Assumptions

The complete state of a company is not directly observable. Companies publish and generate information about themselves, but they rarely state every need, constraint, priority, intention, capability gap, or internal reason behind their actions. Even when substantial information is available, there can still be gaps between what is observed and what someone wants to understand.

Suppose ASUME observes that a company has opened a large number of positions in a particular function. The hiring activity is supported by evidence. To conclude that the company is expanding that function, however, the reasoning may depend on whether those positions represent additional capacity rather than replacements for people who have left.

If the available information does not establish that distinction, ASUME should not silently assume that it does. The missing step can instead be made explicit:

The positions primarily represent additional capacity rather than replacement hiring.

The purpose of the assumption is not to manufacture an answer. It is to identify the part of the reasoning that remains unconfirmed.

Evidence and assumptions play different roles in ASUME's understanding of a company. Evidence has an identifiable basis in information available to ASUME. An assumption appears where that evidential basis does not directly establish something that the reasoning may need.

Consider the following statement:

The company has published fifteen vacancies for implementation engineers.

If that information is present in the available sources, it can serve as evidence.

Now consider:

Most of those positions are net-new roles.

Unless ASUME has information establishing that point, it should not be treated as evidence. It may instead be an assumption relevant to interpreting the hiring activity.

Together, the reasoning might support:

The company appears to be expanding its implementation capacity.

The distinction matters because the final conclusion may sound equally convincing whether the intermediate assumption is visible or not. Keeping the assumption explicit shows where the evidence ends and where additional reasoning begins.

An assumption should not be understood as an arbitrary guess. Its purpose is not to give ASUME permission to speculate whenever information is missing.

A useful assumption is tied to a specific gap in the reasoning and should remain consistent with the evidence and context currently available. If ASUME observes sustained hiring in a new function, the creation of a leadership role for that function, related product changes, and increased discussion of the area in company communications, the evidence may support the view that the function is becoming more important. Any remaining assumption should relate to something the available information still does not establish directly.

This gives assumptions a constrained role. They help ASUME continue reasoning where incomplete information would otherwise create an unspoken gap, but they do not remove the uncertainty created by that gap.

The important question is therefore not simply whether an assumption sounds reasonable. It is whether the assumption is necessary for the conclusion, whether it is compatible with the available evidence, and whether its uncertain status remains visible.

One of the most useful roles of an assumption is to show exactly what the evidence has not established.

Suppose ASUME observes that a company has opened operations in three new countries and is hiring regional sales leaders in each market. Those observations may provide substantial evidence of international activity. A stronger conclusion, such as:

International expansion is a major current strategic priority.

may still require additional interpretation. The evidence may show that expansion is happening without establishing how important it is relative to the company's other priorities.

An assumption can make that missing step visible rather than allowing it to disappear inside the conclusion. This gives users a clearer distinction between what ASUME found, what ASUME had to assume, and what ASUME inferred as a result.

That distinction is particularly important when several pieces of information together create a compelling narrative. Coherence can strengthen an interpretation, but a coherent interpretation is not automatically the same as a directly established one.

Two conclusions can describe the same company state while resting on very different foundations.

Suppose a company explicitly announces that it is opening a new manufacturing facility. A conclusion that manufacturing capacity is increasing may be closely supported by direct evidence.

Now suppose ASUME instead observes manufacturing hiring, procurement activity, and a newly leased site. The same conclusion may still be reasonable, but it could depend on an assumption about what the site will be used for.

The conclusions may look similar when written as natural language, but their support is different. The second interpretation depends more heavily on something that has not been directly established.

ASUME therefore should not evaluate a conclusion only by asking whether it sounds plausible. The path leading to the conclusion matters. Where important assumptions remain unresolved, that should influence how confidently the conclusion is represented and how easily it can change when new evidence appears.

An assumption may matter beyond a single statement.

Suppose ASUME assumes that a large increase in engineering vacancies represents net organisational growth rather than replacement hiring. That assumption could affect several related conclusions: that engineering capacity is increasing, that the company is investing in a particular capability, or that an initiative is becoming more important.

If new information later shows that most of the vacancies were replacements, more than one conclusion may need to be reconsidered.

This is one reason assumptions need to remain part of the maintained understanding rather than disappearing after an output is produced. ASUME needs to know which conclusions depend on uncertain premises so that the relevant parts of the understanding can change when those premises change.

At the same time, a change in one assumption should not make unrelated parts of the company's understanding disappear. Conclusions that remain independently supported can continue to stand.

Assumptions are provisional by design. New evidence can strengthen an assumption, weaken it, contradict it, make it unnecessary, or show that a different interpretation is more appropriate.

Suppose ASUME initially assumes that a company's new cybersecurity vacancies represent additional internal capacity. Later, the company states that the hiring programme is intended to replace an outsourced provider with an internal team.

The new information changes what is known. Part of what previously had to be assumed may now be directly supported, while other conclusions built on the earlier interpretation may need to change.

