What do an IBM i specialist, a nuclear engineer and a Yorkshire stonemason have in common?
They work with very different systems and materials, but each gives us a way to think about expertise that takes time to develop. An organisation might recruit someone with the right qualifications and still need to help them understand the decisions, relationships and practical details that their predecessor learnt over many years.
What interests me is what happens after the new person arrives. Recruitment may fill the post while leaving important gaps in the knowledge available to the person taking it on. Who can explain why an unusual process exists, help them recognise an exception or introduce them to the colleague who knows how to resolve it?
That is why I find knowledge continuity a useful way to think about this. It includes the records people leave, how colleagues learn together and the support that remains available as the organisation changes.
In Fortra's 2026 IBM i survey, 69% selected IBM i skills among their five leading IT concerns. This measures respondents' priorities, rather than knowledge lost through departures. [1] The International Atomic Energy Agency has published guidance on managing knowledge-loss risks, while York Minster describes apprentices learning heritage craft alongside digital technology. [2, 3] Each offers something useful to consider, although the work and the consequences of losing knowledge differ considerably.
Why are things the way they are?
Imagine an IBM i specialist looking at an overnight processing job that appears to run in an unnecessarily complicated sequence. A new colleague can read the code and understand each step, but may still need someone to explain the customer exception or downstream finance process that led to that sequence in the first place.
The nuclear sector makes the importance of that history particularly clear. IAEA guidance on decommissioning calls for critical knowledge to be preserved while a facility is still operating, for use after shutdown. [4] Keeping the drawings and procedures matters, along with the context people need to interpret them.
In our IBM i example, recording the explanation would also give a successor something to question. An old exception may still protect an important dependency, or the circumstances that justified it may have changed. Keeping the reasoning available helps someone make that judgement without having to start from a guess.
When does the usual approach need a closer look?
Other knowledge helps someone recognise what deserves investigation. Historic England's repair guidance explains that a material's appearance needs to be considered alongside its physical properties and compatibility with the existing fabric. [5] For a stonemason learning this work, examining a repair with an experienced colleague could make the reasoning behind the choice visible as well as its finished appearance.
For our IBM i specialist, the equivalent might be a delayed overnight job. Imagine that the standard checks appear to permit a restart, while an unusual combination of activity and incomplete downstream output prompts the specialist to investigate further. Which cues changed their judgement, and how could a colleague learn to recognise them? Discussing that could lead to a clearer runbook, opportunities to work through past cases, or both.
Experience can be valuable without making every established habit correct. A successor needs room to question the explanation, understand the evidence behind it and learn when to pause and seek authorised help.
How does the human organisation work?
A diagnosis might depend on knowing who in finance can explain an unusual transaction. A repair may need another craft specialist involved before work begins. Sometimes the useful relationship is with a receptionist or security guard who knows how to reach the right person when the usual route has failed.
Those relationships are worth including in a handover, but knowing somebody's name does not give a new colleague the trust built through years of working together. An introduction is a start. Working through a real problem together can help the successor learn what each person knows, who has authority to decide and how to explain the issue to the people involved.
What the research tells us
It is tempting to call all of this tacit knowledge, but that can obscure useful distinctions. A customer exception may be straightforward to explain once somebody asks about it. Judging how a material responds while working with it may be much harder to put into words. A review in Knowledge Management Research & Practice challenged the use of the same label for knowledge that can and cannot be articulated. [6] This matters when deciding what a record can preserve and what a colleague will need opportunities to practise.
A study in the Strategic Management Journal examined 122 transfers of best practice within eight companies. Barriers included the recipient's ability to understand and use the knowledge, uncertainty about why a practice worked, and the relationship between the people involved. [7] Although it examined practice transfer within companies, it raises a relevant question for succession: what needs to happen before a colleague can use the information we have made available?
A systematic review in Work, Aging and Retirement examined knowledge retention from older and retiring workers, bringing together the characteristics of the knowledge, the people, their relationships and the organisational setting. [8] It helps explain why continuity depends on the circumstances of the work and the people involved. Someone may leave a thorough explanation of a difficult decision, while a colleague still needs time to work through similar decisions with feedback.
