,

Build vs Buy for Audit Software: Homegrown SharePoint and Power Apps Against a Platform

Most internal audit functions did not choose their first system. They inherited a shared drive, then someone moved the issue log into a SharePoint list because email attachments kept getting overwritten, then a manager who knew Power Apps built a form on top of it so auditees could submit action plans without calling IT. Nobody signed a contract or ran an RFP. Three years later that stack holds every workpaper, every finding and every management response the function owns, and the question is no longer whether to build something, because it already exists. The real question is whether to keep extending it, freeze it and buy a platform instead, or run both at once.

This guide compares what internal audit teams actually build on SharePoint, Power Apps, Power Automate and their neighbors against buying a packaged internal audit platform: where the build genuinely wins, where it quietly becomes a liability, a decision matrix, a total-cost-of-ownership table built on real Microsoft prices, the hybrid pattern most functions land on, and a worked example. It is part of the site’s independent buyer’s guide to internal audit software and follows the evidence rules in how we review audit software. It assumes you already know the case for leaving a plain spreadsheet behind; the site’s guide to audit software vs Excel and SharePoint covers that narrower question. This one is about what happens after you decide to build something more than a spreadsheet: do it yourself on tools you already own, or buy.

How to read this guide

What this guide is for. Working out whether your internal audit function should keep building its own tracking system on Microsoft’s low-code tools or buy a packaged platform, with a cost comparison and a decision matrix to run against your own numbers, not ours.

Evidence. Research-based: Microsoft’s published Power Platform and Microsoft 365 prices, Vendr’s buyer data for packaged platforms, the fit ratings and price evidence from this site’s own platform reviews, and the Global Internal Audit Standards. We have not built the homegrown pattern described here for a real audit function, and we have not used the platforms named on this page hands-on.

Last verified. 27 September 2026.

In this guide

What internal audit teams actually build

“Build” rarely means one tool. In practice it is two or three stacked on top of each other, added in the order the pain showed up rather than designed from a blank page. The table below is the pattern this site sees most often, roughly in the order a homegrown stack grows.

ToolWhat it actually isWhere it fits in a homegrown stack
SharePoint lists and librariesA structured list with columns, views and alerts, plus document libraries with version history, built into the Microsoft 365 subscription most organizations already pay forUsually the first thing built: the issue log, the audit universe, or workpaper storage
Power Apps (canvas or model-driven)A custom form and screen layer on top of SharePoint or Dataverse, built by arranging controls rather than writing codeTurns a list into something that looks like software: guided data entry, role-based views, a mobile layout for auditees submitting action plans
Power AutomateWorkflow automation: routing an item for approval, reminding an owner when an action plan is overdue, posting to Teams when a finding is loggedThe escalation and notification logic a plain list cannot do by itself
Airtable or SmartsheetStandalone spreadsheet-database hybrids, licensed per seat outside the Microsoft tenantChosen by teams outside a heavy Microsoft shop, or that want a faster interface than Power Apps without involving IT
JiraIssue and workflow tracking built for software teams, repurposed with custom fields and boardsCommon where the CAE reports into technology, or the function reuses tooling the second line already runs
Access, or a Python or spreadsheet scriptA single-file desktop database, or scripts that run analytic tests and assemble a reportAccess is a legacy pattern; Python and spreadsheet scripts persist even after everything else is replaced, because they run analytics a packaged platform usually does not

None of these tools was built for internal audit, which is both the appeal and the risk. A SharePoint list will hold anything you put in it, which is exactly why it will hold a poorly designed audit trail just as happily as a good one. The site’s guide to types of internal audit software sets out what a purpose-built internal audit platform, or audit management software, is actually built to do end to end, from the audit universe through issues and follow-up, which is the yardstick the rest of this guide measures a homegrown build against. It is a different yardstick again from compliance automation: if the tracker you are weighing against a purchase is really collecting evidence for a SOC 2 or ISO 27001 certification rather than running internal audit engagements, that is a different comparison, covered in internal audit software vs compliance automation. The site’s guide to tracking internal audit issues from basic Excel to AI-enabled practice traces this same progression for the issue log specifically, and a Power App built to mirror the site’s internal audit work program structure will hold up better than one built to whatever fields seemed useful at the time.

Where a homegrown build wins

A homegrown build wins on four things, and they are real advantages, not consolation prizes for not buying.

