Project scoping questions that make the work clearer

We use project scoping questions to establish the decision a client needs to make before agreeing the output. A request for an AI tool, reporting dashboard or valuation can be a useful starting point. The scope becomes clearer when both parties can explain what the proposed work must change and who will decide whether it has done so.

A scope is an agreed boundary around purpose, inputs, deliverables and responsibilities. Our published Squared Method begins with diagnosis before design and deployment. This note explains the preparation behind that approach and provides a specimen scoping record for a first conversation.

The specimen describes an intended working method. It is not a client case history or a claim about completed engagements. An urgent cash question and a valuation for a fixed reporting date require different investigations, so the initial discussion should establish the evidence needed for that particular decision.

Which project scoping questions come first?

The first question is what made the work necessary now. A board deadline creates a different constraint from a growing exception queue or the departure of a finance employee. We use that event to establish the consequence of delay and the decision that must be supported by the proposed work.

Ask what decision is currently difficult to make. A founder who requests a dashboard may need to know which customer contracts consume cash. An operations manager requesting automation may need a reliable way to route cases that currently depend on one employee. Describe the decision in ordinary language before translating it into a system requirement.

Identify the person who owns that decision. This may differ from the person sponsoring the project or maintaining the existing spreadsheet. The decision owner should explain what information they trust, what remains uncertain and which actions they are permitted to take when the answer changes.

Clarify the boundaries early. A finance reporting project does not automatically include legal advice, tax structuring or a new enterprise system. An AI workflow does not automatically include authority to communicate externally. A written boundary gives both parties a way to recognize when the requested work has expanded.

We request the evidence needed for the next decision

We start with a small sample relevant to the problem. An anonymized report, process description or list of exceptions may establish the next question without requiring a complete data room. Broader access should follow a defined need, with an agreement about who can use the information and for what purpose.

Privacy and confidentiality affect how those samples are shared. The PDPC's guidance on key concepts in the PDPA, reviewed in 2026, explains obligations concerning the purposes and handling of personal data. At the scoping stage, a practical response is to identify why information is needed and avoid requesting personal details that do not affect the assessment.

Look for conflicting definitions. If sales and finance report different revenue figures, the first deliverable may need to establish the meaning and source of each measure. Automating both reports without resolving that disagreement can make the inconsistency arrive faster without making the decision easier.

For valuation work, identify the measurement purpose and applicable reporting context. The IFRS Foundation's IAS 38 overview, reviewed in 2026, illustrates why recognition and measurement questions need defined criteria. A broad request to value intellectual property should be narrowed to the subject, date and intended use of the analysis.

Agree how the client will accept the work

We describe a deliverable in terms the client can inspect. A request to improve visibility might become a reporting pack with agreed definitions and a reconciliation to source records. The scope also needs an owner and a review routine. Those features make the output verifiable while keeping the wider business result separate from what the engagement can control.

Write acceptance criteria before implementation. For a forecast, that might include a reconciled opening balance and a documented update procedure. For a workflow, it might include successful handling of named exception cases and a tested human handoff. The appropriate criteria depend on the consequences of an incorrect result.

NIST's 2023 AI Risk Management Framework gives AI projects a useful reference for identifying and managing risks. A scoping note should specify which risks the proposed work addresses and which decisions remain with the client. Broad references to responsible AI are insufficient if nobody can identify the approval route.

Record assumptions as conditions. Access to a source system, availability of a subject-matter expert and agreement on a data definition can all affect delivery. If a dependency is uncertain, the scope should identify the decision that resolves it and what changes if it cannot be met.

Record the next commitment and its dependencies

The specimen scoping note records the problem, decision owner, baseline, proposed output, exclusions, dependencies and acceptance evidence. It also identifies what remains unknown. A document that makes uncertainty visible gives the next discussion a clear purpose.

Sometimes the next step is a limited diagnostic. Sometimes enough evidence exists to define implementation. In other cases, the business first needs to resolve a policy, ownership or data-access question internally. The initial conversation should make that sequence explicit, including who will provide the missing information.

The cash forecast guide and AI launch checklist show how a broad objective becomes specific evidence and operating responsibility. They are useful preparation when the problem falls within those areas, and they can expose questions worth settling before a proposal is written.

We aim to end the first conversation with a bounded next commitment. That may be a diagnostic, an implementation brief or a specific question the client needs to settle internally. The written record should identify the owner and evidence needed before the next decision. A proposal is useful when both parties can explain those commitments in ordinary language and recognize when the requested work has changed.