Automation Services

Power Automate builds for organisations already on Microsoft 365

If your organisation runs on Microsoft 365, a good deal of automation value is sitting in software you already pay for. The catch is that building it well is a different discipline from clicking through a template.

What is power automate?

Power Automate is Microsoft's automation service for building flows across Microsoft 365, Dynamics 365, SharePoint, Teams and hundreds of other connectors. It suits Australian organisations already committed to the Microsoft stack, where staff data, files and approvals already live and licensing for standard connectors is largely in place.

Get a fixed written quote
Typical timeline
3 to 8 weeks
What drives cost
How many processes are in scope, which connectors they need, and whether approvals and exception handling have to be built around them.
Best for
Microsoft 365 and Dynamics organisations automating approvals, documents and internal processes
You own
Every flow, solution and connection, inside your own Microsoft tenant
Built with
Power Automate, SharePoint, Dataverse, Teams, Azure where flows outgrow the platform
What connects to whatSharePointTeams approvalsDynamics 365Outlook and emailDataverseExcel and ListsPowerAutomate
Flows run under licences you already hold, inside your existing Microsoft tenant.

Your handover

When Power Automate is the right tool

The argument for Power Automate is proximity. Your staff identities are already in Entra ID, your documents are in SharePoint and OneDrive, approvals can appear in Teams where people already work, and the data never leaves your Microsoft tenant. For an Australian organisation with data residency concerns under the Privacy Act 1988 and the Australian Privacy Principles, keeping automation inside a tenant hosted in an Australian region is a materially easier conversation than routing records through a third party service overseas.

  1. 01Tenant and flow audit with owners and failure rates
  2. 02Written licensing position for standard and premium needs
  3. 03Development, test and production environments configured
  4. 04Data loss prevention policies for connector groups
  5. 05Flows packaged in solutions with connection references
  6. 06Error handling, retry policies and failure alerting
  • Custom audit logging beyond default run history
  • Ownership assigned to service accounts, not individuals
  • Administrator documentation and maker training
The rest of the answer

The processes it handles best are the internal ones: leave and purchase approvals, document generation and filing, onboarding checklists that create accounts and assign training, form submissions from Microsoft Forms or Power Apps that need routing, contract review workflows with reminders, and notifications that pull people into Teams rather than another inbox.

It is less compelling when the systems you need to connect are not Microsoft, when volume is high, or when the logic is genuinely complicated. Those cases are covered further down, because pretending one tool fits every job is how organisations end up with three hundred unmanaged flows and no one who understands them.

Microsoft licensing terms change, so we confirm the current position with your reseller rather than relying on what was true last year.

Environments, DLP policies and who owns a flow

The default environment in a new tenant is where ungoverned automation accumulates. Anyone can build there, connectors are wide open unless restricted, and flows are owned by the individual who created them. The predictable outcome is that a finance officer leaves, their account is disabled, and three approval processes stop working with no explanation because the flows were personal.

Getting a single flow working is easy

We set this up properly. Separate development, test and production environments so a change can be tested before it affects anyone. Data loss prevention policies that define which connectors can be combined, which is how you prevent someone innocently building a flow that copies internal documents to a personal cloud drive. Solutions used as the deployment unit with connection references, so a flow moves between environments without being rebuilt by hand. Ownership assigned to a service account or a security group rather than a person, with co owners set on everything.

For organisations with many citizen developers, this governance layer is the whole project. Getting a single flow working is easy. Making it safe for two hundred people to build flows without creating a data leak or an unmaintainable sprawl is the actual engineering, and it is what governance and consulting work exists to establish.

How the engagement runs

How a Power Automate engagement runs

We start with the licensing and governance questions because they constrain everything else, then build the first process end-to-end so your team can see the pattern before we scale it across the backlog.

  1. 01Tenant reviewExisting flows, owners, connectors in use, environment structure and current licensing
  2. 02Process selectionThe two or three highest value processes, checked against the connectors they will need
  3. 03Licensing confirmationStandard versus premium requirements documented and confirmed with your reseller in writing
  4. 04Environment and policy setupDevelopment, test and production, with data loss prevention policies applied
  5. 05Build in a solutionFlows developed with connection references, error handling and logging built in from the start
  6. 06Test and promoteTested with real data in the test environment, then deployed as a managed solution to production
  7. 07HandoverOwnership assigned to a service account, documentation written, and your team trained to extend the pattern
DiscoverDesignBuildTestHandover
Two decisions on your side that keep the project moving

Where an organisation already has a large number of unmanaged flows, the first engagement is often an audit rather than a build. Knowing which flows exist, who owns them, which are failing and which duplicate each other is usually worth more than another flow.

