How we work
Most automation projects fail because they automate the wrong thing. This page explains, with no filters, what happens in each phase, what you get, and where this model fits.
They fail because they start with the tool instead of the process. They fail because a broken process gets automated — and an automated broken process just produces errors faster. They fail because nobody measured what the task cost before, so nobody can prove the savings after.
And they fail, almost always, because the vendor bills for configuration hours instead of defined results.
Amud splits every project into phases — Audit, Design, Implementation and Continuous optimisation — for one very specific reason: to measure the cost of repetitive work before investing a single euro in eliminating it.
30-45 minute session. Free. No commitment.
Its only goal is to understand your operations and find where the hours go. It's not a sales call — it's a diagnostic session.
You don't need process documentation, diagrams or technical knowledge. Just tell us how your team works on a normal day. Mapping the processes is our job — that's exactly what this phase is for.
You talk to the person who will design and implement your automations, not a salesperson. Whoever diagnoses is whoever builds. Zero intermediaries between what you tell us and what gets delivered.
Definition. 1-2 weeks. Price agreed after the audit.
Here we go from "this could be automated" to "this is exactly what will be automated, like this, and it costs this". It's the phase that saves the most money.
The most profitable automation is almost never the flashiest. We rank every workflow by hours saved versus implementation effort, and phase 1 only keeps what pays for itself within months. The rest is phased later or dropped.
Billing for configuration hours rewards slowness. We charge per automation delivered and working, at a fixed price. If it takes us longer than expected, we absorb it. If you change the scope, we re-quote in writing before touching anything.
Sometimes, digging deeper, we find a process shouldn't be automated yet — an API that doesn't exist, volume that's too low, a process that's too unstable. When that happens, we tell you before starting and we don't charge for that workflow's design. You keep the analysis and a clear next step.
Build. 2-6 weeks per phase. Optional continuous optimisation afterwards.
Here we build what was defined in the Design, without disrupting your operations. And once it's live, we keep it measured.
Automations are built in accounts you own — your n8n, your Make, your APIs, your credentials. If we stop working together tomorrow, everything keeps running and any professional can maintain it. No black boxes, no vendor lock-in.
Your team learns how to use each automation, read its alerts and resolve basic incidents. Recorded training session and a practical written guide. The goal is for the automation to be yours, not ours.
Optional and monthly: we monitor every workflow, resolve incidents before you notice them and propose new automations as your business changes. Every month you get a report with hours saved and what we propose to improve. If you'd rather manage it in-house, there's no lock-in.
Industry standard
Audit → Design → Implementation → Optimisation
The operational difference is a 45-minute audit at the start. The financial difference is automating what pays for itself in months — and being able to prove it.
How it happens
The process gets automated exactly as it is, without questioning it. The result: the same errors as always, just produced faster and with nobody watching.
How we prevent it
The Audit reviews the process before automating it. If the process needs fixing first, it gets fixed — sometimes that improvement alone saves more than the automation itself.
How it happens
A workflow is implemented without involving the people who did the task. They don't trust the result, double-check everything by hand and the savings evaporate.
How we prevent it
The Design is done with the person who knows the process, and the Implementation includes validation and training. The automation arrives as help, not as an imposition.
How it happens
Everything is built in the vendor's accounts, with no documentation. When the relationship ends, the company can't touch anything and it all has to be rebuilt from scratch.
How we prevent it
Everything is built in your accounts from day one, documented and with your team trained. If we're gone tomorrow, everything keeps running.
It fits if
It doesn't fit if
If you're not sure which group you're in, the audit is exactly for finding out. 30-45 minutes, free, no commitment.
The plan is phased by design: each phase groups automations that stand on their own. You can implement only phase 1 — the highest-return one —, measure the actual savings and decide later whether to continue.
Your data is processed in your own accounts and tools. When AI is involved, we use paid APIs with no-training clauses — your data is not used to train models. We sign an NDA and a data processing agreement if you need one.
Yes, and it's designed for that: everything stays in your accounts, documented, with recorded training and a written guide. Continuous optimisation is optional, not an obligation.
Every workflow includes automatic alerts and a contingency plan defined in the Design: what gets done manually in the meantime and who acts. With continuous optimisation, we detect and resolve incidents ourselves, usually before you notice.
It depends on the case: some automations run on what you already pay for, others need an orchestration tool (like n8n or Make) or AI credits. Monthly licence costs always appear in the plan, itemised and with no surprises.
No problem. Send us yours before the session or we use a template of ours, whichever you prefer.
Almost no process gets automated 100%, and it doesn't need to be: we automate the repetitive 70-80% and leave the judgement calls to people. That split is defined with you during the Design.