How do you use phased retirement for knowledge transfer?
Phased retirement looks like a long runway for a handover, but in the block model it rarely is: during the working phase the person is fully loaded, and once the release phase begins they are gone from one day to the next. The runway only becomes useful when the last year of the working phase is planned as a project with milestones, and experienced specialists are asked about exceptions and decision logic in short AI-guided interviews, not only in the final weeks.
- The deadline is the release phase. From then on the person is gone, even if still employed.
- The working phase is routine. Rare cases hardly come up there and have to be asked about deliberately.
- Start a year ahead. Short interviews spread over months instead of a manual in the final weeks.
“It is not an issue for us, our people go into phased retirement. That gives us two or three years.” This sentence comes up in almost every conversation about knowledge capture once the baby boomers are mentioned. It sounds reasonable. And it is the reason why, in many organisations, nobody is prepared at the end despite long lead times.
The German scheme behind it is called Altersteilzeit. Comparable phased-retirement arrangements exist in several European countries, and the pattern described here applies wherever the model splits into a working and a release phase.
Why the assumption is so plausible
Phased retirement is predictable. The date is fixed, often years ahead, and someone who has signed the contract will not suddenly resign. Compared with a departure on four weeks' notice, that feels like luxury. HR has the date in the system, managers know it, the person concerned certainly does. What could go wrong?
The answer lies in how the scheme is used in most companies. The block model splits the term into a working phase followed by a release phase. In the first half people work full time, in the second not at all. They remain formally employed, but they are no longer there.
What actually happens in the working phase
The working phase is not a handover phase. It is ordinary working time, and the person in question is usually the one who gets asked most. Precisely because they have been around for decades, the difficult cases land on their desk. Ask them to write a handover manual on top of that and it gets postponed, not out of unwillingness but because day-to-day business is always more urgent.
Then there is succession. In many cases the successor is hired only as the end of the working phase approaches, for budget reasons or because the role is hard to fill. The shared time then shrinks to a few months, sometimes weeks. Three years of runway have turned into a quarter of overlap.
The day the release phase begins
When the release phase starts, access ends abruptly. You can no longer just ask the colleague, and you should not, because they are released from work. Many are still happy to help when the phone rings. Building an organisation on that is not a plan, it is a hope.
We see one pattern again and again. The questions that come up in the first months after release hardly concern daily routine. They concern the rare cases: the line that only causes trouble in winter, the customer who renegotiates every two years, the check that is only required during an audit. Those cases never came up during a working phase that was mostly routine.
The scale of the problem
In Germany around 1.3 million people retire every year while only about 800,000 enter the labour market, leaving a shortfall of roughly 4.3 million workers by 2036 (IW Köln, 2026). Comparable curves apply across much of Western Europe. Part of these departures runs through phased retirement; how large that part is in your organisation, HR knows. How much experience hangs on those people, usually nobody knows exactly.
One qualification belongs here: where phased retirement runs as a part-time model, with reduced hours over the whole term, there really is a longer transition with genuine overlap. That model is less common in practice, but where it applies the problem shifts to whether the shared time is actually used.
What holds instead
The runway is a real advantage when it is treated as a project rather than as waiting time. Three things make the difference.
First, the date. The relevant deadline is not retirement age but the first day of the release phase. Counting back from that date, knowledge capture starts about a year earlier, not in the final weeks.
Second, the scope. Not an entire career gets captured, only the processes on which the person gets asked today and the rare cases only they know. An AI-guided interview needs around 30 minutes of expert time per conversation, spread over months so day-to-day business does not suffer. How to select the critical roles is covered in the piece on the retirement wave.
Third, the result. What gets captured is verified and stored so the successor finds it, whenever they start. In Omnora projects we see 15 to 30% shorter onboarding time and 20 to 40% fewer questions to the people who hold the knowledge. These are our own measurements from projects, not external research.
That leaves one question for the next HR round: who in your organisation starts a release phase in the next twelve months, and for whom would the department notice in the first winter?
How Omnora captures expertise before retirement, alongside daily business and in about four weeks per expert.
Retirement & demographic change →Häufige Fragen
Is the runway of phased retirement enough for a handover?
In the block model, usually not by itself. During the working phase the person is fully loaded, and when the release phase begins access ends abruptly. The runway only helps when the last year of the working phase is planned deliberately for knowledge capture.
When should knowledge capture start with phased retirement?
About a year before the release phase begins. What counts is not retirement age but the first day the person no longer works. Planning runs backwards from that date.
Which knowledge is most likely to be lost?
The rare cases: seasonal faults, negotiations every two years, checks that only come up in an audit. They seldom appear during a working phase full of routine and are therefore missing from every handover list.
How much time does capture cost the experienced specialist?
An AI-guided interview needs around 30 minutes of expert time. Spread over several months, that fits into the working phase without blocking daily business.
What changes with the part-time model?
There is genuine overlap with reduced hours over the whole term. The risk then lies less in the abrupt end and more in whether the shared time is actually used for knowledge transfer.
Quellen
- IW Köln (2026): Retirements, labour market entries and the German workforce gap to 2036 [Study]
- Omnora: around 30 minutes of expert time per AI interview; 15 to 30% shorter onboarding and 20 to 40% fewer questions to knowledge holders [Omnora field value]