Choose the right level

Microsoft 365 licensing: what your seats already include

This is the question that decides the shape of most projects, and it is worth answering before design rather than after. Microsoft 365 seats include use of Power Automate with standard connectors in the context of the Microsoft services you are licensed for. That covers a great deal: SharePoint, Outlook, Teams, OneDrive, Excel, Planner, Forms and similar.

What you want to do

01

Automate SharePoint, Outlook, Teams and Forms

Licensing position

Generally covered by seeded rights in Microsoft 365

Practical implication

No extra licence, which is where most quick wins live

02

Connect to SQL Server, Salesforce or a custom API

Licensing position

Premium connectors, licensed per user or per process

Practical implication

Decide early whether the flow is used by many people or few

03

Store structured data in Dataverse

Licensing position

Premium, with capacity considerations

Practical implication

Often the right call for real applications, overkill for a simple approval

04

Reach an on premises system through a gateway

Licensing position

Premium, plus gateway infrastructure to maintain

Practical implication

Someone has to own the gateway server and its updates

05

Drive a legacy interface with robotic process automation

Licensing position

Premium plus an add on for unattended runs

Practical implication

Fragile by nature, so treat it as a last resort with a maintenance budget

How we work this out during scoping

What is not included is the premium tier. Premium connectors, which include SQL Server, Salesforce, generic HTTP requests, custom connectors and on premises gateway access, require a Power Automate Premium licence for the users involved, or a per process licence attached to the flow itself. Dataverse as a data store, process mining and robotic process automation also sit above the seeded rights, with unattended automation needing an additional add on.

The distinction between per user and per process licensing changes architecture. A flow used by four hundred staff is usually cheaper licensed per process, since the licence attaches to the flow rather than to everyone who triggers it. A handful of power users automating their own work is usually cheaper per user. We work this out during scoping and put the licensing position in writing as part of the proposal, because discovering a premium connector requirement in week five is a genuinely unpleasant surprise. Microsoft licensing terms change, so we confirm the current position with your reseller rather than relying on what was true last year.

Error handling, retries and the limits you will meet

Power Automate applies a default retry policy to most actions, retrying a handful of times with increasing gaps before giving up. That is sensible for transient network faults and useless for logic errors, so we configure retries deliberately per action rather than accepting the default everywhere. A read is safe to retry. A step that creates a record is not, unless the flow checks whether the record already exists.

The full list

Beyond retries, we use scopes to build a try and catch pattern: a block of actions, then a handler configured to run only when the block fails, which logs the failure with context and notifies a monitored channel. Without that, failure notifications go to the flow owner's mailbox and quietly become a rule that files them away. Long running approvals get reminders and a timeout path, because an approval waiting on someone who has left the organisation will otherwise wait forever.

Then there are platform limits, which surprise people at scale. There are request limits per user in a 24 hour period, concurrency and looping limits, message size limits, and throttling on many connectors. Run history is retained for a limited window, so if you need an audit trail beyond that you must write it somewhere yourself, which for us usually means a SharePoint list or Dataverse table populated by the flow. Microsoft will also eventually suspend a flow that fails continuously, and the notice goes to the owner. If that owner is an individual rather than a monitored account, nobody finds out.

  • Retry policies configured per action rather than left at the default
  • Try and catch scopes with a failure handler on every production flow
  • Alerts to a monitored Teams channel, not a personal mailbox
  • Checks that prevent a retried step from creating duplicate records
  • Custom audit logging where run history retention is not sufficient
  • Approval timeouts, reminders and delegation paths for absent approvers

When Power Automate is the wrong layer

Volume is the first honest limit. A flow that runs a few hundred times a day is comfortable. A process running tens of thousands of times a day will meet throttling and request limits, and the visual designer becomes a slow way to work through a problem that a small service would handle without breaking a sweat. At that point Azure Logic Apps, Azure Functions or purpose built code is the better answer, and the fact that Microsoft sells both should make it easier to say so.

