Automation that fits the work, not the other way.
Off-the-shelf tools handle the common cases. Custom automations handle yours. We build workflows that connect the tools you already use, eliminate the copy-paste tasks consuming hours every week, and ship with documentation so they never become another black box.
Six build categories.
One honest sequence.
Every custom automation we deliver starts the same way — by mapping the workflow that exists today, not the one in a process diagram. Once we can describe the manual sequence end-to-end on a single page, we know what to automate, what to leave alone, and where a human still needs to sit in the loop.
Each build below is a real category of work we ship. Most engagements combine two or three of them rather than a single tool in isolation.
Low-code workflow builds
Zapier · Make · n8nFor most small and mid-sized businesses, the right answer is a low-code platform — Zapier for breadth of integrations, Make for branching logic and visual flow inspection, n8n where you want self-hosted control and no per-task billing. We design the flow, configure the modules, handle the authentication, build the error-handling paths, and make sure the trigger conditions match how your team actually behaves rather than the textbook scenario. Low-code is rarely the cheapest option per task, but it is the easiest for your team to maintain after we leave — and that is usually the deciding factor.
Code-led automations
Python · Node.js · BashWhen the workflow needs custom data transformation, runs at a volume that would be punishing on per-task pricing, or has to live behind your firewall, code is the right tool. We build automations in Python or Node where the logic is non-trivial, deploy them on infrastructure you already trust — a small VPS, a serverless platform, or your existing application server — and version-control everything in a Git repository you own. The deliverable is a maintainable codebase with comments aimed at the next developer to touch it, not at us.
API integrations & webhook pipelines
System wiringMost modern business software exposes an API; very little of it talks to other software out of the box. We build the connective tissue — webhooks that fire when a new lead arrives, scheduled jobs that pull invoices into accounting, two-way syncs between a CRM and a project tool, queues that batch and retry when an upstream service is rate-limited. The work includes authentication handling, schema mapping, idempotency so a retried event never charges twice, and observability so a silent failure does not stay silent.
Document & report generation
Output automationAutomated production of contracts, quotes, invoices, weekly reports, monthly client decks, and the dozens of templated documents that quietly take a person an afternoon every week. We build pipelines that pull structured data from your CRM, spreadsheet, or database, apply your existing template, and deliver a finished file — DOCX, PDF, or directly into your e-sign tool — without anyone copying numbers across windows. Where AI helps with summarisation or narrative paragraphs, we wire it in. Where it would invent facts, we leave the writing to a person.
Notification & routing workflows
Right person, right timeA surprising amount of business friction is just slow handoffs — a lead landing in an inbox nobody checks until tomorrow, an urgent invoice waiting on an approver who is on holiday, a support ticket that should have gone to a specialist sitting in a generic queue. We build routing logic that reads the incoming signal, applies your business rules, and notifies the right person on the channel they actually monitor — Slack, SMS, email, Teams — with the right context attached. Quiet hours, escalation paths, and out-of-office handling are all part of the build.
Documentation & handover
Always includedEvery custom automation we ship is accompanied by a written handover document — a plain-English description of what the automation does, where it runs, which credentials it depends on, what to do when something breaks, and how to extend it. The document lives in your knowledge base, not ours. The credentials live in your password manager, not ours. We deliberately design every engagement so a future developer or admin can take it over without needing to call us. A maintenance retainer is available if you would prefer that we keep watch — but it is an option, not a tether.
Built around the work,
not the tool.
The automation industry has a way of inverting the question — starting with the platform, the model, the impressive demo, and then hunting for somewhere it could be useful. We start at the other end. The workflow is the constant; the tool is the variable.
Map the workflow first
Before any tool is selected or any code is written, we sit with the people doing the work and document the existing process. What triggers the task. What inputs they look at. What decisions they make. What they hand to the next person. Most automations fail because somebody automated a workflow they never properly understood — including the small judgement calls and exceptions that keep it functional. We refuse to skip this step, even when the engagement is short.
