,

How to Prepare for an Internal Audit: Step-by-Step Guide

Most weak audit reports were lost before fieldwork started. The objectives were copied from last year’s memo, the scope was whatever the process owner suggested at a thirty-minute kickoff, the document request went out on day one of fieldwork and came back on day nine, and the team spent the first week learning the process instead of testing it. By the time anyone knew which risks mattered, the budget was half gone and the sample sizes had been trimmed to fit. Preparation is where an engagement’s findings are decided: a team that arrives on day one with a risk assessment it can defend, a work program tied to those risks, a request list that came back a week ago, and a sponsor who has already agreed the scope will find what is there. A team that arrives to “get an understanding” will find what it is shown.

This guide is the preparation sequence for the internal audit department, from the day the engagement is confirmed on the annual plan to the morning fieldwork starts. It gives you a T-30 to T-0 timeline with owners and outputs, an engagement risk assessment worked on a real-shaped example, the information request list with 25 line items and due dates, a kickoff agenda and script, a staffing and budget table, a playbook for scope pushback, a table of the questions auditees ask and what to say, and the common failures. It was rewritten in September 2026 to reflect the Global Internal Audit Standards, which put engagement planning under Standards 13.1 to 13.6, and it pairs with the planning memo template and the work program guide, which hold the documents this sequence produces. If you are the one being audited, the companion guide on what to expect during an internal audit is written for you.

In this guide

What preparation decides, and what the Standards require

Preparation decides four things that cannot be recovered later. It decides which risks the engagement is about, which is the difference between a report on the risks that matter and a report on the controls that were easy to test. It decides the population and period, which sets what any sample can say. It decides the budget the team actually has, because hours spent learning the process during fieldwork are hours the testing no longer has. And it decides the relationship with the auditee, which is formed in the first two conversations and rarely changes afterward. A planning memo written on day three of fieldwork can be tidied into something that looks like planning; it cannot recover the sample that was cut, the location that was never visited, or the sponsor who learned the scope from the draft report.

The Global Internal Audit Standards make preparation a set of explicit requirements rather than good habits. Standard 13.1 requires engagement communication with management to be planned; 13.2 requires an engagement risk assessment that considers the objectives of the activity under review, its risks, and its controls; 13.3 requires objectives and scope that are documented and that reflect that assessment; 13.4 requires the evaluation criteria to be set and, where management’s criteria are inadequate, for internal audit to identify suitable ones; 13.5 requires the resources to be appropriate to the objectives and scope; and 13.6 requires a documented work program approved before the work starts. The Domain V guide walks through each. Every item in the timeline below maps to one of those six standards, which is also how an external quality assessor will read your planning file.

One framing rule before the sequence. Preparation is not a longer kickoff. It is a series of decisions, each with a written output, made in an order where each output feeds the next: the risk assessment feeds the objectives, the objectives feed the scope, the scope feeds the work program, the work program feeds the request list and the staffing, and all of them feed the kickoff. Teams that run the sequence backward, kickoff first and risk assessment last, produce memos that describe what they were already going to do.

The T-30 to T-0 timeline

Thirty working days is the right runway for a standard engagement of 300 to 500 hours; a two-week review can compress the same sequence into ten days, and a complex multi-site engagement needs forty-five. The days are working days before the first day of fieldwork (T-0). Owners are the roles in a typical function: the engagement lead (senior or manager), the audit manager who reviews, the CAE who approves, and the staff auditors who execute. The output column is what must exist before the next step starts, and it is the column an assessor will check.

