Power Apps for Small Businesses: When It Replaces Custom Software

Often, yes, for internal tools. If your own staff will use the app, its data can live in SharePoint lists or Excel, and it replaces something like a paper checklist, stock log or job tracker, Power Apps can do it on your existing Microsoft 365 business plan. It's the wrong choice for customer-facing systems, deep integrations or formal records.

The line that decides the cost is the connector. Microsoft 365 Business Basic, Standard and Premium all include Power Apps with standard connectors, which cover SharePoint, Excel files in OneDrive or SharePoint, Outlook and Teams. The moment an app needs a premium connector (SQL, Azure or Dynamics 365), a custom connector to another system's API, on-premises data or a full Dataverse database, everyone who runs it needs a standalone licence.

Follow me on Instagram@sagnikteaches

That licence is Power Apps Premium at $20 a user a month, paid yearly, or pay-as-you-go at $10 per active user per app per month through an Azure subscription. For a 15-person business that's between nothing and $300 a month, depending on which side of the line the app sits. Designing to stay on the included side is usually the single biggest saving.

Connect on LinkedInSagnik Bhattacharya

Power Apps doesn't remove the need for someone who thinks like a developer. It removes the need for them to write most of the code. A practice manager who's comfortable with Excel formulas can build a useful internal app in a few weeks; one who isn't will need help, and that help is where most of the real cost sits.

Subscribe on YouTube@codingliquids

What your Microsoft 365 plan already lets you build

Microsoft's plan comparison lists "Power Apps for Microsoft 365" as included in Business Basic ($7 a user a month, annual), Business Standard ($14) and Business Premium ($22). Microsoft's licensing FAQ adds that there's no limit on the number of apps you can run this way, as long as they use standard connectors. In practice that covers:

• Data in SharePoint lists, the most common home for a small business app's data.
• Excel workbooks stored in OneDrive or SharePoint, fine for small, simple tables.
• Outlook and Teams, so an app can send an email or post a message.
• Dataverse for Teams, a built-in data store inside Teams, included in the plan versions that come with Teams. Microsoft puts its capacity at roughly 1 million rows per team, you can't buy more, and it doesn't support custom connectors.
• Power Automate for Microsoft 365, for flows that run behind the app. On Business Basic and Standard it's limited to cloud flows.

There's also a free Developer Plan for building and testing apps in a non-production environment. It's useful for trying an idea, but anyone using the finished app in daily work needs a production licence.

Where licence costs start

Two details make pay-as-you-go more useful than it looks. An "active user" is someone who opens the app at least once in the month, so staff who don't open it that month cost nothing. And staff with a Microsoft 365 licence using an app with only standard connectors aren't counted at all. Microsoft's licensing FAQ also still lists a per-app plan at $5 per user per app per month, but it no longer appears on the main pricing page, so ask whoever sells you Microsoft licences whether you can still buy it before you plan around it.

Five questions that decide whether Power Apps is enough

1. Who will use it? Staff with Microsoft 365 accounts are cheap to serve. People outside the business are not: Power Pages charges $4 per authenticated active user per website a month, so a portal for 300 active clients would cost $1,200 a month. At that point, specialist software built for the job usually wins.
2. Where does the data live? SharePoint and Excel keep you on the included side. If the app must read your practice-management system, booking platform or accounts software directly, you're into custom or premium connectors and paid licences, and the other system must offer an API (a documented way for software to exchange data).
3. How much data? Canvas apps fetch only the first 500 records for any query the data source can't handle itself (you can raise that to 2,000). Past that, a search can quietly return incomplete results. Lists of a few hundred items are never a problem; lists heading for tens of thousands need careful design.
4. How tangled is the logic? Forms, approvals, checklists and status tracking suit Power Apps. Complex pricing rules, scheduling across many people, or real-time sync with two or more outside systems are signs you need a developer's view.
5. Who will maintain it? If one person builds it and nobody else understands it, you've swapped a software supplier for a single point of failure. Name a second person before you start.

The rule that follows: build in Power Apps when the users are staff, the data sits in SharePoint, Excel or Dataverse for Teams, the lists stay in the low thousands, and two people can maintain it. If any one of those fails, either price the licence line honestly or get a quote for software built for the job. No-code tools vs a custom build covers the wider version of that decision.