  • Cost at small scale. A first system for a small team, one to five auditors, can plausibly run on a couple of SharePoint lists and a Power App for a fraction of what a packaged platform’s own minimum practical spend would consume in licensing alone; the total cost of ownership section below works the numbers.
  • Exact fit. Nobody sells a form that matches your organization’s real risk and control matrix, sign-off chain or committee structure out of the box. A homegrown build can mirror the actual process instead of asking the team to adapt to a vendor’s model of one.
  • No new vendor to review. The data stays inside a Microsoft tenant the organization has already put through its own security and data-residency review for email and files. Adding a new SaaS platform means running that review again, which the site’s audit software due diligence guide covers.
  • Analytics you were going to build anyway. Even functions on a packaged platform typically still run Python, Power Query, Power BI or Alteryx for full-population testing, because most audit management platforms do not ship a scripted testing engine. Homegrown analytics is not a fallback for skipping a platform; it is standard practice either way, which the hybrid pattern below covers.

The common thread is scale and specificity: the smaller the team and the more particular the process, the better a build competes. Both cut the other way as the function grows, which is the subject of the next section. For a function just getting started, the site’s guide to setting up an internal audit function in a small company covers the wider decisions this is only one part of.

Where a homegrown build becomes a liability

The same advantages invert once a function grows past a certain point, or once someone outside the audit team starts asking questions about the system rather than its findings. Six liabilities show up most often.

LiabilityWhat a homegrown build typically hasWhat a packaged platform is built to have
Key-person riskOne or two people understand the security roles, the flow error handling and why a particular field exists; changes stop happening safely once they leaveA support desk, documentation and a release cycle that does not depend on one employee’s memory
Audit trailVersion history on a list or a library, showing that something changed, not a structured trail across a full engagement from planning memo to closed issueField-level history and locked workpapers after sign-off, built as a core feature rather than an afterthought
RetentionWhatever generic retention policy IT set for the Microsoft 365 tenant, rarely written with internal audit’s own retention and legal-hold needs in mindRetention and legal-hold settings scoped to the audit function’s own records
Security reviewOften skipped entirely, because it looks like part of an existing Microsoft 365 tenant rather than a new system in its own rightTriggers a formal vendor security review the first time it is bought, producing documentation the function can point to later
QAIP and methodology evidenceEnforces only what the builder remembered to validate; a step can be skipped in the underlying list even if the Power App discourages itEnforces the workflow: an issue cannot close without evidence, a review step cannot be skipped
Examiner and audit-committee scrutiny“One of our managers built this in Power Apps” is a harder answer to defend than it sounds, even when the app works fineA named, contracted vendor with its own security posture is the answer examiners and external assessors expect to hear

None of these six kill a homegrown build on their own. Together, and especially once a bank charter, a SOX designation or an external quality assessment is on the calendar, they are why functions that start on Power Apps eventually buy something, even when the software itself never stopped working. GIAS 14.6 sets the documentation standard a workpaper has to meet regardless of what system it lives in, and 15.2 expects action plans to be monitored to completion; a platform enforces both by default, while a homegrown build only enforces what someone built it to.

The build-or-buy decision matrix

Run your own function against the seven factors below. A build usually wins on the left, a purchase on the right, and most functions find their answer split across both columns, which is exactly why the hybrid pattern two sections down exists.

FactorLeans buildLeans buy
Team size and platform minimum spendA first system for a small team, where even a mid-market platform’s minimum practical spend is disproportionate to the budgetA mid-size or larger function, where inconsistent workpaper quality across auditors is the visible problem
In-house Power Platform skillA citizen developer or a small BI or IT team already comfortable with Dataverse security roles and flow error handlingNobody owns Power Platform today, and IT will not take on a new production app
Regulatory profileNo SOX program and no external examinerA SOX-heavy public company, where the ICFR program is the center of gravity, or a bank or credit union under examiner scrutiny
Analytics intensityAn analytics-heavy team already running Python, Power BI or Alteryx for full-population testing, which most platforms would not replace anywayAnalytics is not the gap; the workflow, workpaper and issue layer is
Growth and GRC ambitionStaying a single-purpose audit trackerConsolidating audit with risk and compliance across the three lines onto one platform, the question the site’s GRC suite vs standalone audit software comparison works through
Budget shapeInternal developer time is available and is not competing with higher-priority workA software budget line is easier to get approved than diverting a scarce Power Platform developer, the case the site’s business case for audit software guide helps build
Time to a defensible systemComfortable with an iterative rollout, a first version live in weeksNeed an enforced, defensible system quickly, ahead of an external quality assessment or a new examiner relationship

