Menu
Nederlands

Process automation with AI

A lot of office work is moving information from one place to another and making small decisions along the way. I automate those steps, with AI for the parts that need reading or writing and a person for the parts that need judgement.

Tell me about your project

The problem

Think of the incoming email that has to be sorted, the report someone puts together by hand every week, or the request that passes through three people before anything happens. Each step is small, but together they take hours.

Classic automation tools break as soon as the input isn't perfectly structured. Language models can handle messy input, but they need a reliable process around them.

How I approach it

  1. Map the process

    We walk through it step by step and mark which steps are repetitive, which need judgement and where things go wrong.

  2. Automate the dull parts

    Scheduled jobs, queues and integrations move the data between systems, and retry when a system is down.

  3. AI for reading and writing

    Language models sort emails, pull out details and draft replies or summaries, always in a fixed format that the next step can check.

  4. People where it matters

    Approvals and exceptions go to the right person, with everything they need in one place.

  5. Visible and measurable

    Logs and simple dashboards show what ran, what failed and what's waiting, so the process doesn't turn into a black box.

What to watch out for

  • Automating a bad process only makes the mistakes faster. Simplify first.
  • Every integration can fail. Plan for retries, alerts and a manual fallback.
  • Not every decision needs a model. Fixed rules are cheaper, faster and easier to explain.

Projects like this

  • B2B supplier
  • 2026

Expert blog articles in many languages

The client wanted a steady flow of expert articles on its website, in many languages. A blog engine picks a topic, researches it on the web, writes the article with sources, checks it and translates it. It also creates the images and delivers the article to the CMS. An editor approves every article before it goes live.

  • After the first version, I worked on the engine's source research, its scale and the delivery to the CMS.
  • Every article has references readers can check. An article without sources doesn't get through.
  • Validation steps check the structure, brand rules and factual claims before delivery.
  • A delivery queue sends each article to the CMS and retries when the CMS returns an error.
  1. Topic profileWeb sources
  2. Write the article with sources
  3. Check facts and structure
  4. Translate and create images
  5. Article in the CMS, ready for approval
  • Financial services company
  • 2025

Automatic news checks for every client

Staff used to search the news for each client by hand. Now a scheduled job collects articles every night, matches them to the client list and classifies them. In the morning, the team sees only what matters, such as an insolvency.

  • Every night, the job pulls in articles from many news sources.
  • It flags insolvencies and other important events and filters out routine news.
  • The team now reads one short list each morning instead of searching for hours.
  1. News sources, each night
  2. Match to the client list
  3. Classify the event
  4. Morning list of important items
  • Investment company
  • 2025

A personal opening line in each newsletter

The client sends a newsletter to its business contacts, and a generic opening doesn't get much attention. Now every contact gets a short, personal opening line that refers to a recent LinkedIn post by the person or their company, or to the last email exchange.

  • The tool pulls the contact list and company data from the CRM.
  • It finds recent LinkedIn posts by each person and each company.
  • A language model writes one or two sentences per contact, in the client's tone of voice.
  1. CRM contact listLinkedIn postsLast email exchange
  2. Find a relevant topic
  3. Write the opening line
  4. Personal opening in each newsletter

Frequently asked questions

Do you use tools like Zapier, Make or n8n?

Where they fit, yes. For processes with more volume, complex logic or sensitive data, custom code is usually more reliable and cheaper to run.

Which systems can you connect?

Anything with an API, and many systems without one through exports, email or a database. Common examples are CRMs, accounting software, SharePoint, ticketing systems and webshops.

What happens when something goes wrong?

The process retries where it can, alerts someone when it can't, and never silently drops an item.