DayTaskOwnerOutputStandard
T-30Confirm the engagement against the annual plan: entity, process, why it was planned, hours, and the risk-assessment score that put it thereAudit managerEngagement charter line: one paragraph with the plan rationale and the budget13.5
T-29Notify the executive sponsor and the process owner in writing; request the first meetingEngagement leadNotification email with the planned window and the names of the team13.1
T-28 to T-25Desk research: prior reports and open issues, external audit and regulatory findings, policies and procedures, org charts, system inventory, volumes and values, prior-year narratives and RCMs, incident logs, management informationEngagement lead and staffBackground pack, no more than ten pages, with a “what changed since last time” section13.2
T-24Sponsor meeting: the sponsor’s concerns, recent changes, known problems, what a useful report would answerEngagement lead, audit managerMeeting note; three to five sponsor concerns recorded verbatim13.1, 13.2
T-23 to T-21Process owner meeting and a first process overview; identify systems, locations, key people, and reportsEngagement leadDraft process map at the level of major steps; list of interviews and walkthroughs needed13.2
T-20 to T-18Engagement risk assessment: risks scored on likelihood and impact, existing controls noted, residual view, fraud considerationsEngagement lead; challenged by audit managerScored risk table (see below) with the rationale for each score13.2
T-17Draft objectives, scope, exclusions, period, locations; select criteriaEngagement leadDraft planning memo sections 1 to 413.3, 13.4
T-16 to T-13Draft the work program: procedures per risk, populations, sample sizes, analytics, evidence expected, hours per procedureEngagement lead with staffDraft work program with hours that foot to the budget13.6
T-14Data request to IT or the data owner for full populations (the request that takes longest)Engagement leadData request with fields, period, and format; delivery date agreed13.6
T-12Information request list (PBC) issued to the process owner with due dates staggered from T-7 to T-2Engagement leadPBC list, numbered, with owners and due dates13.1
T-11Planning memo and work program reviewed and approvedAudit manager; CAE for high-risk engagementsApproved memo; approved work program; budget confirmed13.3 to 13.6
T-10Staffing confirmed; specialist or co-source engaged; access requests submitted (systems, buildings, shared drives)Audit managerStaffing sheet; access tickets with reference numbers13.5
T-9Kickoff meeting with sponsor, process owner, and key staffEngagement leadKickoff note; scope confirmed or the disagreement recorded13.1
T-8 to T-3PBC items arrive and are logged; data populations received and completeness-checked; walkthroughs scheduled with performersStaff; engagement leadPBC tracker with received dates; population completeness check documented13.6
T-5Team briefing: each auditor can explain every procedure they own, its risk, its population, and its evidenceEngagement leadBriefing note; procedure ownership confirmed on the work program13.5
T-2Chase outstanding PBC items; escalate to the sponsor anything more than three days lateEngagement leadOutstanding-items email to the process owner, copied to the sponsor if needed13.1
T-1Final readiness check: access works, data loads, interview calendar full for week one, workpaper file structure createdEngagement leadReadiness checklist signed off13.5
T-0Fieldwork starts with testing, not learningTeamFirst walkthrough at 9:00 on day one

Two dates on that table carry most of the risk. The data request at T-14 is the one that slips: full-population extracts from an ERP take a data owner a week when nothing goes wrong, and a second week when the first extract arrives without the fields you asked for. Send it before the work program is final, from the draft, because the population you need is knowable before the sample size is. The PBC list at T-12 is the other; a request list issued at T-12 with staggered due dates arrives before fieldwork, and one issued at T-0 arrives in week two and turns week one into waiting.

The engagement risk assessment, worked

The annual risk assessment decided that the process deserved an engagement; the engagement risk assessment decides what the engagement is about. It is a different exercise at a different altitude: the annual assessment scores auditable entities, the engagement assessment scores the specific things that could go wrong inside one of them, and the scores drive which procedures get hours. Done well it takes two days and a page and a half; done badly it is a list of generic risks copied from a framework, each rated “medium,” which tells the work program nothing.

The method has five inputs: the process objectives (what the process is supposed to achieve, in the owner’s words); what could go wrong against each objective, stated as an event rather than a category (“counts are adjusted without approval,” not “inventory risk”); likelihood and impact scored on a scale the function uses everywhere, with the rationale written next to the score; the controls management says exist, noted rather than assessed, because assessing them is fieldwork; and a residual view that says where the engagement’s hours should go. Fraud gets its own line for every process that handles assets or numbers people are paid on; the fraud red flags library is the prompt list. The worked example below is the FY27 warehouse inventory engagement at MidState Beverage, the site’s running example: a three-state distributor with 12 depots, 300 routes, a 2013 ERP, two acquired distributors not yet integrated, and a six-person audit function that planned 380 hours for the engagement.