A veterinary practice's ward app, costed three ways

Consider a veterinary practice with four vets, seven nurses, three receptionists and a practice manager: 15 people on Microsoft 365 Business Standard. Hospitalised animals are tracked on clipboards clipped to kennel doors (feeding, medication times, observations) plus a whiteboard showing kennel occupancy. Handovers depend on handwriting. The consumables stock count is a spreadsheet emailed round every Friday.

The practice manager wanted one tablet app for the ward: today's inpatients by kennel, a check screen for each animal, an alert to the on-duty vet when a nurse flags a concern, and a simple stock count. As an illustration, a freelance developer quoted $9,500 to build a small web app for the same job plus about $100 a month for hosting and support. The practice then priced the Power Apps routes:

Route A won for the first three months. At about eight admissions a day, the receptionist's retyping cost around 16 minutes a day, far cheaper than any connector. The app was built on three SharePoint lists (Inpatients, Checks, Stock items), a canvas app for the ward tablets and one cloud flow posting a Teams message to the duty vet when a check was marked "Concern". The practice manager built it in about 30 hours over five weeks, with another 10 hours of testing on real shifts.

Route C stays on the table for later. If the retyping starts causing mistakes, a connector to the practice-management system and pay-as-you-go licences for the staff who actually open the app would cost around $110 a month, still well short of the custom quote. Anything that forms a formal clinical record stayed in the practice-management system throughout; the ward app is a working tool, not the record.

Where Power Apps beats a developer, and where it doesn't

A dental practice's lab-work tracker is a textbook fit. Each crown, bridge or denture sent to the lab gets a row in a SharePoint list with the patient reference, lab, date sent and date due. A one-screen app shows what's overdue, and a daily flow emails the practice manager anything due tomorrow that hasn't arrived. It's internal, small, and needs no outside system.

A nursery's parent booking portal is not. Parents are outside users, so a Power Pages site would be billed per authenticated monthly active user, and a nursery with 120 active parents would pay $480 a month before anyone has built anything. The same nursery would be better served by specialist nursery software with a parent app built in, plus Power Apps for internal jobs such as room ratios or staff training records.

A tutoring agency's tutor timesheets sit in between. Tutors are on Microsoft 365, so a timesheet app on a SharePoint list costs nothing extra. But if the agency wants hours to flow straight into its payroll software, it needs that software's connector, which may be premium. The cheap design is two apps: tutors submit hours through a standard-connector app on the SharePoint list, at no extra cost, and the two office staff who approve hours and send them to payroll use a second app with the premium connector, on pay-as-you-go at $10 each a month.

If what you need is an automation rather than an app (no screens, just "when this happens, do that"), Power Automate for small businesses is the better starting point. And if the job is an AI assistant for staff or customers, Copilot Studio is Microsoft's tool for that, priced separately.

Letting Copilot draft the app, and the licensing catch

Power Apps can build a starting point from a plain-English description. Under "Start with data", you describe what you want to track and Copilot proposes tables and relationships, then opens an app for you to edit. The practice manager tried:

Create tables to track hospitalised animals, including patient
name, species, owner surname, kennel number, admission date,
feeding times, medication due times and check status.

An illustrative result: three tables (Inpatient, Medication schedule, Ward check), linked so each inpatient has many checks and many medication times, with sample rows filled in. Two things needed fixing. The owner's surname had no business in a ward app, so it came out. And more importantly, Microsoft's documentation says this feature generates Dataverse tables and needs access to a Dataverse database. Dataverse is a Premium capability, so an app built this way needs Premium or pay-as-you-go licences for everyone who runs it.

The practical answer was to use Copilot's draft as a design (which tables, which columns, how they link) and then create the same structure as SharePoint lists, keeping the app on the included side. Five minutes of Copilot saved an hour of planning; building on its tables directly would have cost $300 a month.