The third case is a stack that is not Microsoft

Complexity is the second. Once a flow has forty actions, nested conditions and several parallel branches, you have written a program in a medium that offers poor version control, awkward testing and no meaningful code review. Anything with real algorithmic content, complex data transformation or transactional integrity across systems belongs in written code or behind a proper API, with Power Automate calling it rather than containing it.

The third case is a stack that is not Microsoft. If your core systems are a non Microsoft CRM, a self-hosted database and half a dozen software as a service tools, you are buying premium connectors to reach all of them and losing the licensing advantage that made the platform attractive. Compare it against a self hosted n8n build where you control the runtime, or Zapier for a small number of mainstream connections. If you have not decided which layer belongs where, the process view in workflow automation comes first. Larger organisations juggling several of these at once usually need the governance conversation we describe in our enterprise work.

How we scope it

Four ways to scope your Power Automate project

We do not publish package prices, because the same brief can be a short build or a long one. These are the shapes the work usually takes. Tell us which one sounds like you and you will get a fixed written quote that spells out exactly what it covers.

First workflows

The two or three processes costing the most time now

Fixed written quote, agreed before work starts

  • Tenant and flow audit with owners and failure rates
  • Written licensing position for standard and premium needs
  • Development, test and production environments configured
Request a quote
Most common

Connected stack

The systems you already pay for, talking to each other

Fixed written quote, agreed before work starts

  • Everything in First workflows
  • Data loss prevention policies for connector groups
  • Flows packaged in solutions with connection references
  • Error handling, retry policies and failure alerting
Request a quote

Operations platform

Operations running on automation you can see and audit

Fixed written quote, agreed before work starts

  • Everything in Connected stack
  • Custom audit logging beyond default run history
  • Ownership assigned to service accounts, not individuals
  • Administrator documentation and maker training
Request a quote

Automation care

Watching, fixing and extending as the processes change

Rolling monthly, quoted in writing

  • Every workflow monitored, with alerts that reach a person
  • Fixes when an upstream system changes its behaviour
  • New workflows added from your backlog each month
  • Rolling, cancel with 30 days notice
Request a quote

These are shapes, not menus. Most quotes end up somewhere between two of them, and we will say so when the honest answer is the smallest one. Describe the problem and we will tell you which it is.

Questions buyers usually ask

Frequently asked questions

Ownership and handover

Do we need extra licences to use Power Automate?

It depends on what the flows touch. Automating SharePoint, Outlook, Teams and Forms is generally covered by the rights included with Microsoft 365 seats. Reaching SQL Server, a custom API, an on premises system or Dataverse requires premium licensing, either per user or per process. We confirm the exact position with your reseller during scoping and put it in the proposal in writing.

What happens to our flows if the person who built them leaves?

Nothing, if they were built correctly. We assign ownership to a service account or security group rather than an individual, add co owners, and store flows in solutions that can be exported. That is precisely the failure we are most often called in to repair, where a departed employee's disabled account took three business processes with it.

Detail and edge cases

How long does a Power Automate project take?

Usually 3 to 8 weeks for a first engagement covering two or three processes plus governance setup. Environment and policy configuration takes the first week or so. Where an on premises gateway or premium connector approval is needed, internal IT scheduling tends to set the pace rather than development, so we raise those requests immediately.

How do we know when a flow has failed?

Every production flow we build has a failure handler that logs the error with context and posts to a monitored Teams channel with the record identified. We do not rely on the default notification, which goes to the owner's mailbox and is usually filtered. We also set up a weekly summary of run failures across the tenant so gradual degradation is visible before it becomes an incident.

Can Power Automate connect to our on premises server?

Yes, through the on premises data gateway, which requires premium licensing and a server to run on that someone must patch and monitor. It works well when set up properly. We will always check first whether the on premises system exposes an API or can be reached another way, because a gateway is an extra piece of infrastructure with its own failure modes.

We already have hundreds of flows nobody manages. Can you help?

That is a common starting point and the first job is visibility. We inventory every flow with its owner, connectors, run frequency and failure rate, identify duplicates and orphans, and give you a prioritised remediation list. Usually a portion can simply be switched off, a portion needs reassigning to service accounts, and a small number need rebuilding properly with error handling.

Get the licensing answer before you build

Tell us which systems the flows have to touch and how many people will use them. We reply within one business day with the licensing position and a fixed written quote.