Pick the simplest stack that fits
Sometimes the right answer is a single Zapier flow your office manager could maintain. Sometimes it is a Make scenario with branching logic and error retries. Sometimes it is a Python script on a small VPS, scheduled by cron. We do not have a default platform — we have a default question: what is the simplest stack that solves this without locking you in or charging you per task on a workflow that runs ten thousand times a day? The answer changes per engagement.
Maintainable by your team
An automation only your consultant can fix is a liability dressed up as an asset. Our builds favour mainstream platforms, plain-language naming, comments where logic is non-obvious, and a written handover document for every engagement. When custom code is necessary, it lives in a Git repository you own, with a README that points to the runtime, the credentials, the failure modes, and the next developer’s first move. Independence is the goal; ongoing dependency is the failure mode.
Observable in production
Every automation we ship has a way to know whether it is still working — a Slack channel that gets a daily heartbeat, a status dashboard, error alerts that go to a real person rather than into a void. Automations that fail silently are worse than no automation at all, because the team trusts them and stops checking. We treat observability as a build requirement, not a nice-to-have, and we test the failure paths the same way we test the happy path.
Workflows where the
copy-paste is obvious.
Custom automations are the right next step when you can describe the manual workflow in a sentence, the volume is high enough to feel painful, and the rules are stable enough that a machine can follow them most of the time. They are not the right next step when the workflow is genuinely judgement-driven or changes weekly.
A few recurring patterns where a custom build pays back inside the first quarter.
- i Lead-to-CRM workflows that drop informationForm submissions, inbound emails, and chat enquiries that someone manually pastes into your CRM, often inconsistently and often late. A clean automation here is usually the highest-impact build a sales team can commission.
- ii Invoice, contract, and quote generationAnywhere a deal closes and the next step is “open the template, fill in the fields, save as a new file, send for signature.” This category alone often pays for the entire engagement in saved hours per month.
- iii Cross-tool sync that two products refuse to do nativelyCRM to project tool. Project tool to accounting. Calendar to a custom database. Anywhere a person’s job is “keep these two systems in agreement” — that person should be doing other work.
- iv Reporting roll-ups assembled by handWeekly client decks, monthly performance reports, internal dashboards stitched together from four different sources every Monday morning. Automation does not have to draft the narrative — it just has to deliver the numbers, charts, and tables to where the writer starts.
- v Notification routing that is failing quietlyLeads going to inboxes nobody monitors. Urgent issues sitting in generic queues. Approvals stalled on holiday. Routing logic with the right channel, the right escalation, and the right context attached usually changes response times by an order of magnitude.
Where this does not fit: workflows that change weekly, processes that genuinely require a person’s judgement at every step, and one-off tasks that will not be repeated enough to earn back the build time. The first deliverable of every engagement is an honest assessment of which of your candidate workflows belong in each bucket — automate, redesign, or leave alone.
Curious how Google sees your site?
Send us your URL. We’ll send back a Premium SEO Report, prepared by hand, within 48 hours — domain authority, keyword rankings, backlinks, competitor gap, and the quick wins worth chasing first.
No sales call required.
The best automation is the one your team forgets is there. It respects the workflow that exists, removes the friction that doesn’t need to, and asks for nothing back.— The Aureole Practice —
Questions we get
about custom builds.
If a question is missing here, the contact link at the foot of the page goes straight to the person who would answer it. No ticket queues, no funnels.
i When do you build on Zapier or Make versus writing custom code?
ii What does a typical engagement timeline look like?
iii Do we own the automation, or are we locked into your account?
iv What happens when something breaks six months later?
Where custom automations
fit in the whole.
A custom build sits in the middle of the workflow practice — informed by the audit upstream, supported by tool selection, and only successful when the team training that follows is taken seriously. The link below returns to the parent service; the pills extend laterally to the sister sub-disciplines that compound with this work.
Parent service
Sister sub-disciplines
Adjacent services
Ready to ship
the first build?
Tell us about the workflow that wastes the most hours each week. We’ll respond within one business day with an honest read on the right stack, the rough timeline, and whether a custom build is the right answer or a smaller move would do.