#Risk (what could go wrong)Objective affectedLikelihood (1–5) and whyImpact (1–5) and whyControls management describesResidual view and hours
R1Physical inventory differs from the perpetual record because cycle counts are incomplete or adjustments are posted without investigationAccurate inventory records; correct cost of sales4 — the ERP cycle-count module was disabled at four depots after the FY24 upgrade; counts are done on spreadsheets there4 — $18.6M inventory at cost; a 2% error is $370K, above the external auditor’s materiality for the segmentWeekly cycle counts; adjustments over $2,000 approved by the depot manager; month-end reconciliation by HQHigh. Full-population adjustment analytics plus counts at three depots: 140 hours
R2Product is diverted from depots (theft by staff or drivers) and covered by adjustments, write-offs, or breakage claimsSafeguarding of assets3 — FY26 breakage write-offs rose 41% at two depots with no change in volume; the FY26 route cash theft showed how weak detection is at depot level4 — loss plus the control environment signal to the boardDriver load-out sheets signed by the warehouse lead; breakage requires a photo and supervisor sign-off; monthly write-off report to the VP OperationsHigh. Write-off and breakage analytics by depot, route, and approver; load-out sheet testing: 90 hours
R3Receiving records quantities that were not delivered, or receives against the wrong PO, so payables are overstated and inventory is wrong from day oneAccurate records; valid liabilities3 — receiving at nine depots is keyed by the depot clerk, not by a separate warehouse role3 — the three-way match relies on it; overstatement rather than lossReceipt keyed against PO in the ERP; carrier delivery note filedMedium. Receipt-to-delivery-note sample of 40, weighted to the nine single-clerk depots: 50 hours
R4Slow-moving and short-dated product is not identified and written down, so inventory is carried above net realizable valueValuation3 — no aging report exists in the 2013 module; the controller runs an annual spreadsheet3 — beverages have 6- to 12-month shelf life; the FY26 year-end write-down was $212KAnnual review by the controller; quarterly depot self-report of short-dated stockMedium. Rebuild an aging from receipt dates for the full population; test the Q4 self-reports: 40 hours
R5The two acquired distributors count and value inventory on their own systems with unknown methods, and consolidation adjusts blindlyAccurate consolidated records4 — nobody at HQ has seen either count procedure3 — $2.1M combined inventoryMonthly inventory certification signed by each acquired-entity managerMedium. One-day site visit to each; walkthrough of the count and the certification: 30 hours
R6ERP inventory roles allow the same user to receive, adjust, and approve adjustmentsSafeguarding; accurate records3 — the FY27 user access engagement found role conflicts in route accounting; inventory roles were out of its scope3 — enabler of R1 and R2 rather than a loss in itselfRole design reviewed at go-live in 2013Medium. Role-conflict analysis for all 84 inventory users: 20 hours
Reserve for what the walkthroughs find, reporting, and reviewPlanning 30, reporting and review 60, contingency 20: 110 hours; total 380

Three things about that table are the method. The likelihood rationale cites a fact each time: a disabled module, a 41 percent rise, a role that was never re-examined. A score without a fact next to it is an opinion, and the audit manager’s challenge at T-18 should be to ask for the fact behind every 4. The impact rationale names a number and what it is compared with, so that “material” means something. And the hours column allocates the budget before the work program is written, which is the only way the work program ends up reflecting the risks rather than the habits of whoever wrote last year’s. When R1 and R2 take 230 of 380 hours, the memo can say why, and the sponsor can disagree with a decision rather than discover one.

Objectives, scope, and the planning memo

The objective is a sentence that says what the engagement will conclude on, written so that the report’s conclusion can be read against it. “To assess whether controls over depot inventory are designed and operating effectively to ensure that inventory records are accurate, product is safeguarded, and inventory is valued at the lower of cost and net realizable value” is an objective; “to review inventory” is a subject. Each objective should trace to the risks in the assessment, and each risk in the assessment should be covered by an objective or explicitly excluded, with the reason.