What breaks after launch, and how you'd notice

• Lists outgrow the query limit. The stock list reached 2,300 items once batch numbers were added. A search that the data source couldn't handle itself started showing only the first 2,000, and a nurse couldn't find an item that was on the shelf. The editor had flagged it with a blue underline, a delegation warning; rewriting the search as a delegable filter fixed it.
• The builder leaves. Make at least one other person a co-owner of the app and its flows, and keep a one-page note of what each list and screen does.
• Connectors get reclassified. Microsoft has previously moved SQL, Azure and Dynamics 365 connectors from standard to Premium. Check your app's connectors against Microsoft's current list once a year.
• Flows switch themselves off. Power Automate turns a flow off after 14 days of continuous failure, and after 90 days without being triggered unless the owner has a Premium licence. A seasonal flow, such as an annual stock-take reminder, can quietly disappear this way.
• AI features change how they're paid for. AI Builder credits are being replaced by Copilot Credits, and the seeded AI Builder credits stop with new or renewed licences from 1 Nov 2026. If your app reads forms or invoices with AI Builder, check how it will be billed at your next renewal.

Checking the app earned its place after a month

An internal app replaces a habit, so judge it against the habit it replaced, not against a feature list. The practice checked four things after four weeks, with the practice manager pulling the numbers from the SharePoint lists and the rota:

1. Handover time. The evening handover went from about 20 minutes of reading clipboards to about 12 minutes of scrolling the kennel view.
2. Missed or late checks. Every check now has a time stamp, so late ones are visible. In week one, 9% were more than 30 minutes late; by week four it was 3%, mostly on the busiest surgery day.
3. Stock-outs. Two in the month before the app, none in the month after, because the count now happens on the tablet at the shelf rather than from memory on Friday afternoon.
4. Alerts that reached the wrong person. One did. The flow sent "Concern" alerts to a named vet, and on her day off the message sat unread for 40 minutes until a nurse phoned. The fix was to post alerts to a ward channel in Teams that whoever is on duty watches, instead of to one person.

If none of those numbers had moved, the honest conclusion would have been that the clipboards were fine, and the app should be retired rather than polished. Build time is spent either way; running a tool nobody needs costs attention every week.

A one-page brief to write before anyone builds

Whether you build it yourself, hand it to a freelancer or ask a developer to quote, write this first. The practice's version:

App:            Ward checks (tablet, two devices on the ward)
Replaces:       Kennel-door clipboards, occupancy whiteboard,
                Friday stock spreadsheet
Users:          4 vets, 7 nurses (all on Microsoft 365)
Outside users:  None
Data:           SharePoint lists: Inpatients, Checks, Stock items
Records:        About 8 admissions a day; 400 stock items now,
                up to 2,500 within two years
Must do:        Inpatients by kennel; check screen per animal;
                "Concern" alerts the duty vet in Teams; stock count
Nice to have:   Pull admissions from the practice system (later)
Formal records: Stay in the practice-management system
Licence check:  Standard connectors only; no Dataverse
Owner:          Practice manager   Co-owner: head nurse
Review:         After 3 months: retyping errors? connector worth it?

A brief like this does two jobs. It tells the builder what to make, and it shows you, before anyone spends an hour, which side of the licence line the app will land on. If the brief itself turns out to need a developer, briefing a developer on a custom workflow takes it from there, and the five-year cost comparison of per-seat fees vs one custom tool helps with the long-run sums.

Further reads

• Google Workspace vs Microsoft 365 for AI: Which Suits Your Business? — If you're still choosing between the two suites.
• Microsoft 365 Business Premium vs Standard for AI Security — What the pricier Microsoft 365 plan adds for security.
• AI for Veterinary Practice Owners: Costs, Tools, and First Steps — Where else AI fits in a vet practice, with costs.
• Who Owns the AI Workflows a Consultant Builds for You? — Who owns an app a freelancer builds for you.
• Copilot Chat vs Microsoft 365 Copilot: What Do You Get Free? — What the free and paid Copilot tiers add inside Microsoft 365.
• Is Microsoft 365 Copilot Worth It for a Small Business? — Whether the paid Copilot licence earns its fee.
• Airtable vs Google Sheets for Business Data AI Can Use — Which home suits business data you want AI to read: Sheets for flat lists, Airtable for linked records, with prices, limits and a worked school example.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Microsoft Power Apps pricing page; Microsoft 365 business plan comparison page; Microsoft Learn articles 'Power Platform licensing FAQs', 'Pay-as-you-go meters', 'Understand delegation in a canvas app' and 'Build apps through conversation with Copilot' (checked September 2026).