TechRam For Executives and GMs
How do you turn SharePoint into the system your operation runs on?
Start with the register rather than the folder. Every operation tracks a handful of things — incidents, assets, renewals, approvals, complaints — and each one is a list with an owner and a set of states, not a folder. SharePoint is bought as a file server, which is why it ends up as one.
The platform isn’t the problem
Almost every Australian mid-market organisation we walk into already owns SharePoint. It came with Microsoft 365, it has been there for years, and it is being used as a place to put files.
That’s the gap. SharePoint is perfectly good at being a file share, and being a file share is worth almost nothing. The value is in the layer above it: the registers, the approvals, the states and the owners that turn a pile of documents into a system somebody is accountable for.
Nobody buys that layer. It has to be built, and it’s built out of things the licence already includes.
What “runs on” actually means
A system your operation runs on has four properties a document library does not.
Every item has a state. Not a filename convention, not a folder called “In progress”. A field, with a defined set of values, that a person or a flow changes deliberately.
Every item has an owner. One name, visible on the record, who is accountable for moving it to the next state. When the owner is a team, nothing moves.
The record is written as the work happens. Not reconstructed at the end of the month by someone reading their sent items. If the evidence is assembled afterwards, it’s a story about the work, not a record of it.
Something notices when it stops. A reminder, an escalation, a report that shows the queue getting longer. A system nobody is told about when it fails is a system that fails quietly.
Where organisations get stuck
Two patterns account for most of it.
The first is the intranet that became a filing cabinet. It was built once, around the org chart, and it hasn’t been touched since. People don’t look there because the last three times they did the answer was out of date. The fix isn’t a redesign — it’s structuring around the work people actually do, and putting an owner and a review date on every page that matters.
The second is the process that lives in an inbox. Requests arrive by email, get triaged when someone has time, and are tracked in a spreadsheet one person maintains. It works right up until that person takes leave. The fix is a list with states and owners, and a form that writes to it.
Neither of these is an AI project. Both of them are what has to be true before an AI project is worth funding.
The order it usually happens in
Start with the read. An Environment Baseline tells you what you have: permissions, external sharing, sprawl, storage, and what each is costing. Two to three weeks, read-only, and it ends in a prioritised plan you own. It is boring and it’s the reason everything after it can be scoped against evidence rather than a guess.
Then governance, because permissions are the guardrail. Access granted by group rather than by forwarded link, external sharing under a policy, and the sites that sprawled consolidated with an owner named against what stays.
Then the operational layer: the registers, approvals and trackers your people are currently running in spreadsheets. This is where the hours come back, and it’s usually where the business case pays for itself.
Then, and only then, AI. An agent can only answer from what it can reach. Once the permissions are known and the content has one current version, it has something to answer from. Before that, it will answer confidently and you’ll have no way to tell which answers to trust.
What it looks like when it lands
Achieve Australia ran thirty onboarding and offboarding events a month across 1,700 staff, spread over five systems with no single record of what had happened. The pipeline that replaced it starts with one request and ends with an audit record: 120 hours a month back, and $126K a year recovered. The full story names the systems.
At EPP Law, every claim started on a printed form that was scanned, emailed, filed by hand and keyed into a spreadsheet, in English only. It’s now one online form in three languages, and a single submission files the claim to SharePoint, the register and the client at once, with nothing typed twice. The full story shows the form.
Neither of those is a new platform. Both are the licences the organisation already held, arranged into something with owners and states.
Where to start
If you don’t know what is in your tenant, start with the Environment Baseline. If you already know and the problem is a specific process, the catalogue lists what we have built and for whom.
And if you want to see the current state yourself before talking to anyone, the free tools deploy into your own tenant and answer one question each.
Common questions
01. What makes something a system rather than a document library?
Four properties. Every item has a state, set deliberately rather than by a filename convention. Every item has one named owner. The record is written as the work happens rather than reconstructed at month end. And something notices when it stops.
02. Where do organisations usually get stuck?
Two patterns. The intranet that became a filing cabinet, built once around the org chart and not touched since. And the process that lives in an inbox, triaged when somebody has time and tracked in a spreadsheet one person maintains, which works right up until that person takes leave.
03. Does the AI work come before or after this?
After. An agent can only answer from what it can reach, so until the permissions are known and the content has one current version it will answer confidently and you'll have no way to tell which answers to trust.
04. What order does the work usually happen in?
The read first — permissions, external sharing, sprawl and storage, and what each is costing. Then governance, because permissions are the guardrail. Then the registers and approvals, which is where the hours come back. Then AI.