fee and it is the point — every business has its own definition of urgent. It goes live on your word, not on a timer.

Frequently asked questions

What does it cost?

Priced per inbox, per month. Watch is ₹1,500 and gives you triage, the WhatsApp alert and the daily report, touching nothing. Triage is ₹3,000 and adds filing, with full undo. Managed is ₹5,000 and adds drafted replies in your own wording plus monthly tuning. Setup is ₹15,000 once, covering the connection, your rules, and the two supervised weeks. Internationally: $49, $109 and $179 per inbox per month, $499 setup. Four or more inboxes take 20% off, ten or more 35%. Minimum two inboxes.

Which inbox should we start with?

The shared one that everybody reads and nobody owns — usually info@, sales@ or orders@. That is where things go missing. Personal inboxes are worth adding second, once you have watched the first one for a month.

Do you need our email password?

No, and for Gmail or Outlook nobody can use one any more: Google stopped accepting passwords for this in March 2025 and Microsoft has done the same. You approve access on Google or Microsoft’s own screen, inside your own account, and you can withdraw it yourself at any time without asking us. For older mailboxes on cPanel or Zoho, where a password is still the only option, we use an app-specific password you can revoke on its own.

Who can read our email?

The automation reads it to sort it. It runs against your own Google or Microsoft account rather than a Paizah mail server, so your mail does not move. We sign a data processing agreement before anything is connected, and for UK and EU clients that includes the GDPR paperwork naming every party involved. Nothing in your mailbox is used to train any model. A named person at Paizah reads the daily report during setup and at the monthly review — the report, not the mail.

What happens when it gets something wrong?

It will, and the service is built so that being wrong is cheap. Nothing is deleted, everything is reversible with one command, and the report shows you every action rather than hiding it. The first fortnight is the tuning period; after that, if it is still wrong, you stop paying. It is month to month.

Does this replace a person?

No, and you should be suspicious of anyone who says it does. It replaces the twenty minutes somebody spends scanning an inbox for the one thing that cannot wait, and it does that scan every few minutes instead of twice a day. The judgement, the reply and the customer relationship stay with your people.