If you land solidly in the buy column, the site’s vendor-neutral RFP method and demo script cover the process from requirements through a live demo. Skipping a real comparison in favor of whatever a colleague used at their last job is one of the 15 buying mistakes this site sees most often, and it cuts both ways: teams talk themselves into buying without ever pricing out what they already have the tools to build.

Total cost of ownership: a homegrown build against a mid-market platform

Neither side of this comparison has a single true cost, so treat the table below as a worked illustration with every assumption labeled, not a quote for your organization. It compares MidState Beverage, the site’s recurring fictional six-person audit function, building on the Power Platform against buying Onspring, the mid-market platform whose fit rating for a function that size is Strong across both a first system and a mid-size function, and whose Vendr-sourced price evidence is public. LogicGate, with a Vendr median of $53,784 a year, and Resolver, with a low-confidence Vendr average of about $32,471 from few deals, sit on either side of Onspring’s $33,808 median if you want a wider range; the site’s Onspring vs LogicGate vs Resolver comparison lines the three up directly, and the full Onspring review covers the product this table prices.

Cost elementHomegrown Power Platform buildOnspring (mid-market platform)
Year 1 Microsoft licensing: Power Apps Premium for 6 users at $20 a user a month, Power Automate Premium for 2 maker-admins at $15 a user a month; Microsoft 365 itself assumed already owned for email$1,800—
Year 1 build labor (assumption: about 200 hours to build the audit universe, RCM, workpaper flow, issue tracker and reporting, at a $100 to $150 an hour blended rate; no public rate card exists for this)$20,000 to $30,000—
Year 1 subscription—$33,808 (Vendr’s buyer data, viewed 26 September 2026; no separate implementation fee found in public sources)
Years 2 to 5 Microsoft licensing, per year$1,800—
Years 2 to 5 maintenance and enhancement labor, per year (assumption: about 60 hours a year at the same blended rate)$6,000 to $9,000—
Years 2 to 5 renewal, per year (assumption: flat; no public renewal-escalation figure found for Onspring)—$33,808
Five-year total, using the midpoint of the labor rangesabout $64,000about $169,000

Two things distort a straight read of that gap. First, the homegrown total assumes someone can actually deliver 200 hours of build time without it either sitting in a queue behind fieldwork or being billed at a contractor’s rate well above $150 an hour; the site’s implementing audit management software guide covers what the first 120 days of either path realistically costs in people’s time, not just license fees. Second, the gap does not price in the six liabilities from the section above: an examiner or an external quality assessor does not accept that a system was cheaper as an answer to a question about audit trail or retention. If the homegrown build would fail a security review the first time anyone looked closely at it, the true comparison is not $64,000 against $169,000; it is $64,000 plus the cost of fixing that later, which is rarely small once workpapers already live inside it.

Microsoft’s own packaging is not static either. Power Apps sold a per-app plan for years, cut from $10 to $5 a user a month for a single app in 2021; a third-party Microsoft-licensing tracker reports that Microsoft ended its sale in January 2026, leaving only a pay-as-you-go usage meter at $10 an app, alongside the unlimited per-user Premium plan the table above uses. A build plan priced against the old per-app rate would already be wrong. Unattended automation and process mining move the homegrown side’s ceiling further still: a single unattended bot runs $150 to $215 a month and the Process Mining add-on is priced at $5,000 a tenant a month, figures that erase a build’s cost advantage fast once the ambitions grow past what a per-user plan covers.

The hybrid pattern: platform of record plus homegrown analytics

Most functions that have both a real budget and a real Power Platform habit do not pick a side. They buy a platform of record for the parts of the liability table above that are hardest to defend after the fact, and keep the homegrown layer for analytics, where a platform adds little anyway.

