← Services · 06 / 06

Automation & portals

Client portals, billing, documents and support that run without anyone chasing.

  • Retain
  • In production: Zesty, Soma, BidMorning
  • Audit trail on every step

Why it matters

Automation is only useful if you can trust it. When it touches money or stock, every step has to be recorded, repeatable and safe to retry.

In Zesty every change to an order, from a guest’s QR scan to a refund, goes through one place that checks the rules and records who did what. Money and stock update together or not at all, and messages go out from a queue that retries safely.

The same discipline runs Soma’s GST invoicing and BidMorning’s tender alerts: failures are loud and reviewable, never silently patched.

What we build

01Client portalsOne login for customers to see orders, invoices, documents and support tickets.
02Billing and invoicingGST-compliant invoices, credit notes and settlements generated from the record, not retyped.
03DocumentsAgreements, quotes and reports generated, signed and stored where people can find them.
04Alerts and digestsEmail and WhatsApp alerts on the events that matter, like a new tender or a failed payment.
05Support desksTickets with owners, status and history, connected to the customer’s record.
06Custom automationsThe repetitive steps between systems, run on schedule, retried safely and logged.

How we build automations

  1. 1Name every stepWrite down the transitions (bill, settle, refund, renew) and the rules for money and stock as checks we can test.
  2. 2One door for changesEvery change goes through one place that applies the rules, so the portal, staff and payment webhooks agree.
  3. 3Record and retry safelyEach action is logged with who and when. Messages and syncs go through a queue that retries without doubling up.
  4. 4Make failures visibleAlerts on stuck jobs, failed payments and backlogs, reviewed by a person instead of hidden.
  5. 5Launch and stayDeployed with a runbook, a tested backup restore and error tracking watched after release. The same team keeps it patched and running.

The stack we reach for

Postgres
One recordMoney, stock and audit entries saved together in one transaction.
Queues · workers
Background workEmails, alerts and syncs delivered by workers that retry safely and report what failed.
Dokploy · AWS
Running itDeployed with health checks, tested backup restores and alerts on what actually breaks.
Built forFinance teamsOperationsCustomersSupportOwners

Automation in production

See all work →

Questions

Something else on your mind? Ask it on the free 30-minute call.
01What can be automated first?

Start with the step people repeat most and dislike most: invoices, reminders, reconciliations or status updates. We automate that, measure it, then move on.

02Is it safe to automate billing?

Yes, when it is built properly. Totals come from the record, payments are confirmed by verified events, and every change is logged so it can be checked.

03Can customers see their own records?

Yes. A client portal gives each customer their orders, invoices, documents and tickets, and nothing belonging to anyone else.

04What happens when something fails?

It is retried safely, and if it still fails it lands in a list your team reviews. Nothing fails silently.

Stop chasing. Start running.

A free 30-minute call. Tell us which task you would automate first.
Pick a dayMon – Sat
Request Tue 20 Oct →