Businesses are under pressure to make better use of what their most capable people know.
That pressure is understandable. In almost every organisation, a small number of people carry a disproportionate amount of operational knowledge. They:
- Know why a customer is unhappy before the issue appears in a report.
- Can spot the missing detail in a contract.
- Understand which exception matters, which process can be bent, and which shortcut will create trouble later.
When those people are unavailable, leave, or simply have too much work, the organisation slows down. Others wait for answers. Decisions are repeated. Customers receive inconsistent responses. Important work accumulates around the people who already have the least spare capacity.
AI and automation make a tempting promise: capture what the best people do, turn it into a system, and make that capability available to everyone.
There is a problem with that framing.
It can sound like the organisation wants to extract expertise from people, convert it into software, and make the people themselves less valuable. Workers are right to be cautious when that is the unstated plan.
The better ambition is not to make experts replaceable. It is to make expert judgement available without requiring an expert to personally intervene in every routine case.
Expertise is more than repeatable activity
The most visible part of expertise is often the easiest part to automate.
An experienced operations manager may know how to check a request, find the relevant documents, apply a standard rule and prepare a response. Those are useful practices to make more available.
- They reduce delays.
- They help newer team members work with greater confidence.
- They give customers more consistent answers.
But the person’s real value is rarely limited to those steps.
Expertise also includes judgement. It includes the ability to recognise when the rule does not fit, when the information is incomplete, when an unusual pattern is emerging, or when a customer situation needs a different response. It includes accountability for a decision and an understanding of the consequences that may not appear in any process map.
A system can support that work. It should not pretend to erase it.
The design goal is to separate routine access to proven practice from the judgement that only experience can provide.
A better use of intelligent systems
A well-designed intelligent system can handle the work that repeatedly interrupts skilled people without adding much value to their day.
It can:
- Retrieve the right current information.
- Prepare a first draft.
- Guide a colleague through a known process.
- Identify missing inputs.
- Apply defined rules and show the evidence behind an answer.
- Escalate when the situation falls outside agreed boundaries.
That changes the expert’s role.
Instead of answering the same question twenty times, they can review difficult cases.
Instead of searching for basic information, they can improve the operating method.
And, instead of being the only person who knows how the work gets done, they can mentor others, solve emerging problems and shape better decisions.
The organisation becomes less dependent on one person being available at precisely the right moment. The person becomes more available for the work that genuinely requires their capability.
That is an extension of expertise, not a reduction of it.
The worker must be a co-designer, not a source to be mined
This distinction depends on how the work is done.
If an organisation asks an experienced employee to document everything they know, builds a system from it, and then uses the system to reduce their role, it has treated knowledge as something to extract. The likely result is distrust, defensive behaviour and a poorer system. The employee will protect the nuance that the process missed, often because they know that nuance is what prevents mistakes.
A stronger approach makes the expert a co-designer and steward of the capability.
This is an important distinction because the expert is there to help define what good work looks like and identify where the process is safe to standardise and where it must stop.
They specify exceptions and can also decide what evidence a system should show before a decision is accepted and review failures and improve the method over time.
Their contribution is not “training the replacement”. It is designing the conditions under which more people can work safely and well.
That is a more senior and more valuable role.
The questions that reveal the real intent
Businesses should be direct about why they are introducing intelligent systems.
The following questions help separate a capability-building program from a cost-cutting exercise disguised as knowledge management:
- Will experts spend less time on low-value repetition after the system is introduced?
- Will they have more time for complex cases, customer relationships, coaching and improvement?
- Do they have authority to define boundaries, exceptions and escalation points?
- Can they see, challenge and improve the system’s recommendations?
- Is the organisation creating recognised career paths for practice leadership and knowledge stewardship?
- Are productivity gains being shared through better tools, less interruption, meaningful development or improved service capacity?
- Is management being honest about any workforce implications?
There is no virtue in pretending that technology has no effect on roles. It will. But trust depends on whether people can see a credible future for their contribution.
Design for dignity as well as efficiency
Organisations often treat efficiency and human dignity as competing objectives. They are not necessarily in conflict.
A team member who spends hours each week copying information between systems, answering the same basic questions or reconstructing the history of a case is not being honoured by that work. Removing it can be a genuine improvement.
The issue is what replaces it.
If the answer is more surveillance, tighter targets and less discretion, the system has reduced the quality of work. If the answer is greater capacity for judgement, customer care, learning and improvement, the system has raised it.
That is a design choice. It should be made consciously.
From knowledge capture to capability architecture
At Semantia, we think of this as capability architecture.
The task is not to capture everything an expert knows. It is to identify the parts of a workflow that can be made more reliable and accessible, connect them to current sources of truth, define safe boundaries, and preserve a clear path back to human judgement.
The result should make the organisation more resilient without making its people feel disposable.
The test is simple:
Does the system remove routine interruption from skilled people and create more room for the work only they can do?
If it does, the organisation is extending expertise.
If it merely converts experienced workers into unpaid trainers for their replacement, it is extracting value from people while pretending to build capability.
Those are not the same thing. The distinction should shape every intelligent automation program from the start.