Scope is the set of boundaries: the process from where to where, the period, the locations and entities, the systems, and the exclusions. The exclusions are the part most memos skip and the part that causes the most trouble later, because an exclusion stated in the memo is a decision and one discovered at the closing meeting is a gap. Write them with the reason: the two acquired distributors are in scope for a walkthrough only because they are on separate systems and the integration engagement covers them in Q3; pricing and margin are out of scope because they belong to the revenue engagement; the period is the twelve months to 30 June because year-end counts fall outside it. Criteria come next: the policies, procedures, standards, and control-design principles the process will be measured against, named by document and version, so that a finding can quote its criterion. Where management has no adequate criteria for a risk, Standard 13.4 requires you to identify suitable ones and say so in the memo, which is where COSO’s 17 principles or an industry practice standard come in.

The memo assembles those decisions into the document the sponsor signs and the assessor reads. The planning memo template gives the full structure with model language; the checklist below is the review test before it goes for approval at T-11.

Memo elementThe test at reviewThe usual failure
BackgroundQuantified: volumes, values, locations, systems, what changed since the last engagementTwo pages of process description with no numbers
Why nowStates the plan rationale and the sponsor’s concerns, verbatim where possible“Per the annual audit plan”
Risk assessmentEvery score has a fact; the hours follow the scoresAll risks “medium”; hours split evenly
ObjectivesOne sentence each, conclusion-shaped, traceable to risksObjectives are activities (“review,” “assess,” “evaluate”)
Scope and exclusionsBoundaries with reasons; period and locations stated; exclusions listedNo exclusions; period not stated
CriteriaDocuments named with versions; gaps filled with named external criteria“Company policies and best practice”
ApproachProcedures summarized by risk with populations and sample sizes“Interviews, walkthroughs, and testing”
Resources and timingNames, hours per procedure, milestones, specialist needsTotal hours only; no milestones
Communication planWho gets what, when: status updates, preliminary findings, draft, finalAbsent; the sponsor first hears of findings in the draft
ApprovalAudit manager and, for high-risk engagements, the CAE, dated before T-0Approved after fieldwork started

The work program and the information request list

The work program converts each risk into procedures with a population, a sample or analytic, the evidence expected, and the hours; the work program guide covers the procedure-writing craft and the sample-size guide the sizing. Two preparation rules matter here. First, write the data request from the draft work program at T-14, not from the final one at T-11, because populations are knowable before sample sizes are and the extract takes longest. Second, make every procedure’s owner on the team able to explain it at the T-5 briefing: the risk it addresses, the population, the evidence, and what an exception would look like. An auditor who cannot explain the procedure will test what is convenient.

The information request list is the document that decides whether week one is testing or waiting. It goes out at T-12, numbered, with a named owner and a due date per item, staggered so that the items needed for walkthroughs arrive first. Ask for documents by their real names, as the process owner meeting revealed them, and ask for system reports by report name and parameters. Every item should be traceable to a procedure; an item nobody can trace is a fishing request and will be resented. The list below is the one MidState’s team issued for the inventory engagement; the due dates are relative to T-0.

