AI automation
Automate the copy-paste.Keep the decisions.
For teams that move data from one app to another every day: we connect the systems you already use and automate the repetitive steps between them, from invoices and quotes to notifications and reports. AI steps in where information needs to be interpreted, and decisions stay with people.
Your team is doing the work of a system that doesn't exist yet.
In many companies, apps don't talk to each other, so people become the “connector”: they move data from one place to another, chase approvals by email and rebuild the same reports every month.
Data entered twice
The order comes into the online store, then gets copied into the invoicing software, then into the delivery spreadsheet. Every manual transfer is a chance for error.
Approvals lost in the inbox
A quote, an expense or a contract is waiting for someone's sign-off, but nobody knows exactly who has it or for how long.
Reports rebuilt by hand
At month end, someone exports data from three apps, combines it in Excel and writes the same kind of commentary as last month.
Documents filled in from scratch
Contracts, quotes and onboarding forms are filled in by hand using the same information that already exists in the CRM or in a form.
Automations built in a hurry
Some workflows were put together quickly, stop without warning and have no documentation. They are worth reviewing and describing clearly so they can be maintained.
Classic automation plus AI, each where it belongs
AI automations are workflows that run on their own between your apps, following rules we agree together, with artificial intelligence steps where information needs to be interpreted. Unlike an AI agent, an automated workflow always follows the same path, which makes it predictable and easy to check. Not every step needs AI: clear rules (if the payment is confirmed, issue the invoice) are automated simply and predictably. AI comes in where information arrives unstructured: an email that needs to be understood, a document that data must be extracted from, a text that needs summarising. In between, we put human checkpoints.
Process map
We draw the current workflow step by step: who does what, in which app and how long it takes. Often the map alone reveals steps that can be removed.
Connectors between apps
We link systems through APIs or automation platforms such as n8n, Make or Zapier, so data flows without manual intervention.
AI steps for unstructured data
Classifying emails, extracting data from PDFs or drafting a first version of a text, with the result handed to a person for review.
Human approvals and checks
The workflow can pause automatically for approval: a notification in Slack or by email, a confirmation button, then the process continues.
Monitoring and alerts
Every run is logged. If a step fails, the person responsible gets an alert and no data is lost.
What automation can bring
- 01
Time won back from repetitive tasks
Hours spent copying and checking data can be redirected towards clients, analysis or growth.
- 02
Fewer transcription errors
When data flows directly between systems, there are fewer chances to get a figure, a code or an address wrong.
- 03
Processes visible end to end
You can see what stage a quote, an approval or an onboarding is at, without asking anyone.
- 04
Faster response
Clients can receive confirmations, documents or next steps right after they act, not when someone finds the time.
- 05
Growth without multiplying manual work
A higher volume of orders or requests can be handled without each new client adding the same amount of admin work.
The real effect depends on how stable and documented the process is, the quality of the data and the apps involved. We estimate it during the analysis stage, without guaranteeing it.
Common scenarios
The examples are conceptual and illustrate possible types of workflow. They do not describe implementations delivered for specific clients.
- Online store
From order to invoice and delivery
For every paid order in WooCommerce or Shopify, the invoice is issued in SmartBill, the AWB (shipping waybill) is generated and the customer receives an email with the details. Orders with mismatches go to a colleague for review.
Conceptual example - B2B services agency
Quotes generated from the CRM
When an opportunity reaches a certain stage in HubSpot or Pipedrive, a quote is prepared automatically on the company template, with the client's details. The manager approves it before it is sent.
Conceptual example - Distribution company
Processing supplier invoices
Invoices received by email are read with AI, the data is extracted and compared with the purchase order, and any differences are flagged to the accountant before the invoice is booked.
Conceptual example - Growing company
Onboarding new employees
When the contract is signed, the necessary accounts are created, welcome documents are sent and first-week tasks are scheduled, with confirmations from each person responsible.
Conceptual example
What we can automate
We usually start with a single workflow with visible impact, then expand based on what works.
Workflows and apps
- Workflow automation
- Automations between apps
- Data synchronisation between systems
- CRM/ERP integration with your other apps
Documents and reports
- Document automation: generating, filling in, archiving
- AI data extraction from PDFs and emails
- Assisted report generation, with summaries
Sales and administrative processes
- Sales process automation
- Quote automation
- Administrative process automation
- Client or employee onboarding
Control and review
- Multi-level approval systems
- Human review steps before sensitive actions
- Error alerts and a log of every run
Indicative prices
Packages and prices
Price benchmarks for this service, in euro, excluding VAT. The final offer is set after a conversation about the project.
See all pricesAI assistant on your website
+ from €90/month operation · 2–4 weeks
from€1,500
- Answers from your documents and pages
- Takes requests and passes them to the team
- Installed on the website, WhatsApp or Messenger
- Monthly monitoring and improvement
- Recommended
Automation
n8n or Make · 1–3 weeks
from€900per workflow
- Analysis of the process and the data
- Connecting the applications you already use
- Parallel testing before going live
- Documentation and training
- Optional operation, from €150/month
AI voice agent
+ per-minute costs · 4–6 weeks
from€2,900
- Handles calls, bookings and frequent questions
- Transfers to a person when the situation requires it
- Integration with your calendar or CRM
- Disclosure and consent in line with the law
AI provider costs (OpenAI, Anthropic, telephony) are billed on consumption, transparently. The online AI assessment is free.
