Abstract
A transformation can concentrate on one part of an organisation while depending on conditions elsewhere. This article examines the relationships between people, processes, data and technology, using hypothetical ERP and agentic AI examples. It argues that a focused delivery scope still requires attention to relevant dependencies, shared information and cross-functional decisions. Drawing on socio-technical research and AI risk-management guidance, it distinguishes understanding those connections from assuming that everything must change. For executives and transformation leaders, the implication is to define the capability the organisation needs, examine the arrangements supporting it and make proportionate choices about what to change or preserve.
enterprise transformation; people processes data and technology; operating model; ERP implementation; agentic AI; cross-functional dependencies
A focused change can have wider consequences
When I ask whether you can transform one part of an organisation without changing the others, people may think first about departments. Can finance transform without manufacturing? Can supply chain change without involving sales? Those are useful questions, but I am also interested in what sits beneath the organisation chart.
People, processes, data and technology are closely connected in the work an organisation performs. A transformation may concentrate its investment on one of those areas. It still needs to examine what that change will require of the others.
That does not mean every initiative must become an enterprise-wide programme. It means the boundary drawn around a project cannot, by itself, establish where its consequences will stop. The task is to understand the relevant dependencies well enough to make deliberate choices about what needs to change and what can remain.
People experience the connections through their work
A person completing an order is following a process, using information, interacting with technology and exercising judgement. They may also depend on someone in another function to confirm a requirement or resolve an exception. From their perspective, these are connected parts of getting the work done.
I use people, processes, data and technology as practical lenses for examining those connections. They are not four independent workstreams or an exhaustive account of an organisation. Leadership, decision authority, incentives and culture influence how each operates.
This perspective is consistent with the socio-technical approach examined by Baxter and Sommerville (2011). Their work connects organisational change with systems development and considers the practical challenges of bringing those perspectives together. It supports examining technical and organisational arrangements in combination; it does not prescribe this particular four-part lens.
An apparent technology problem may therefore require attention to the process or information that surrounds it. Equally, a capable employee may struggle because responsibility for a decision is unclear. I would want to understand those relationships before treating the most visible difficulty as the whole problem.
ERP configuration involves choices about how the business operates
An enterprise resource planning system, or ERP, is a useful example. Consider a hypothetical manufacturer replacing fragmented systems to improve the reliability of orders, purchasing and production planning. The software introduces opportunities to connect information that previously sat in separate places.
It also introduces business decisions. Which product details are required before an order can proceed? Who maintains lead times? How should changes to a customer commitment reach production planning? Which exceptions require approval, and who can provide it?
The implementation team will need answers to configure the system. Yet the consequences extend into the responsibilities, processes and information that people will depend on after the team has finished. An unanswered business question can become a configuration assumption with a long operational life.
Markus (2004) describes technology-enabled organisational change as requiring a technical solution alongside complementary organisational changes. Her analysis highlights the risk of misalignment with organisational characteristics, including incentives. That is a useful basis for asking whether an implementation decision fits the business arrangements around it.
Standard ERP processes may be an appropriate choice. The judgement is whether they support the intended way of operating, what changes their adoption requires and which differences matter. Neither keeping every existing variation nor accepting every software default resolves those questions automatically.
AI brings these dependencies into sharper view
The same questions become important when AI is introduced into operational work. A proposed use may rely on information whose definitions, ownership or limitations have not been settled. A technically impressive output leaves open whether it is appropriate for the decision the organisation wants to make.
In this article, an agentic system means an AI-enabled system authorised to carry out defined actions through connected tools or systems. That description is deliberately bounded: the actions available depend on the particular implementation and its permissions. A demonstration of an action is not evidence that it is reliable in ordinary operation.
Imagine the manufacturer considering an agentic system that can initiate replenishment requests. Ambiguity about available stock could then influence an action that affects procurement and production. The organisation would need to decide the system’s authority, the situations requiring human involvement and how errors or exceptions would be handled.
