Layer 3: Workaround Intelligence
The return you approved is paying for work your system cannot see
v1.0 · 14 November 2026 · First issued 14 November 2026 · Custodian: BusinessGPS · Version index and changelog
The system went live and the manual work your business case retired did not stop. It moved outside the system, and part of your return is paying for it.
Your new system is live. The business case said a block of manual work would stop. It did not stop. It moved.
Transactions now run through the system. Your people still reconcile its outputs, repair its data and clear its exceptions outside it, in spreadsheets and inboxes. The activity has moved. The effort remains.
That effort matters most at one moment: the next spend request. A phase 2, an automation add-on or an extended support contract: each asks you to fund more change. Each sits on top of a return you approved and have not yet seen. Before you sign the second cheque, find out what the first one is paying for.
The saving can exist only in the business case
Your financial case rests on things you can count: lower overtime, less contractor spend, recruitment avoided or more output from the same team. Each one depends on capacity being released. If the people are still doing the work, by another route, the capacity was never released, and the benefit has no way to arrive.
Here is how large an ignored workaround can be. Take a hypothetical process that handles 1,000 exceptions a month. Each exception needs 12 minutes of manual work, so the process uses 200 hours a month.
At an assumed fully loaded cost of £45 an hour, the hypothetical cost is £9,000 a month. At the same volume, that hypothetical cost is £108,000 a year. These are illustrative inputs, not a client result and not a savings claim.
The point is the shape, not the sum. A cost of that kind hides behind the words “just something the team does”, and it runs every month.
Some of that effort protects the business. Some of it is avoidable. Finance needs to know which is which before it accepts the benefit forecast or funds the next fix.
The same hidden work can also limit growth, because each extra transaction brings more manual intervention with it. And at handover it can pass to operations unresolved, as extended support.
Removing the wrong workaround can destroy value
The answer is not to remove every spreadsheet or automate every manual step. Workarounds come in three kinds:
- some make up for a defect in the system
- some hold the judgement that prevents a costly mistake
- some bypass a control that should stay in place
Take a hypothetical planner who adjusts delivery promises in a local workbook. The system holds the order. The workbook adds how reliable each supplier has really been. The planner knows which customer commitments can never be treated as routine.
We place this work in the Five-Layer Activity Model, our map of where work sits, from standard software to AI agents. Layer 3, Contextual Workarounds, is the workbook, the reconciliation and the informal coordination. Layer 4, Managerial Judgment, is spotting the risk, weighing the options and deciding when to step in.
Remove the workbook without replacing what it knows, and the business makes promises it cannot keep. Automate its steps without the judgement, and the error travels faster.
So the investment question changes. What cost can you remove without removing the capability that protects the outcome? There are three treatments:
- repair the defects that create avoidable effort
- preserve, or strengthen, the expertise that protects performance
- constrain the unsafe bypasses, and fix what caused them
Research on workarounds supports asking why each one arose. Its existence alone is not proof of waste or resistance.[1][2]
The evidence you need is not all in the system
System records show corrections, delays and repeated exceptions. They cannot show why an experienced employee rejected a recommendation or trusted a different source.
So follow actual cases. Open the local files. Ask the people who do the work three things: what they noticed, what would have gone wrong and how they checked the result. Structured methods such as applied cognitive task analysis help bring out the judgement behind the visible actions.[3]
Our paper, Layer 3: Workaround Intelligence, sets out the evidence on hidden work in full. It shows how to tell the three kinds of workaround apart, and what to do with each.
This kind of evidence depends on trust. If people expect what they disclose to be used against them, you miss the very knowledge you need.
So explain what the work is for. Be honest about possible changes to roles. Let people challenge how their work is described.
Transformation Capability Management, the proposed standard we hold as custodian, names this in its sixth domain: social licence and absorption trust. It matters here because a sound investment decision needs sound evidence from the people doing the work.
Four questions show whether a promised benefit is real
Choose one material benefit that depends on less manual work. Ask finance and operations, together, to show you what has actually changed.
- Which work has stopped, and which work has moved outside the system?
- What does the remaining effort cost, and what outcome does it protect?
- What would fail if the key expert were unavailable?
- Which intervention would release usable capacity, and how would that become a measurable benefit?
If those questions have no answer, the next investment rests on a gap in the evidence. Close that gap before you assume another automation purchase will deliver the missing return.
Start with what you already hold. Before you ask operations where the work went, see what your programme’s own documents can prove.
The Baseline grades the first 9 of our 27 steps from your programme’s documents, including whether value is tracked against the business case.
Applying costs nothing. The fee is fixed and printed on the Baseline page. Cancel within 30 days of paying and you get the fee back in full, and you keep the report.
References
[1] Alter, S. (2014). Theory of Workarounds. Communications of the Association for Information Systems, 34, Article 55. https://doi.org/10.17705/1CAIS.03455
[2] Malaurent, J. and Karanasios, S. (2020). Learning from Workaround Practices: The Challenge of Enterprise System Implementations in Multinational Corporations. Information Systems Journal, 30(4), 639 to 663. https://doi.org/10.1111/isj.12272
[3] Militello, L. G. and Hutton, R. J. B. (1998). Applied cognitive task analysis (ACTA): A practitioner’s toolkit for understanding cognitive task demands. Ergonomics, 41(11), 1618 to 1641. https://doi.org/10.1080/001401398186108
The research behind this post
Layer 3: Workaround Intelligence — The Evidence on the Hidden Work of Enterprise Systems
Version index and changelog
- v1.0 (14 November 2026): first issued.