The Work That Was Not Logged: Finding Hidden Work and Shadow Processes With AI
Some of the most important work in an organisation is not visible in the system of record. It happens in side chats, inboxes, spreadsheets, personal notes, copied templates, manual workarounds and quiet fixes that keep the business moving.
Every organisation has two versions of how work gets done
Every organisation has two versions of how work gets done.
There is the official version: the process map, the CRM, the workflow, the ticketing system, the project plan, the policy and the dashboard.
Then there is the real version.
The real version includes the work people do to make the official version function. The checking, chasing, interpreting, copying, reconciling, reformatting, reminding, escalating and fixing that never appears in the system.
This is the work that was not logged.
It is the hidden work sitting underneath business-as-usual. It is often invisible to leaders, undercounted in process reviews and missing from transformation plans. It is also one of the best places to look for practical AI and automation opportunities.
Key point: Hidden work is not always waste. Sometimes it is the organisation’s survival mechanism. The risk is that nobody is measuring it, designing for it or deciding whether it should exist.
What hidden work actually means
Hidden work is the work people do that is necessary to keep things moving but is not formally captured, measured or understood.
It may not appear in the process map. It may not show up in the CRM. It may not be counted in workload reports. It may not be recognised in job descriptions.
But without it, the system would slow down or fail.
This work is often carried by the most capable people in the team. They know where the process breaks. They know who to call. They know which system cannot be trusted. They know the workaround.
The problem is that the organisation often mistakes this hidden labour for normal performance.
That is why hidden work should be treated as a serious signal in process improvement, workflow automation and AI strategy.
Why hidden work matters
Hidden work creates several business risks.
It hides the true cost of delivery
If a process looks efficient on paper but only works because people are doing unpaid coordination, manual checking or repeated correction, the organisation is underestimating the cost of delivery.
This matters when leaders make resourcing, pricing, technology or service-level decisions.
A process that appears to take 20 minutes may actually take two hours once the undocumented work is included.
It creates dependency on key people
Hidden work often depends on individuals who know how the system really works.
When those people leave, move roles or become overloaded, the organisation suddenly discovers that the process was not as stable as it looked.
The knowledge was not in the system. It was in people’s heads.
It makes automation risky
If you automate the official process without understanding the hidden work beneath it, you may automate the wrong thing.
The organisation can end up with faster handoffs, cleaner dashboards and worse outcomes because the human workaround that made the process safe has been removed.
This is why Changeable does not treat workflow automation as a tool-selection exercise. The real work has to be understood first.
It weakens decision-making
Leaders often rely on system data to understand workload, performance and bottlenecks.
But if the real work is happening outside the system, the data is incomplete.
The dashboard may say the process is fine while the team is quietly drowning in rework.
It increases burnout risk
Hidden work is exhausting because it is rarely recognised.
People absorb complexity, protect customers, correct errors and bridge gaps without the organisation formally seeing what they are carrying.
Over time, that becomes a workload, wellbeing and retention issue.
WorkSafe New Zealand’s guidance on managing psychosocial risks at work is a useful reminder that workload, role clarity, support and poorly managed change all matter. Hidden work often sits right in the middle of those risks.
Useful distinction: A process issue is not always where the official workflow breaks. Sometimes it is where people have quietly learned to compensate for the workflow.
Shadow processes: the unofficial operating model
Shadow processes are the unofficial workflows people create when the formal process does not meet the needs of the work.
They are not always malicious. In fact, they are often created by people trying to do the right thing.
A team creates a spreadsheet because the system report is too slow. A manager keeps a personal tracker because approvals get lost. Staff use a shared chat thread because the workflow tool does not reflect how decisions actually happen. Someone creates an unofficial checklist because the documented process is incomplete.
These shadow processes are signals.
They tell you where the official process is too slow, too rigid, too unclear or too disconnected from operational reality.
This is similar to the logic behind shadow IT. Gartner has described shadow IT as technology being acquired or used outside formal IT visibility, often because teams are trying to solve real work problems quickly. The same pattern applies to shadow processes. People build unofficial ways of working because the official system is not keeping up.
The risk is not that people are trying to solve problems. The risk is that the organisation cannot see the solutions they are relying on.
Why AI makes shadow work more visible
AI can help organisations see hidden work because it can analyse large amounts of unstructured information.
Hidden work often lives in places that traditional process reviews do not analyse properly.
These sources contain patterns. They show repeated questions, recurring clarifications, manual corrections, delays, decision bottlenecks and workarounds.
AI can help identify those patterns faster than a manual review, especially when the organisation has too much text-based operational material for a person to read line by line.
This is not about spying on staff. It is about understanding how work actually happens so the organisation can improve it.
That means AI use needs to be governed carefully, with privacy, transparency and purpose clearly defined. The Office of the Privacy Commissioner’s Privacy Act 2020 principles are an important reference point where personal information may be involved.
What AI can look for
AI can be used to detect operational signals that are difficult to see manually.
Practical rule: AI should be used to find patterns in work, not to blame individuals for how they have adapted to broken systems.
The difference between useful hidden work and waste
Not all hidden work should be eliminated.
Some hidden work is valuable judgement.