The NIST Artificial Intelligence Risk Management Framework treats AI as socio-technical and addresses organisational context, data-related risks and human responsibilities (Tabassi, 2023). Its governance guidance includes defining roles and oversight arrangements. This is voluntary risk-management guidance rather than evidence that a particular AI deployment will deliver benefits.
My concern is the fit between the proposed capability and the organisation expected to use it. Introducing AI can create an opportunity to improve information and processes. It should not be taken as evidence that those underlying arrangements have already been resolved.
Interdependence does not require everything to change
Some organisations have capable people, reliable information and workable processes constrained by outdated technology. Their most significant investment may reasonably be in system replacement. Elsewhere, a useful system may already exist while inconsistent responsibilities or processes limit its value.
Recognising those differences is part of keeping a transformation proportionate. Understanding the connections should help narrow the work where appropriate, as well as identify changes that were initially missed. It should not become an argument for replacing arrangements that are already effective.
A useful distinction is between the scope of the change being delivered and the dependencies that need to be understood. A project can remain focused while examining a neighbouring process, confirming a data assumption or resolving an approval responsibility outside its direct delivery boundary.
The required depth depends on the intended outcome and the consequences of getting the assumptions wrong. Dependencies can also become clearer during implementation. A bounded programme needs a way to address those discoveries without either ignoring them or expanding automatically whenever an issue appears.
Someone must be able to resolve cross-functional choices
An integrated view of the work will sometimes expose competing interests. One function may benefit from standardisation while another loses flexibility. A faster process may shift checking work downstream. Changing ownership of information may place a new demand on a team whose capacity is already committed.
These are choices about how the organisation operates. A delivery team can surface the consequences and develop options, but it may lack the authority to settle priorities across functions. Executives need to be clear about who can make those decisions and what outcomes should guide them.
I would also want the people doing the work involved in understanding the consequences. Their experience can reveal an exception or dependency that an executive description overlooks. Participation contributes evidence; accountable leaders still need to make and support the resulting decisions.
Begin with what the organisation needs to be capable of doing
The question I would bring forward is what the organisation needs to be capable of doing differently when the transformation is complete. That creates a starting point for examining the work, the information it needs, the responsibilities involved and the technology that can support it.
It also changes how an apparently local improvement is assessed. A new capability in one function is useful when its contribution to the intended outcome is understood, including consequences for other functions and the people using it. Delivery activity alone leaves those relationships uncertain.
So, can you transform one part of an organisation without changing the others? Sometimes much of the surrounding organisation can remain as it is. What matters is establishing which connections affect the outcome and whether the proposed change can work within them. That understanding allows the transformation to be focused without treating the rest of the organisation as irrelevant.
References
Baxter, G. and Sommerville, I. (2011). Socio-technical systems: From design methods to systems engineering. Interacting with Computers, 23(1), 4–17. https://doi.org/10.1016/j.intcom.2010.07.003
Markus, M. L. (2004). Technochange management: Using IT to drive organizational change. Journal of Information Technology, 19(1), 4–20. https://doi.org/10.1057/palgrave.jit.2000002
Tabassi, E. (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0). NIST AI 100-1. National Institute of Standards and Technology. https://doi.org/10.6028/NIST.AI.100-1
Publication and disclosure notes
Adapted from Dr Liz Wilson’s article “Can You Transform One Part of an Organisation Without Changing the Others?”, originally published on thedrliz.com. https://thedrliz.com/can-you-transform-one-part-of-an-organisation/
The manufacturer, ERP, stock-definition and agentic replenishment scenarios are hypothetical illustrations. They are not reported client cases, tested configurations or claims about the performance of a particular product.
The author is Founder & Editor of Transformation Expert Insights and works professionally in behavioural and organisational transformation.
AI assistance: OpenAI ChatGPT assisted with adapting the original article, restructuring the argument, drafting wording and identifying and checking references. The author is responsible for the published content.