n8n automation
Connect the systems your business actually runs on.
I use n8n to build the workflows between CRM, email, documents, spreadsheets, APIs and internal systems — especially when standard integrations stop being flexible enough.
When to use n8n
When a simple integration stops being enough.
Not every automation needs n8n. If a native integration between two tools already does what the business needs, it should stay simple. n8n is one of several approaches I use across business process automation, CRM automation, marketing automation and AI workflows.
It becomes useful when a process combines several of the things a point-to-point integration was not built for: more systems, more logic and more ways for the data to be incomplete or wrong.
Signs a process may need n8n
- Multiple systems in one process
- APIs and webhooks
- Custom business logic
- Conditional routing
- Data transformations
- Validation before data moves on
- Email or document processing
- Retries and error handling
- AI-assisted steps
- Writes to a CRM
- Human checkpoints
Choose the architecture based on the process, not the tool.
Connected systems
One workflow can cross the tools your teams use every day.
n8n acts as the orchestration layer: it receives a trigger, reads from one system, decides what should happen and writes to the next. These are the kinds of systems my workflows connect.
CRM
- Salesforce
- HubSpot
- Connexys on Salesforce
- Find existing records
- Create or update records
- Route leads
- Create job applications
- Enrich data
- Validate before writing
Email
- Outlook
- Shared mailboxes
- Monitor incoming messages
- Recognise relevant emails
- Process replies
- Extract content
- Prepare responses
- Keep the conversation context
Documents and data
- CVs
- PDFs
- Word documents
- Google Sheets
- Data tables
- Extract information
- Normalise it
- Validate it
- Pass structured data to another system
APIs and webhooks
When a standard integration does not cover the process, n8n can call a system’s API directly or react to a webhook, so the workflow is not limited to the connectors a tool happens to offer.
AI
- Gemini through Vertex AI
- Extract or analyse unstructured information
- Classify messages
- Assess information against defined criteria
- Prepare a response
In production
From candidate email to structured Salesforce data.
For Logi-technic I built Flo, an n8n workflow around a shared recruitment mailbox. Staffing agencies send candidate proposals by email; Flo recognises the relevant ones, processes the email and supported CV or document attachments, and structures contact details, work experience, education and qualifications.
Logi-technic’s detailed technical role profiles are the input for the assessment. AI proposes a core profile and supports the assessment; predefined workflow logic controls how that output is used in the process.
What the workflow handles
- Follow-up is prepared from existing Dutch and French templates, as Outlook drafts rather than sent automatically.
- Replies from agencies are connected back to the original conversation.
- Information is validated before anything is written to Salesforce/Connexys.
- Existing candidate and application records can be reused where appropriate.
- Candidate and application details and CVs can be carried into Salesforce/Connexys.
- Error handling, validation and record checks reduce unnecessary or duplicate work.
Flo, step by step
- 01 · EmailMonitor the shared recruitment mailbox and recognise candidate proposals.
- 02 · ExtractRead the email and supported CV or document attachments.
- 03 · StructureContact details, experience, education and qualifications.
- 04 · AssessAI-supported assessment against the technical role profiles.
- 05 · PrepareOutlook drafts from Dutch or French templates.
- 06 · ValidateCheck the details and look for existing records.
- 07 · SalesforceCarry candidate, application and CV into Salesforce/Connexys.
AI + automation
Use AI where judgement helps. Keep the process controlled.
AI is good at specific tasks that are hard to express as fixed rules. I use it for those tasks inside a broader workflow, not as a replacement for the workflow.
Tasks AI can take on
- Extract information from unstructured text
- Classify messages
- Interpret documents
- Propose categories
- Assess information against defined criteria
- Prepare a response
What the workflow still controls
- When AI is called
- Which data it receives
- What happens with the output
- Validation
- Retries
- Writes to other systems
- Exception handling
The goal is reliable automation, with AI where it adds judgement.
Production workflows
Automation should be designed for the moment something goes wrong.
A production workflow is more than one node connected to the next. APIs time out, emails arrive incomplete and records already exist. How much protection a workflow needs depends on what it touches and what a failure would cost the business.
Depending on the workflow, that can include
- Input validation
- Retries
- Handling API failures
- Error logging
- Notifications
- Checkpoints
- Checks for duplicate or existing records
- Fallback logic
- Controlled writes
- Manual intervention where appropriate
n8n, Zapier, Make or a native integration?
Each has its place. The decision depends on the process.
- Native integration
- Best when the process is simple and the built-in connection already does what the business needs.
- Zapier or Make
- Can be a good fit for straightforward workflows where speed and simplicity matter more than deep custom logic.
- n8n
- Particularly useful when a workflow involves multiple systems, APIs, transformations, custom logic, branching, validation or more technical orchestration.
- Custom code
- Sometimes the better choice, when a workflow platform adds unnecessary abstraction or the requirement is better solved directly in software.
The right question is not “Which automation tool is best?” but “What is the simplest reliable architecture for this process?”
Questions about n8n
n8n, answered briefly
- What is n8n?
- n8n is a workflow automation platform used to connect systems, APIs and data flows. It is particularly useful when a process needs more logic or flexibility than a standard point-to-point integration provides.
- When should a business use n8n?
- n8n becomes useful when a workflow crosses multiple systems or needs custom logic, data transformations, validation, API calls, conditional routing or controlled error handling.
- What can n8n automate?
- Examples include processing incoming emails, extracting information from documents, synchronising data between systems, updating CRM records, routing leads or tasks, calling APIs and adding AI-assisted steps to an existing process.
- Can n8n connect HubSpot and Salesforce?
- Yes, n8n can work with both HubSpot and Salesforce. The exact architecture depends on what data needs to move, which system owns that data and what should trigger the workflow.
- Can n8n process incoming emails and documents?
- Yes. A workflow can monitor a mailbox, identify relevant messages, process supported documents, structure extracted information and pass validated data into another system. The Logi-technic workflow is a real example of this approach.
- Can n8n use AI inside a workflow?
- Yes. AI can handle defined tasks such as extracting, classifying, analysing or preparing content while n8n controls the surrounding workflow and system interactions.
- Is n8n suitable for GDPR-sensitive workflows?
- It can be, but GDPR compliance depends on the complete architecture rather than n8n alone. Hosting location, subprocessors, connected systems, the data being processed, retention and access controls all need to be considered.
- n8n vs Zapier or Make: which should I use?
- It depends on the process. Simpler workflows may be well served by native integrations, Zapier or Make. n8n becomes particularly useful when more technical orchestration, APIs, transformations or custom logic are required.
- Can you improve an existing n8n workflow instead of rebuilding it?
- Yes. Existing workflows can be analysed for fragile steps, unnecessary manual work, error handling, data quality and maintainability. Rebuilding only makes sense when the current architecture is the problem.
- Who builds n8n automations in Belgium?
- VNCK builds n8n automations from Willebroek, Belgium. Projects include workflows connecting email, documents, AI-assisted processing and CRM systems such as Salesforce. Projects can be delivered remotely, with on-site collaboration where useful.
Have a process in mind?
Start with the process, not the automation tool.
Tell me what your team is doing manually, which systems are involved and where the process gets stuck. I’ll help determine whether n8n — or a simpler approach — is the right fit.