What I automate
E-mails and enquiries, copying data between systems, client reminders, reports, extracting data from documents and answers to frequent questions on the website.
- E-mails and enquiries. Sorting by type, logging to a CRM with a summary and a draft reply that a person checks.
- Website enquiries. A form or configurator stores all the answers and sends a complete brief, not a single sentence.
- Reminders. For example the appointment reminders sent by the booking system for the Gynkomed practice.
- Documents. Extracting data from invoices and receipts and writing it to a spreadsheet.
- Reports. A regular overview of enquiries or sales by e-mail.
- Answers to frequent questions. A chat with FAQs is part of my SaaS Marxyo for medical practices.
How it looks in my own work
I use AI automation every day in my own work: for enquiries, e-mails and processing lists of companies.
An enquiry from my website configurator arrives as a clear brief with all the answers, so I do not have to send follow-up questions.
Incoming e-mails are in my own CRM and AI prepares a draft reply in my style. I read the draft, edit it and send it myself.
A script cleans my company lists of duplicates and verifies the e-mail addresses. Work that would take hours by hand runs in the background.
Which tools I use
For simple integrations there are platforms such as Make, Zapier or n8n. When more control is needed, I write my own scripts in Python or Node.js that call language models, for example Claude from Anthropic or open source models through Groq.
- Make, Zapier, n8n. Ready-made connectors for e-mail, spreadsheets and CRM. Suitable for simple flows with a few steps.
- Language models. Claude, models through Groq and similar services read, sort and summarise text and prepare draft replies.
- Own scripts and database. Python, Node.js and PostgreSQL for large volumes, custom rules or when you do not want to pay for every step on a platform.
- Web application. When automation needs its own interface and overviews, I build a custom web application.
How to start
We list the repetitive tasks, pick the one that saves the most time, run the automation alongside manual work first and only then add the next one.
- List of routine tasksFor a week you note down what you do repeatedly and how long it takes.
- First processWe pick a frequent task with a clear outcome. You get a solution design within 2 working days.
- Parallel runThe automation runs alongside manual work until we confirm it does what it should.
- Next stepOnce the first process works, we add another one.
What to watch out for
Do not automate chaos. Let a person approve replies to clients at first, handle personal data carefully and set up an alert for when the automation fails.
More examples by industry in the article AI automation for small businesses.
Frequently asked questions
Do I need a developer for AI automation?
Simple integrations can be built in Make, Zapier or n8n without programming. For more complex processes and larger volumes a custom solution pays off.
Which processes should I automate first?
The ones that repeat every day and have clear rules: sorting e-mails and enquiries, copying data and reminders.
How much does AI automation cost?
We start with one process, so the investment is small. You get the solution design and a price estimate for free.
Will AI replace employees?
In a small business it rather takes routine work off people. Decisions and communication with customers stay with a person.
Is my data safe?
I send to an AI service only the data the task needs. I do not send sensitive data, such as health data, to external AI services.
I will send you a design within 2 working days
Send me a few sentences about what you need. The design is free and with no obligation.
Contact me