#Item requestedOwnerDueProcedure it serves
1Inventory policy and the depot count procedure, current versions with revision historyDirector of Route AccountingT-8Criteria; all
2Org charts for warehouse and depot operations, all 12 depots, with vacanciesVP Operations officeT-8Interviews; R6
3Inventory balance by depot and SKU at 30 June, at cost, from the ERP (report INV-104)ControllerT-8Populations; R1, R4
4All inventory adjustment transactions, 1 July FY26 to 30 June FY27, with user, approver, reason code, quantity, and value (report INV-221)IT data ownerT-7R1 analytics
5Cycle count schedules and completed count sheets for the period, all depots, including the spreadsheet counts at the four depots without the moduleDepot managers via Route AccountingT-7R1
6Breakage and write-off log with photos and approvals, by depot, for the periodDirector of Route AccountingT-7R2 analytics
7Driver load-out sheets for a week to be named at T-3 (we will select the week)Depot managersT-3R2 testing
8Goods receipt transactions for the period with user, PO, and delivery-note reference (report PUR-310)IT data ownerT-7R3 population
9Carrier delivery notes for the 40 receipts we select (list issued T-4)Depot clerksT-1R3 sample
10Receipt dates by SKU and depot to rebuild an aging (report INV-118 or an extract with first-receipt date)IT data ownerT-6R4
11The controller’s FY26 year-end short-dated and slow-moving review spreadsheet and the write-down entryControllerT-6R4
12Q1 to Q4 depot short-dated self-reportsRoute AccountingT-6R4
13Monthly inventory certifications from both acquired distributors, all months in the period, with the count procedure each usesIntegration leadT-5R5
14ERP inventory role definitions and user-role assignments for all inventory users (84 expected)IT securityT-6R6
15Month-end depot inventory reconciliations, HQ to depot, for the period, with preparer and reviewerControllerT-5R1
16Prior-year external audit management letter points on inventory and their statusControllerT-8Background
17Open internal audit issues on inventory from the issue log, with current statusInternal audit (own)T-8Background
18Inventory insurance claims and police reports for the period, if anyRisk managerT-5R2
19Depot layouts and access control lists (badge access to warehouse areas)FacilitiesT-4R2 site visits
20Cycle-count module configuration and the FY24 upgrade change record that disabled it at four depotsITT-5R1 cause
21Management information pack pages on inventory sent to the executive team for the last six monthsFP&AT-5Reporting controls
22Vendor delivery schedules for the three depots to be visited (for planning counts)PurchasingT-3Site visits
23Names and availability of depot managers and warehouse leads at the three visit depots for the week of T+5VP Operations officeT-4Logistics
24ERP read-only access for the two auditors named (ticket reference)IT securityT-6All
25Warehouse safety induction requirements and PPE for site visitsFacilitiesT-4Logistics

Log every item as it arrives, with the date, and check the populations for completeness on receipt rather than in week two: row counts against a control total, value against the ledger, period boundaries against the request. An extract that is short by a month, discovered on day eight, costs the engagement a week; discovered at T-6 it costs an email. The IPE testing guide covers what “complete and accurate” means for a system report you intend to rely on.

The kickoff meeting: agenda and script

The kickoff is not where scope is discovered; by T-9 the memo is approved and the request list is out. It is where the sponsor and the process owner hear the scope from you, in one room, and either confirm it or disagree on the record. Forty-five minutes is enough. Invite the sponsor, the process owner, the people who will host the walkthroughs, and the person who owns the data extracts; do not invite the whole department, because a kickoff with twenty people produces no disagreement and no information. Send the agenda two days ahead with the scope paragraph from the memo attached, so that nobody is reading it for the first time in the meeting.

MinutesItemWhat you are trying to get
0–5Why this engagement, why now, in the plan’s words and the sponsor’sShared understanding that the engagement is about the sponsor’s risks, not audit’s habits
5–15Objectives, scope, period, locations, exclusions, criteriaExplicit confirmation or explicit disagreement, recorded either way
15–25Approach: what will be tested, which populations, which sites, what analytics will run on their dataNo surprises later; the owner knows a full-population adjustment analysis is coming
25–32Timeline: fieldwork dates, status updates, preliminary findings meeting, draft and final report dates, who gets eachAgreement on the communication plan from Standard 13.1
32–38Requests: the PBC status, the outstanding items, the walkthrough schedule, accessNamed owners and dates for anything late
38–43How findings are handled: facts cleared with performers, ratings against the published scale, management responses in the report, tracking in the issue logThe owner knows the rules before the first finding exists
43–45Questions and the one thing each attendee wants the engagement to answerThree sentences that go into the memo’s sponsor-concerns section if they are new

Opening. “This engagement is on the FY27 plan because inventory is $18.6 million at cost across twelve depots, four of them count on spreadsheets since the upgrade, and breakage write-offs at two depots rose 41 percent last year without a volume change. Those are the three things the board’s risk assessment flagged. The objective is to conclude on whether the controls over depot inventory are designed and operating to keep the records accurate, safeguard the product, and value it correctly. We are not looking at pricing or margin; that is the revenue engagement in November.”