FunctionUsually the platform of recordCommon to keep homegrown
Engagement workflow, workpapers, sign-offYes, since audit trail and retention live hereRarely; this is the highest-liability area to homegrow
Issues and follow-up, management self-reportingYes, since examiners and the audit committee expect this to be systematicSometimes a simple escalation email, layered on top
Reporting to the audit committeeYes, using export and dashboard features built for this audienceA Power BI dashboard is common even on top of a platform, pulling data through an API
Full-population testing and continuous monitoringNo; most platforms, Onspring included, do not ship a scripted testing engineYes: Python, Power Query, Power BI or a dedicated tool, covered in the site’s five data analytics tools every internal auditor should know
AI drafting and summarizationIncreasingly inside the platform: Onspring’s AI runs on Anthropic’s Claude models, LogicGate’s on named OpenAI modelsOr a Copilot Studio agent built on the homegrown side, using Microsoft 365 Copilot, the enterprise add-on at $30 a user a month; a smaller function on a Business plan would instead see the Microsoft 365 Copilot Business add-on, $21 a user a month list, $18 on promotion through December 2026
Risk and compliance for the other two linesSometimes, if consolidating GRCRarely worth building for a function that only owns audit

The pattern holds because the two sides of the table have opposite economics. A platform’s workflow, sign-off and retention features are expensive to replicate well and cheap to buy once, which is why so little of the liability table above gets rebuilt on top of a purchased platform. Analytics is the reverse: a platform’s built-in version is usually thin, Python or Power BI is a skill the team may already have, and the site’s evaluating AI in audit software guide covers how to tell a platform’s genuine AI testing capability from a feature list padded for the sales deck, which matters before you decide a platform has made homegrown analytics redundant.

Is your team ready to build? A readiness checklist

Answer these six honestly before the first Power App goes into production, not after auditees start depending on it.

Build readiness checklist

Maker capacity. At least one person who can own Dataverse security roles, connector permissions and flow error handling without a consultant on retainer for every change, plus a named backup who could take over without starting from zero.

Protected maintenance time. A manager willing to protect ongoing hours for fixes and small enhancements after launch, not just the initial build sprint; homegrown systems that stop getting maintained are how orphaned records happen inside the very tool meant to catch them.

An audit-trail and retention design, decided before you build. Field-level history, who can edit what after sign-off, and how long records are kept, documented on purpose rather than left to whatever Microsoft 365’s default tenant policy happens to be.

A security review, even though nobody is making you run one. The same due-diligence questions a new vendor would face, run against your own build, before workpapers start living inside it.

A named trigger for revisiting the decision. A headcount, a SOX designation, a bank charter or an external quality assessment that would flip the answer from build to buy, written down now rather than argued about later.

An exit plan. How the workpaper and issue history would export if the environment were ever retired or Microsoft changed the licensing terms again, the way it did to the per-app plan in 2026.

Worked example: MidState Beverage weighs build against buy

MidState Beverage is the site’s recurring fictional worked example, used here only to make the sections above concrete: a three-state drinks distributor with a six-person internal audit function, never a real company. For three years MidState’s audit trail has lived in an Excel issue log and a SharePoint document library for workpapers, built by an audit manager who taught herself Power Apps well enough to add a submission form for management action plans. It has worked, mostly. Two things changed this year: MidState is adding a fourth state through a small acquisition, and the lender behind its credit facility added a covenant requiring an annual internal audit function assessment, putting an external reviewer’s eyes on the system of record for the first time.

Run against the decision matrix above, MidState splits down the middle. Team size and budget shape lean build: six auditors is still closer to a first system than a mid-size function, and there is no existing software line to redirect. In-house skill leans build too, cautiously, since the self-taught auditor is competent but not a developer and has no backup. Regulatory profile leans buy: the lender covenant is not a SOX program, but it is exactly the kind of external scrutiny the liability table above describes, and “one of our managers built this in Power Apps” is a weak answer to a reviewer’s question about audit trail and retention. Analytics intensity leans build, since MidState wants to start testing accounts payable and inventory data and a platform would not replace that layer regardless.

