Engagement types and risk-based planning
14 min read
Lesson objectives
- Select an assurance scope from a risk inventory and distinguish assurance from bounded advice
Please sign in to use all the tools
Once you sign in, the reading tools are yours: larger or smaller text, typeface, line spacing, focus mode, highlighting and bold, and notes saved to your account and gathered in My study.
Imagine an audit committee has ten days of testing while the IT inventory contains forty systems. Giving every system a quarter of a day feels even-handed but may provide little assurance where it matters. Risk-based planning begins with an audit universe: activities, services, and entities that could be audited. Link each to business objectives, possible impact, likelihood of control failure, environmental change, and previous audit results. A red risk rating is a prompt for inquiry, not an explanation. Ask what would happen to customers, services, and data if the control failed, and what evidence supports the estimate.
The engagement type follows the question being answered. In assurance work, the subject, criteria, and scope are agreed and the auditor obtains enough evidence for an independent conclusion. In advisory work the auditor may analyze choices for management but does not make management's decision or offer assurance that was never tested. A focused compliance review compares practice with a requirement that applies in its context. A design review asks whether a control could meet its objective if operated as designed. An operating test asks whether it actually worked over a defined period. These approaches can coexist, but the distinction matters: finding a written policy does not prove it operated.
For prioritization, make a small record for each subject: asset or service, effect on a business objective, exposure signal, existing controls, last review, and effort. Rank candidates and test the ranking's sensitivity. Would the decision change if likelihood moved by one level? A product of likelihood and impact does not create statistical precision when its inputs are judgmental. Exceptions also matter. A contractual obligation can require a review even when the comparative risk score is lower. A newly migrated service may need attention despite a favorable old audit. Then state the engagement objective, boundaries, period, organizational units, criteria, and likely evidence sources. Discuss the scope with stakeholders while recording reasons for any exclusion rather than letting a manager silently remove an awkward area.
Test the plan by asking whether another reviewer could understand why one system was included and another was not, and what a concluding report could legitimately say. A ranking without criteria, period, or access limits is not a completed engagement plan. The conclusion will be bounded by the systems and tests actually covered, not a blanket assurance over all enterprise technology.
A newly acquired business reaches the audit committee with an urgent request. Its manager can provide contract records only next month, while the board wants an opinion next week. Begin with the decision the board will make from your work. Does it need assurance that controls are designed, or that they operated throughout the year? Those questions require different evidence. Design may be examined through system descriptions and authorized access tests. A conclusion about a year of recurring reviews needs records from that year. Establish the distinction before allocating team days, because it determines the engagement that can actually be completed.
Build the audit universe around traceable services. “Contract system” may name one application, but the service can depend on supplier records, document storage and payment approval. Record an owner, objective, recent change and evidence source for each service. Then identify the failure that matters to the decision maker. The concern might be signing with a prohibited supplier rather than losing the search screen. This combines business understanding with technology understanding. A dependency map can reveal a small component controlling a major decision and prevents effort being allocated simply by server size or the application’s prominence.
Apply declared prioritization criteria and then review the result with professional judgment. Suppose impact, speed of harm and service change are separate considerations. Payment processing has a large impact that may emerge within minutes. An archive is less urgent operationally but holds evidence for an active dispute. That dispute can justify testing the archive despite a lower numerical ranking. Record why a ranking changed, who authorized the change and which information supported it. Priorities become open to meaningful challenge when readers can see the basis for promoting a subject instead of receiving an unexplained composite score.
Negotiate scope while preserving engagement objectivity. Managers can explain the effort of extracting records and propose practical interview times. The audit committee can set priorities within its authority. Removing a high-risk path merely because it is uncomfortable requires explicit treatment. Obtain the reason, assess the effect of exclusion on the objective and escalate the restriction to the appropriate authority. The engagement objective may be narrowed when the original one cannot be achieved, provided the report clearly reflects the revised objective. Agreement on a comfortable title does not make missing evidence sufficient for broad assurance.
State boundaries by period, path and organizational unit. “Review access” is broader than “test finance employees’ access grants during the third quarter in the new application.” The latter reveals exclusions such as local accounts, service identities and earlier periods. If fieldwork reveals that a service identity can grant the same privileges, reassess scope. Add a focused test or document a significant limitation. The decision follows the path’s effect on the objective rather than its account count. One powerful service identity may alter the conclusion of an engagement covering hundreds of ordinary users.
In an exam item, read the requested action before comparing alternatives. “First” asks for a step that establishes the basis for later decisions. Discovering real data in an environment labeled experimental calls for risk reassessment before expanding a sample built on the old assumption. “Best conclusion” asks for wording supported by current evidence. Requesting another test may be professionally sensible yet answer a next-step question instead of a conclusion question. Separate facts, objective and constraints, then evaluate each alternative against them. This helps solve the actual item; it is not a shortcut based on option length.
Two team members may disagree about impact. One sees payment interruption as immediate loss; another assumes a two-hour delay can be absorbed. Separate disagreements about facts from disagreements in judgment. Ask the process owner about transaction volume, an available manual alternative and authority to use it. If an alternative exists, test its capacity before lowering priority. When impact remains uncertain, retain a range rather than a definitive number. The discussion should identify evidence that resolves the disagreement. Averaging two scores without understanding their assumptions can conceal that one reviewer imagined an alternative that does not exist.
Connect engagement days to identifiable steps. Two days for a service may include understanding the path, reconciling the population, testing and reviewing results. If understanding consumes most of the allowance, reassess the remaining work before promising completion. Reserving review and communication time matters because a late evidence limitation can prevent a useful report. Identify dependencies between steps: population reconciliation precedes sampling, and criteria confirmation precedes exception classification. This exposes when one delayed file blocks several procedures, allowing early contact with its owner or resequencing work that is still feasible.
Keep a revision record alongside the plan. A row can state the date, new information, decision and effect on evidence and opinion. It explains changes to the board and helps a member joining after fieldwork began. Distinguish continuing an existing procedure from expanding the engagement to a new objective. The first may require rescheduling; the second may require different authorization and resources. At completion, compare performed procedures with declared objectives and attach either a result or an explicit limitation to each objective. Planning is complete when readers can follow that path from the board’s need to the final conclusion.
Lesson summary
Test the plan by asking whether another reviewer could understand why one system was included and another was not, and what a concluding report could legitimately say. A ranking without criteria, period, or access limits is not a completed engagement plan. The conclusion will be bounded by the syste