Scope confirmation. “The period is the twelve months to 30 June. We will visit Dayton, Columbus, and Fort Wayne in the week of the 14th; the other nine depots are covered by the analytics on the full adjustment and write-off populations, and by the reconciliation testing. The two acquired distributors get a one-day walkthrough each, not testing, because the integration engagement covers them in Q3. Is there anything in that scope you would draw differently, and is there a risk you would add?”

Approach. “Three things to know in advance. We will run analytics on every adjustment posted in the period, so if a depot has a pattern, we will see it before we visit. We will select the load-out week ourselves and tell you three days ahead. And we will count a sample at each visited depot on the day we arrive, unannounced as to which SKUs, so please do not prepare a count for us.”

How findings work. “When we find something, the first conversation is with the person who performs the step, about the facts, and the second is with you, before anything is written as a finding. Ratings follow the scale in the appendix, which you have seen. Your response goes into the report verbatim, and where we disagree, both positions are in the report. Nothing reaches the audit committee that you have not seen first.”

Closing. “Before we finish: what is the one thing you would want this engagement to tell you that you cannot find out yourself?”

The last question produces the best material in the engagement about a third of the time. Depot managers have said “whether the breakage rule is being gamed at Toledo” and “whether my counts are worse than everyone else’s or the module is wrong,” both of which became procedures. Record the answers verbatim in the kickoff note and add any new risk to the memo with a note that it came from the kickoff; an assessor reading the file should see that the risk assessment was still alive at T-9.

Staffing, budget, access, and logistics

Standard 13.5 asks for resources appropriate to the objectives and scope, and the honest test of that is whether the hours in the work program foot to the budget and the people on the engagement can do the procedures assigned to them. Match skills to risks rather than availability: the auditor who has never run a population analytic should not own R1, and the one who has never been in a warehouse should not lead the site counts. Where the function lacks a skill, the decision to bring in a specialist or a co-source partner is made at T-10, not in week two; the co-sourcing guide covers how to scope a specialist’s piece so it integrates with your file rather than arriving as a separate report.

RolePersonProcedures ownedHoursSkill check
Engagement lead (senior)Senior auditor ARisk assessment, memo, R1 analytics design, R5 site visits, reporting150Has run full-population analytics; led two prior inventory engagements
Staff auditorAuditor BR1 count testing at three depots, R3 receipt sample, R4 aging rebuild120Comfortable in the ERP; needs a half-day on the aging query with A
Staff auditorAuditor CR2 write-off analytics and load-out testing, R6 role analysis70Ran the FY27 route-accounting SoD analysis; reuse the query
Audit managerManager DReview at T-11, T-5, week two, and draft; sponsor relationship30
SpecialistNone needed0Inventory valuation is within the team’s competence; no co-source
Contingency and adminTravel, delays, the walkthrough surprise10
Total380Foots to the plan

Access is the logistics item that fails silently. Submit system access requests at T-10 with the specific roles (read-only inventory, read-only purchasing, report execution), because a generic “auditor access” request comes back as whatever IT last gave an auditor, and confirm on T-1 that each login works by running the report you will need on day one. Physical access to depots needs badges, safety inductions, and PPE, all of which have lead times, and a count at a depot needs the depot manager to know the day but not the SKUs. Book the walkthrough performers at their desks, in transaction order, for the first three days of fieldwork; the walkthrough guide explains why you walk with the clerk and not the owner. Create the workpaper file structure before T-0 from the work program’s numbering, so that every procedure has a home on day one; the workpaper best practices guide gives the indexing convention.

What auditees will ask, and how to answer

The questions below arrive in the sponsor meeting, the kickoff, and the corridor. Each has a good answer and a bad one, and the bad ones are usually the reflexive ones. Answer plainly, in the first person, and never with “that’s our methodology,” which is the phrase that ends cooperation.