The total-cost-of-ownership table above puts the five-year gap at roughly $64,000 for a build against $169,000 for Onspring, but MidState’s real maker capacity is thin: 200 hours of build time has to come from a contractor, not a stretched employee’s spare afternoons, which pushes the build side toward the top of its range rather than the midpoint the table uses. Weighed against a lender covenant that specifically wants a defensible system of record, MidState’s audit committee lands on the hybrid pattern above: buy Onspring, whose fit rating is Strong for both a first system and a mid-size function, as the platform of record for workpapers, issues and the reporting the lender and the audit committee will actually see, and keep the accounts payable analytics homegrown, running the tests from the site’s accounts payable analytics catalog in Python against an extract from the ERP, since Onspring would not have replaced that layer either. The risk and control matrix that used to live only in a SharePoint list moves into Onspring’s audit universe, built from the structure in the site’s risk and control matrix template, and the Power Platform skill that built the original system becomes the person who owns the Python scripts instead, which is a better use of it than being the one person who understands a homegrown app’s security roles.

Questions about build vs buy for audit software

Is SharePoint alone enough for a small internal audit team?

For a genuinely small function, often yes, at least to start. A SharePoint list with a few views can hold an issue log and a lightweight risk and control matrix for one to five auditors, which is why it rates a workable first system in the decision matrix above. It stops being enough the moment the team needs a structured audit trail across a full engagement, more than one reviewer’s sign-off, or evidence for an external quality assessment; the site’s audit software for small teams guide covers the real prices once you outgrow it.

Can Power Apps and Power Automate meet SOX documentation requirements?

Nothing stops a well-built Power App from documenting a control test the way SOX requires; the question is whether it will, consistently, once more than one person is entering data. GIAS 14.6 sets the documentation standard a workpaper has to meet regardless of the system underneath it. A packaged platform enforces that standard through workflow; a homegrown build only enforces what its builder remembered to validate, which is why most SOX-heavy public companies land in the buy column of the decision matrix above.

How much does a homegrown audit tracker actually cost?

The total cost of ownership table above works a labeled example: about $64,000 over five years for a six-person function building on the Power Platform, against roughly $169,000 for a mid-market platform over the same period. Most of that gap is Microsoft’s published per-user prices plus a stated assumption about build labor hours, not a quote for your organization; run your own numbers against the table’s line items rather than trusting either total directly.

When does a bank or SOX-heavy company need to buy instead of build?

Once examiner scrutiny or an ICFR program is the center of gravity, the audit-trail, retention and security-review liabilities in the table above stop being theoretical and start being questions a real examiner or an external quality assessor will ask. That is not a claim that a homegrown system cannot meet the bar; it is that meeting it convincingly, on a schedule you do not control, is usually more expensive than the license fee the decision matrix compares it against.

What happens to a homegrown system when the person who built it leaves?

Whatever that person did not document leaves with them: which flows fire on which trigger, why a particular security role exists, how to recover from a failed sync. Nothing about Power Apps or SharePoint prevents good documentation, but nothing enforces it either, which is the key-person risk in the liability table above. The build readiness checklist’s named backup is the cheapest insurance against this, and it is worth having before the first version ships, not after someone resigns.

Can we build now and buy a platform later without wasting the build?

Yes, and it is common: a homegrown build that started as an issue log often becomes the analytics or reporting layer that sits alongside a platform bought later, rather than being thrown away, as the hybrid pattern above describes. What does not carry over cleanly is the workpaper history inside the old system; plan a migration, not just a cutover, and the site’s implementing audit management software guide covers what that move actually takes.

internalauditguide.com has no commercial relationship with any vendor named on this page. We take no vendor money, run no affiliate links and accept no sponsored placements, and no vendor saw this page before publication. Product and company names are the trademarks of their owners. Corrections: desk@internalauditguide.com.

Sources and verification

Free PDF · 15 one-page cheat sheets

The Internal Audit Desk Reference, free by email.

GIAS 2024 on one page, the sample-size tables, the risk and control matrix, the five-C finding, the SOX deficiency ladder, the CIA exam at a glance, and nine more. Printable. Written by working auditors.

Free. Confirm your address once; the PDF link follows. Unsubscribe anytime.

Cover of The Internal Audit Desk Reference: fifteen one-page cheat sheets and templates (PDF)
Cover of The Internal Audit Desk Reference: fifteen one-page cheat sheets and templates (PDF)

Free PDF

The desk reference: 15 cheat sheets, one page each.

GIAS map, sample sizes, the RCM, the five-C finding, SOX deficiency ladder, CIA at a glance. Free, printable.

Confirm once; the PDF link follows. Unsubscribe anytime.

Comments

Leave a Reply