This is why an assumption does not become permanently valid simply because it has existed in the understanding for some time. Its status remains connected to the evidence available around it.

A maintained understanding should therefore be able to preserve useful assumptions without allowing them to harden silently into facts.

When an assumption no longer holds, ASUME needs to reconsider conclusions that depended on it.

If the interpretation that a hiring programme represents net expansion turns out to be wrong, ASUME should not remove that assumption while leaving every downstream conclusion untouched. Statements that relied materially on the assumption may need to be weakened, revised, or removed.

The change should remain proportionate to what the assumption actually affected. If the assumption concerned one capability, unrelated evidence about the company's products, customers, markets, or other capabilities does not automatically become invalid.

This gives maintained understanding both continuity and revisability. ASUME can change the parts of its understanding whose basis has changed without treating every new piece of information as a reason to reconstruct the entire company.

Assumptions become particularly important when the same evidence can support several interpretations.

Suppose a company suddenly increases hiring. That observation might reflect organisational expansion, replacement hiring, the creation of a new capability, internalisation of work that was previously outsourced, restructuring between teams, or a combination of these changes.

Different interpretations may depend on different assumptions. If ASUME assumes that most of the positions are new, expansion becomes more plausible. If it assumes that the roles largely replace departures, the same hiring activity may support a different understanding.

Where the available evidence cannot distinguish between these possibilities, ASUME does not need to force one interpretation into certainty. The competing explanations, and the assumptions on which they depend, can remain visible until better information becomes available.

This helps prevent a reasonable possibility from being silently promoted into the only explanation simply because the system needs to produce an answer.

An assumption is not necessarily valid everywhere or indefinitely.

Its usefulness may depend on a particular time period, geography, product, business unit, market, organisational context, or another condition relevant to the company.

For example:

The company is prioritising cost reduction.

might be a reasonable assumption for one business unit during a restructuring period. That does not mean cost reduction should automatically be treated as a permanent company-wide priority.

ASUME therefore needs to keep the context of an assumption connected to the reasoning in which it is used. An assumption that is appropriate for one part of the company should not silently spread to unrelated areas, and an assumption that was reasonable at one point in time should not remain applicable simply because nothing explicitly removed it.

Scope helps keep assumptions useful without allowing them to become broader than their basis supports.

Assumptions about the underlying state of a company can matter across more than one objective.

Suppose an assumption affects ASUME's understanding of a capability gap. A Revenue objective might use that understanding when evaluating a commercial opportunity, while a Partnerships objective might use the same underlying state when considering strategic complementarity. If the assumption changes, both outputs may need to change because they depend on the same company understanding.

This follows the broader AUM principle that ASUME understands the company before applying a particular objective. The objective determines which parts of the understanding matter for the task, but it does not create a separate version of the company.

Assumptions therefore belong to the maintained company understanding where they concern the company itself. Their relevance to a particular output depends on how that output uses the underlying state.

Because an assumption can materially affect a conclusion, it should be possible to inspect its role in the understanding. The useful questions are practical: what is being assumed, why is the assumption needed, what evidence exists around it, which conclusions depend on it, where does it apply, and what kind of new information could strengthen or challenge it?

This does not require exposing hidden chain-of-thought or every internal computational step. The purpose is to make the relevant dependency visible. A user should be able to understand when a conclusion depends materially on something that ASUME has not directly established.

This also makes assumptions challengeable. A user may know something that ASUME does not, new evidence may become available, or a previous interpretation may no longer fit the company. When the assumption can be identified, the corresponding part of the understanding can be reconsidered rather than leaving an unsupported premise hidden inside the output.

This contributes directly to ASUME's approach to traceability.

The current status of an assumption is important, but sometimes the way it changed is useful as well.

An assumption may have been reasonable when ASUME had only limited public information and later become unnecessary when more direct evidence appeared. Another assumption may have initially fitted the available evidence and then been rejected after the company changed or contradictory information became available.

Preserving relevant changes allows ASUME to explain not only what the current understanding depends on, but also why that understanding changed. This is particularly useful when two outputs produced at different points in time differ because the underlying evidence or assumptions changed between them.

The purpose is not to preserve every intermediate thought indefinitely. It is to preserve enough revision context that a meaningful change in the company's understanding does not appear arbitrary.

An assumption is not evidence, and it is not a fact. It does not become true because it sounds plausible, because several conclusions depend on it, because it has remained unchallenged for a long time, or because an output based on it happened to be useful.

An assumption is also not an unsupported guess inserted to force ASUME toward an answer. Its role is the opposite: to expose where reasoning depends on something the available evidence has not established.

When that dependency is visible, the resulting conclusion can be interpreted more appropriately, challenged more easily, and revised when better information becomes available.

That is why assumptions are central to AUM. They allow ASUME to reason beyond what can be directly observed without hiding where the understanding goes beyond the evidence.