What they askWhat they meanWhat to say
“Why us, why now?”Am I being singled out?The plan rationale in one sentence with the numbers, and the fact that the process was last audited three years ago. Never “it came up in the rotation.”
“How long will this take my team?”What will it cost me?A real estimate: the request list is 25 items, the walkthroughs are six people for 45 minutes each, the site visit is two days with one host. Then keep to it.
“Can you use the counts we already do?”Please do not duplicate our work.Yes, as evidence of the control operating, and we will also count ourselves, because the objective includes whether the records are right, which only an independent count answers.
“We already know the module is off at four depots. Do you have to write that up?”Can known problems be exempt?A known problem with an approved remediation plan is reported as such, with the plan; a known problem with no plan is a finding. Tell me which one this is.
“Who sees the report?”How much exposure is this?The distribution list from the memo, by name, and the fact that they will see your response next to every finding.
“What if we disagree with a finding?”Do I have any say?Facts are cleared with your people before anything is written; ratings follow the published scale; if we still disagree, both positions go in the report and the audit committee decides.
“Can we see the findings before the closing meeting?”I do not want to be ambushed.Yes: preliminary findings weekly during fieldwork, in writing, and the full list before the closing meeting. That is the communication plan in the memo.
“Why do you need the whole population? Just take a sample.”That extract is a lot of work for IT.Because a sample of 40 adjustments tells us about 40 adjustments; the population tells us which depot has a pattern, and that is the question the board asked. The extract is one report with four parameters; we have named it.
“Is this about the theft last year?”Is this an investigation?No. The FY26 loss is background, and this engagement tests whether the controls would catch the next one. If we find evidence of a specific loss, it goes to the General Counsel under the charter, not into the report as a finding about a person.
“Can we move it to next quarter? Year-end is busy.”Please go away.The timing was set in the plan the audit committee approved; I can move the site visits by a week to avoid the count week, and I will tell the CAE you asked. A change of quarter needs the sponsor and the CAE.

When management pushes back on scope

Pushback is information. A sponsor who wants a location out of scope, a period shortened, or a risk dropped is telling you something about that location, period, or risk, and the first response is to ask why in a way that lets them answer honestly. Some reasons are good: the depot is closing in October, the system is being replaced in Q3, the risk belongs to another engagement already on the plan. Some are not, and the difference is whether the reason is about the organization’s interests or the sponsor’s comfort. Either way, the decision belongs to internal audit under its charter, the reason and the decision go in the memo, and the sponsor sees both. What never happens is a scope change made verbally in a kickoff and discovered by the CAE in the draft report.

The pushbackReasonable responseThe line you hold
“Leave the acquired distributors out; integration is handling it”Confirm the integration engagement’s scope and timing in writing; reduce to a walkthrough with a stated hand-offA risk on the annual assessment is covered by someone, somewhere, this year, in writing
“Don’t visit Toledo; the manager is new”Ask what the previous manager left; a new manager is a reason to visit, not to skipLocation selection is audit’s, based on risk indicators, and the indicators say Toledo
“Use a three-month period; the earlier data is messy”Ask what “messy” means; a system conversion in the period is a scope note, not an exclusionThe period covers a full cycle unless a documented event makes earlier data unreliable, in which case the unreliability is itself reported
“Don’t run analytics on adjustments; it will look like an investigation”Explain the difference: a population analytic is a test of a control’s operation, reported at process level; a named-person pattern goes to counsel under the charterFull-population testing of the risk rated highest in the assessment is not negotiable
“We’ll fix the module before you start, so drop R1”Good; test the fix as an implemented action and keep the period’s exposure in scopeRemediation during the engagement is reported as remediation, not as absence of a problem
“The CFO wants revenue in scope too”Explain the budget arithmetic; offer a scoped addition with hours from contingency or a separate engagementScope creep without hours produces two half-engagements; the CAE decides plan changes

When a pushback is accepted, write the exclusion into the memo with the reason and the name of the person who requested it, and where it removes coverage of a risk on the annual assessment, tell the CAE at the time, because the annual plan’s coverage statement to the audit committee just changed. When it is refused, record that too. The memo’s exclusion section is the audit trail of every one of these conversations, and it is the section an external assessor reads to see whether the function’s independence is real.