A senior administrator may spot a risk before it becomes a problem. A customer support person may rewrite a response so it lands better. A project coordinator may know which stakeholder needs a phone call instead of an email. A business analyst may translate unclear requirements into something a delivery team can actually use.
That kind of hidden work is not waste. It is expertise.
The goal is not to automate everything that is invisible.
The goal is to distinguish between:
This distinction is important for identity-safe automation. If AI is used to remove the work that gives people meaning, it can damage trust and adoption.
If AI is used to remove the repetitive work that gets in the way of people’s expertise, adoption becomes much easier.
How hidden work becomes AI opportunity
Hidden work is one of the best sources of practical AI use cases because it shows where people are already compensating for the system.
That means the need is real.
The organisation does not have to invent a use case. It can find one by looking at where people are already spending time.
These are strong candidates for AI use case discovery because they are grounded in real operational pain.
The key is to start small and measure whether the hidden workload actually reduces.
Why process mining is not always enough
Process mining and workflow analytics can be powerful, but they usually depend on system event logs.
That means they are strongest when the work already happens inside structured systems.
But much hidden work does not.
It happens in conversations, documents, spreadsheets, meetings, inboxes and personal judgement. It happens before the official task is created or after the task is marked complete.
That is why AI analysis of unstructured material can complement more formal process mining.
It can help reveal the context around the event log:
The event log may show what happened. The hidden work often explains why.
The governance problem
Using AI to find hidden work requires care.
Employees may reasonably worry that AI analysis of emails, chat messages or notes will become surveillance.
If the project is framed as monitoring people, trust will collapse.
The governance model needs to define:
Good AI governance makes this kind of analysis safer and more useful.
The point is to improve the work system, not judge people for keeping it alive.
A practical method for finding hidden work
A practical hidden-work review does not need to be overcomplicated.
Pick a process with visible pain
Start with a workflow where people already know something is wrong.
This might be customer enquiries, onboarding, reporting, approvals, complaints, procurement, contract management, case handling or internal service requests.
Gather the official process view
Document the formal process. Identify the system steps, roles, handoffs, decision points and expected outputs.
This gives you the official version of the work.
Gather the lived-work evidence
Look at the material around the process: emails, notes, tickets, spreadsheets, chat threads, templates, meeting notes and manual trackers.
Only use sources that are appropriate, approved and governed.
Use AI to identify patterns
AI can help cluster repeated issues, extract common phrases, identify recurring manual tasks and surface workarounds.
The output should be treated as a discovery aid, not a final conclusion.
Validate with the people doing the work
The team should review the findings.
They can explain which hidden work is valuable, which is waste, which is risk and which workaround exists because the official process does not work.
Decide what to fix, formalise or automate
Some hidden work should be removed. Some should be turned into a formal step. Some should be automated. Some should be protected because it represents expert judgement.
This is where process improvement, workflow automation and AI agents can be designed properly.
What hidden work tells leaders
Hidden work gives leaders a more honest view of the organisation.
It shows where process documentation is inaccurate. It shows where people are absorbing complexity. It shows where systems do not fit the work. It shows where customers, staff or stakeholders are being protected from process failure.
It also shows where AI can be useful.
Not in the abstract, but in the actual work.
That is where the best AI opportunities usually live: close to the friction, close to the people doing the work and close to the repeated manual effort nobody has had time to fix.
The best AI use cases often start where people are already doing invisible work to keep the process alive.
What Changeable helps with
Changeable helps organisations find the gap between documented process and operational reality.
That includes identifying hidden work, shadow processes, manual workarounds and practical AI opportunities.
Start with a Decision Clarity Session
A Decision Clarity Session is a no-obligation conversation where we listen to what you are trying to achieve, what is getting in the way and whether hidden work, shadow processes, process improvement, AI or automation is the right next step.
Frequently asked questions
What is hidden work?
Hidden work is work that people do to keep the organisation moving but which is not formally captured, measured or recognised. It often includes manual checking, chasing, fixing, reformatting, interpreting and coordinating outside official systems.
What are shadow processes?
Shadow processes are unofficial workflows created by teams when the formal process does not fit the real work. They may include spreadsheets, side chats, personal trackers, informal approvals or workaround routines.
Why is hidden work a problem?
Hidden work can hide true workload, create dependency on key people, distort process data, increase burnout risk and make automation projects fail because the real work has not been understood.
Can AI find hidden work?
Yes. AI can help identify patterns in unstructured work material such as emails, tickets, meeting notes, documents and workflow comments. It can surface repeated issues, manual fixes, bottlenecks and shadow processes for human review.
Is using AI to analyse work a privacy risk?
It can be if handled poorly. Organisations need clear governance, transparency, data minimisation, purpose limits and privacy controls. The goal should be to improve work systems, not monitor or blame individuals.
Should all hidden work be automated?
No. Some hidden work represents valuable expertise and judgement. The goal is to distinguish between hidden work that protects quality and hidden work that exists because the process is broken.
How does this relate to AI strategy?
Hidden work reveals practical AI opportunities. Instead of asking “how can we use AI?”, organisations can ask “where are people already doing repetitive, invisible work that AI could safely support?”
Find the work your systems are not showing you.
Changeable helps organisations uncover hidden work, shadow processes and manual workarounds so AI, automation and process improvement can be grounded in how the work actually happens.