Build on the methods already in use
There is already thoughtful work on this in the IBM i community. Fortra's retirement guide recommends overlap so experienced staff can explain local business logic and mentor new colleagues. COMMON's N2i programme connects newcomers with mentors and helps them develop professional networks. [9, 10]
Automation also deserves a serious place in continuity planning. Fortra recommends introducing it while experienced staff can help encode familiar processes. [9] A working automated process can reduce dependence on a person carrying out its routine steps. Alongside that, someone needs to understand its assumptions, recognise when they have changed and take responsibility for a safe revision.
The question I would bring to an organisation is how its existing arrangements work together. Documentation, mentoring, automation and external support may cover much of what it needs. Looking at one important dependency could confirm that the provision is strong, or show where a successor needs more explanation, practice or access to other people.
A final handover can contribute, particularly when colleagues have time to work through questions together. I would be less comfortable relying on one conversation to reveal everything that matters. It depends on what the departing person remembers, what they think is worth explaining and what the other person knows to ask. IAEA guidance includes exit interviews alongside methods such as mentoring and cross-training. [2] Beginning earlier gives people time to return with another question and try the work together.
What would give us confidence that continuity is working?
For a strategic leader, I would begin by asking where the loss of one person's accumulated knowledge would materially affect the organisation's ability to operate, decide, deliver or recover. That helps prioritise the work without attempting to record every detail of everybody's role.
In a selected area, I would ask the people involved to show me how the work happens, including the exceptions that do not fit neatly into the process description. Which decisions need explanation? Who else understands them? Where does a colleague need practice, and where would a better record be enough?
We could then look at what the successor can do with what is available. In our overnight-processing example, a safe rehearsal could ask them to explain the business consequence of changing the sequence, identify a reason to pause a restart and work out who needs to be involved. The experienced colleague could observe and help afterwards, making the gaps useful learning material.
The exercise would need to fit the work and the person's responsibility. They should have access to the tools, colleagues and authorised support they would normally use, including knowing when to ask for help. What I would want to understand is how well those arrangements work when the departing expert is unavailable.
This is what I mean by knowledge stewardship: making time for people to explain, question, practise and maintain what the organisation needs. Counting interviews or stored documents tells us how much material we have collected. Observing colleagues use it gives us a better basis for deciding what support they still need.
Where AI could help
I can see a useful role for AI in supporting this work. With permission and appropriate access controls, an organisation could use it to help organise repeated interviews and debriefs, connect related explanations and make relevant passages easier to find. Someone joining months later could then have access to parts of the earlier conversation that would otherwise have been unavailable to them.
The answers would need links to their original records, with dates, context and a way for someone knowledgeable to correct what has become outdated. A fluent answer about an old process still needs scrutiny before anyone treats it as current guidance. I would also want to understand how the team manages when its AI assistance is unavailable.
Those conversations and records could give a successor a better starting point, while mentoring, observation and practice help them develop their own judgement. If I were looking at this with a leadership team, I would choose one important area of work and ask: if our most experienced person left, what would the next person need to understand, and what would give us confidence that they could use it?
Sources
1. Fortra, 2026 IBM i Marketplace Survey, pp. 2, 4, 13. Vendor-produced survey; 320 completed responses overall. Respondents selected their five leading concerns.
2. IAEA, Knowledge Loss Risk Management in Nuclear Organizations, NG-T-6.11 (2017), sections 3.3-3.5 and Table 7.
3. York Minster, York Minster seeks its next apprentice, institutional account of heritage apprenticeship and digital technology.
4. IAEA, Training and Human Resource Considerations for Nuclear Facility Decommissioning, NG-T-2.3 Rev. 1 (2022), section 5.3, pp. 26-27.
5. Historic England, Materials for Historic Building Repairs, updated 9 March 2026.
6. Gourlay (2006), Towards conceptual clarity for 'tacit knowledge': a review of empirical studies, Knowledge Management Research & Practice, 4, 60-69. The conceptual distinction used here is supported by the publisher's abstract.
7. Szulanski (1996), Exploring internal stickiness, Strategic Management Journal, 17(S2), 27-43.
8. Burmeister and Deller (2016), Knowledge Retention From Older and Retiring Workers, Work, Aging and Retirement, 2(2), 87-104.
9. Fortra, What to Do When Your AS/400 Talent Retires, vendor guidance on overlap, mentoring and automation.
10. COMMON, About the N2i Mentorship Program, programme description.