Common preparation failures

FailureWhat it looks likeWhy it mattersFix
Last year’s memo with the dates changedObjectives, scope, and risks identical to the prior engagement; no “what changed” sectionThe engagement tests the process that existed three years ago; new systems and new people are missedDesk research at T-28 produces a written “what changed” list before the risk assessment starts
Generic risk assessmentTen risks from a framework, all rated medium, no factsHours are spread evenly; the risks that matter get the same 30 hours as the ones that do notA fact behind every score; hours allocated by score before the work program is written
Request list issued at T-0Week one is spent waiting; the team “gets an understanding” insteadA quarter of the budget produces no evidencePBC at T-12 with staggered due dates; data request at T-14
Populations not checked on receiptThe adjustment extract is missing two depots; nobody notices until the analytics look odd in week twoA week lost, and a sample drawn from an incomplete population is worthlessRow counts and value totals against a control total on the day the extract arrives
Kickoff as discoveryScope is worked out in the meeting; the memo is written afterward to matchThe sponsor sets the scope; the memo is a record, not a decisionMemo approved at T-11; the kickoff confirms it
Staffing by availabilityThe analytic is owned by the auditor who is free, not the one who can do itThe highest-rated risk gets the weakest testSkills-to-risks matching at T-10; a skill gap is a co-source decision, not a hope
Verbal scope changesA location dropped in a corridor conversation; the CAE learns from the draftCoverage reported to the committee is wrong; independence looks negotiableEvery scope change written into the memo with reason and requester; CAE told at the time
No communication planThe sponsor first hears of findings in the draft reportThe closing meeting becomes a negotiation about surprises rather than a discussion of factsWeekly written status; preliminary findings in writing; the plan in the memo and the kickoff
Access assumedERP login fails on day one; badge for the depot not issuedDays lost, and the auditee sees a team that was not readyAccess requests at T-10; every login tested at T-1

Adapting the sequence: small functions, co-sourcing, remote, regulated

Small functions and one-person shops

A one- or two-person function runs the same sequence with the same outputs and fewer meetings. The risk assessment is still a scored table, but it is a page; the memo is still approved before fieldwork, by the CAE who wrote it, with the audit committee chair copied for high-risk engagements as a substitute for a second reviewer; the request list is still numbered and dated. What a small function cannot do is compress the runway. The data request still takes IT a week, the PBC still takes the owner ten days, and a solo auditor who issues them at T-0 has nothing to do for two weeks. The small-company internal audit guide covers how a new function builds the templates once so that preparation is a fill-in rather than a rewrite.

Co-sourced and specialist engagements

When a co-source partner or a specialist owns part of the work, preparation includes a scoping document for their piece that names the procedures, the evidence standard, the workpaper format, the review process, and the dates, and it is agreed at T-10. The partner joins the T-5 briefing and the kickoff. What goes wrong is the partner running their own methodology in parallel and delivering a separate report in week four that has to be reconciled to yours; the fix is one work program, one file, one report, and the partner’s procedures numbered inside it.

Remote engagements

Remote preparation front-loads evidence. Everything that would be picked up in a site visit, the shared-drive spreadsheet, the count sheets in a binder, the badge list on the wall, has to be requested by name, so the process-owner meeting at T-23 becomes an inventory of where records physically live. Walkthroughs are scheduled with screen sharing of the live application and recorded with consent. Counts, cash, and custody steps are not remote; if the engagement’s risks include them, a site visit goes in the plan or the objective is narrowed and the memo says why.

Regulated entities and ICFR support

In a bank or insurer the preparation sequence gains a step: the regulatory expectations for the area, from examination manuals and outstanding supervisory findings, become criteria alongside management’s policies, and the memo maps them. The financial services guide sets out what examiners expect the third line’s planning file to show. For engagements that support the SOX program, the population and period decisions have to line up with management’s assessment timeline and the external auditor’s reliance plan, which means the T-24 sponsor meeting includes the SOX lead; the SOX scoping guide covers the mechanics.

Where preparation ends

Preparation ends when the first walkthrough starts, and the measure of whether it worked is the week-one status report. A prepared engagement’s first status report lists procedures completed and preliminary observations. An unprepared one lists documents received. From there the sequence continues into fieldwork with the walkthrough template, the evidence standards, the rating scale the kickoff promised, and the report whose objectives are the ones written at T-17. Every guide on the site is indexed on the Topics page.

Related guides

Comments

Leave a Reply

Discover more from internalauditguide.com

Subscribe now to keep reading and get access to the full archive.

Continue reading