Recording a sick leave notification seems simple: name, date, reason, expected duration, done in the system. That's precisely why many companies wonder whether AI couldn't simply take this over. The answer is partly, and it's useful to know exactly why that part can and the other part cannot.
The task as defined here -- recording a notification with date, reason and expected duration, up to and including the registration -- is fairly structured. There's a fixed format: a date, a text field for the reason, an estimate of the duration. It therefore scores high on structuredness. The volume is often substantial: any company with more than a handful of employees has ongoing sick leave notifications, and those add up to a fair number of hours of registration work per year. Physically, too, there's nothing to it: this is purely text work, not an action someone needs to perform on site. And customer contact doesn't come into play -- the notification comes from an employee or manager, not from an external customer.
Against that, compliance tips the balance, and not in favor of automation. A sick leave notification is medical information, or directly touches on it. The reason for absence often falls under special categories of personal data, with its own rules on who may see it, how long it's retained and who has access to it. That's not a matter of 'AI can process text, so this can be automated' -- it's a matter of whether the link with the absence protocol and the privacy safeguards are in order. Without that being secured, registration by AI is not responsible, however simple the task looks on paper.
Discretion therefore also scores low: there's little to weigh in the recording itself, but the sensitivity of what is being recorded calls for careful setup beforehand. And the cost of errors is not negligible: an incorrectly registered date or reason can have consequences for building absence files, for reintegration, and ultimately for decisions an employee is entitled to.
What AI can handle today is the text: converting a notification that comes in via a message, email or form into the right fields in the absence system, filtering out the date and expected duration and filing it neatly. That saves an HR employee or manager from manually typing out every notification. This concerns the typing, not assessing the notification, contacting the employee about the reason, or estimating what the notification means for staffing. That last part remains human work, and that is also precisely where the task in practice often stops: at a registered notification, not at the follow-up steps.
The outcome depends heavily on how a company has set up its absence process. An organization with an absence system that is already linked to the absence protocol and where the privacy rules around medical information are technically secured, can have the typing of notifications proceed considerably faster than a company that still works with loose forms and emails. Volume also makes a difference: at a company with hundreds of employees and a lot of short-term absence, automating the input work pays off sooner than at a company with ten employees, where a manager can type in the notification themselves just as quickly.
This is one of many places in a company where the same shift becomes visible: AI takes over the input work, people retain responsibility for what is sensitive or uncertain. You see that same dividing line in processing leave requests, where the request itself is easy to process but the judgment on exceptions is not. Also in preparing payroll processing, the emphasis lies on structured input, with human control over the outcome. And in drawing up schedules, a similar division applies: the calculation work can be automated, the exceptions and an employee's context cannot.
Why this already works this way at one company and not yet at another depends not only on the task but also on the setup of systems and processes around it -- this is elaborated further on the page about why the same work turns out differently per company. For HR tasks in a broader sense, and how they relate to, for example, IT work, the starting point can be found on the page about work AI can take over in the IT department, where a similar line of reasoning is applied to a different domain.
This page makes no statement about personnel decisions that might follow from a lower workload in absence registration: separate statutory requirements apply to that, independent of what is established here about the task itself.
If you want to know how this applies to your own absence process, the free quickscan is a first step: twelve questions, no account required, with an indication of what portion of the hours in this profile can be taken over by AI today. The full work scan, which maps your company's work task by task including FTE calculation, is still under construction and not currently available.
Vraag maar. Ik ken de kennisbank van deze site; wat ik niet weet, zeg ik erbij.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.