Service
Automation and integrations
I connect the company's systems so work passes from one to the next without anyone typing: the sale on the website becomes an order in the ERP, and billing and messages go out on their own. The integrations do not duplicate (idempotency, queues and reprocessing) and warn you when a flow stops. The pilot usually goes live in 2 to 4 weeks.
Who it is for
Companies where someone copies data from one system to another every day, and the errors and delays already cost sales. It fits operations that grew with several disconnected tools, and anyone with an automation that duplicates, forgets or stops without anyone noticing.
Problems it solves
- A sale made on the website that somebody retypes into the ERP, with delays and typos.
- An integration that creates the same order or sends the same message twice because the notice arrived repeated.
- A flow that stopped days ago and was only discovered when a customer complained.
- Reprocessing data after a failure and fearing to charge or notify the customer twice.
- Billing, invoices and messages that depend on one person remembering to do them.
What I deliver
- Integration between your systems (website, ERP, CRM, billing, WhatsApp, spreadsheets), by API or webhook.
- Protection against duplicates: the same event, even when received twice, is processed only once.
- Queues and retries for what fails, with safe reprocessing after an outage.
- An alert when a flow stops or fails, with a record of what happened.
- A record of each run, to check what came in, what went out and when.
- Documentation of the flow in plain language, so your team can understand and maintain it.
How it works
Diagnosis
One conversation to map what gets copied by hand today, how much time it takes and what happens when it goes wrong. No cost, no commitment.
Design
I define what triggers the flow, what counts as a repeat and what happens on each failure, before writing any code.
Pilot in production
In 2 to 4 weeks, a real flow running, with the execution log and alerts turned on.
Scale
With the pilot measured, I bring the same care to the next flows.
Typical timeline
Pilot in production 2 to 4 weeks after the diagnosis, on one real flow. The timeline changes with the number of systems and with how soon the accesses are released.
Frequently asked questions
How do you stop an integration from duplicating an order or a message?
Each event received carries a key that identifies it, and the system records which keys were already processed. If the same notice arrives again, it is recognized and ignored. That is idempotency: repeating the operation gives the same result as doing it once. Platforms such as Meta deliver the same event more than once, so this protection is not optional.
What if the flow goes down halfway?
What failed goes to a queue and is tried again, and the system records where it stopped. After an outage, reprocessing is safe, because only what is safe to repeat is repeated and the rest goes through a check. When the flow stops for good, you get an alert instead of finding out from a customer complaint.
Do you use n8n or code?
Both, depending on the case. For simple flows that your team will edit, n8n is usually enough. For high volume, complex rules or a critical integration, I write code (TypeScript or Python), with tests. I explain the choice in the proposal.
How much does it cost and how long does it take?
The pilot usually goes live in 2 to 4 weeks. The price depends on the systems involved and the volume, and I quote it after the diagnosis, which is free. Third-party costs (platforms, APIs) get a monthly estimate before we start.
Let's talk?
The diagnosis is a conversation, free and with no commitment. You leave with the list of processes that eat the most hours of your team.
Talk on WhatsAppOther services
Want this running in your operation?
Tell me in one sentence. I reply with what can be done, how long it takes and what it costs.
Tell me the problem