From manual process to automated workflow
Step 01
Process analysis
We talk to the people who actually do the work and identify the repetitive steps, their volume and the points where errors happen most often.
Step 02
Prioritisation
We choose the workflows with the best ratio of effort to impact and agree what stays manual, at least at first.
Step 03
Workflow design
We define the triggers, the rules, the AI steps, the approval points and what happens when something goes wrong.
Step 04
Build and testing
We implement the workflow in a test environment, with anonymised real data or simulated cases, and check it together with your team.
Step 05
Parallel launch
For a while, the automated workflow runs alongside the manual one, so we can compare results and build confidence before switching over completely.
Step 06
Documentation and maintenance
We leave a clear description of each workflow and agree who monitors it, so the automation doesn't become a “black box”.
The apps we can connect
We work with established automation platforms or with integrations written directly against the API, depending on complexity, volume and how much control over the data you need.
Whether an integration is available depends on each app's API and licence plan, and is checked at the start of every project.
- n8n
- Make
- Zapier
- Google Workspace
- Microsoft 365
- Slack
- HubSpot
- Pipedrive
- Salesforce
- SmartBill
- WooCommerce
- Shopify
- Airtable
- SQL databases
Security & control
Automation with brakes
An automated workflow repeats the same action many times, including a mistake. That's why every automation has limits, checks and a clear way to stop it.
Personal data and external providers
We identify which personal data passes through the workflow and on what GDPR legal basis. We send external AI services only the information needed for that step, under the provider's data processing terms.
Dedicated accounts and permissions
Automations use separate technical accounts, with access only to the data and actions they need, and easy to revoke.
Human approval for sensitive steps
Payments, messages to clients or bulk changes can be set up to wait for confirmation from a responsible person.
Logs, alerts and retries
Every run is logged. Errors trigger alerts, and failed steps can be retried without duplicating operations already completed.
Uncertain AI results
When an AI step cannot extract or classify information with enough confidence, the case is sent for manual review rather than pushed further along.
Useful guides
Before you decide, read
- AutomationHow to choose the first process to automate in your companyA practical method for choosing your first process to automate: an inventory of tasks, scoring on clear criteria, a small measurable case, human validation.Read the guide
- Automationn8n vs Make vs Zapier: which automation platform should your business choose?n8n vs Make vs Zapier, compared in terms a manager can follow: hosting, cost model, where data goes, complexity, AI steps and error handling. Plus who should choose which.Read the guide
- Automatione-Factura automation: what you can automate in RO e-Factura and what you check by handRO e-Factura can be integrated with your invoicing software, accounting and internal workflows. What can be safely automated, what stays checked by a person and what to ask before you start.Read the guide
- AutomationWhat is n8n? The automation platform you can host yourselfWhat n8n is, explained for managers: how workflows run, what the fair-code licence means, when self-hosting is worth it, what it can do with AI and where its limits are.Read the profile
Frequently asked questions
Which processes are worth automating first?
The ones that repeat often, follow fairly clear rules and involve moving data between apps. A good starting point is a task someone does every day and finds tedious.
What does AI process automation mean?
It means the repetitive steps in a process, such as copying data between apps, issuing documents or sending notifications, run automatically according to clear rules. AI is used only for steps where information arrives unstructured, for example an email or a PDF, and the result can be checked by a person.
What is the difference between classic automation and AI automation?
Classic automation follows fixed rules and is highly predictable. AI is added where information is not structured, for example to understand an email or extract data from a document. We combine the two, and the AI steps have extra checks.
When should I choose an automation and when an AI agent?
Automation fits when the steps are known in advance: the same trigger, the same rules, the same outcome. An AI agent fits when the request is phrased freely and the next step has to be chosen from a set of approved actions. Many projects start with an automation and later add an agent for the part that needs interpretation.
What happens if the AI doesn't know the answer or gets it wrong?
AI steps are configured to flag uncertain cases and send them for human review. In addition, sensitive actions can be blocked until approved, and the log makes it possible to find and correct errors quickly.
Do we have to change the apps we use?
Usually not. The aim is to connect your existing systems. We only discuss alternatives if an app does not allow integration at all.
How is data protected?
We use technical accounts with minimal permissions, send external services only the data they need and keep a log of every run. Together we agree the legal basis for processing personal data and which providers are acceptable for your company.
How do I measure how well the automations work?
Before implementation, we record the current time and volume of the task, then track the number of runs, the error rate and the cases sent for review. The comparison shows concretely what has changed.
What do the cost and duration of an automation project depend on?
On the number of apps involved, API availability, how complex the rules are, whether there are AI steps and the level of control required. A simple workflow between two apps is much faster than a CRM/ERP integration with several approvals. After analysing the process, we propose an estimate.
Can’t find the answer? Write to us and we’ll reply directly.
Request a process automation review.
Briefly describe a process that takes up your team's time. We'll tell you what can be automated, what should stay manual and which apps we'd work with.
Related services
- Custom software & integrationsApplications, platforms and integrations built around your processes, so data flows between systems on its own.
- AI agentsAI assistants that follow the steps you set, use only approved tools and complete concrete tasks, with a person making the decisions.
- Finance AI & business intelligenceInvoice data extracted automatically, reports that update themselves and clear figures for decisions.