Ask a legal assistant what happens when a new matter comes in and you will hear a version of the same answer. Create the folder. Copy the structure from the last matter — or from the one before that, if the last one was done wrong. Look up the next matter number, hope nobody else took it in the past ten minutes, and write it on the whiteboard. Email the responsible lawyer. Put the limitation date in a calendar. Fill in the metadata columns, or skip them, because the file needs to be open in four minutes.
Twenty minutes, several times a week, done slightly differently by every person who does it. Then a year later somebody asks how many active real estate matters the firm has open in Alberta, and the honest answer is that nobody knows — because the metadata was optional, and optional means empty.
This is the most automatable process in a law firm, and the tooling to fix it is already inside the Microsoft 365 subscription you are paying for. Here is exactly how it works, including the parts most articles skip.
What the automated version looks like
One person fills in one form. Everything below happens without anyone touching it.
1. The intake form
A Microsoft Form — or a SharePoint list form, if you want it to live in the same site as the files — captures client name, matter type, practice area, responsible lawyer, jurisdiction, and opposing party. That is the last time anyone types this information. Every step that follows reads from it.
2. The conflict check, routed and recorded
The flow raises an approval task with the conflicts lead using the built-in Approvals action, which appears in both Teams and Outlook. Rejected, and nothing is created — the requester is notified with the reason. Approved, and provisioning continues.
The valuable part is not the routing. It is that the decision, the approver, and the timestamp are recorded against the matter automatically. A conflict check that lives in somebody’s sent items is not a record. This is.
3. The matter number
The flow reads a dedicated numbering list, takes the next value, and formats it to your firm’s convention — 2026-0147, or whatever you already use. Because one flow owns the sequence, you cannot get duplicates from two people opening files at the same moment, and you cannot get gaps from a number reserved and never used.
4. The document set
Now the flow creates the matter itself: a document set containing your standard structure — Pleadings, Correspondence, Discovery, Accounts, Closing — plus any starter documents pulled from your precedent library.
Document sets are worth understanding properly, because this is the piece that makes everything else work. A document set behaves like a folder, but it is really a content type with its own columns and its own welcome page. Every matter is created from the same definition, so every matter is organised identically — not "mostly identically, except the ones opened in 2023."
5. The metadata, applied once and inherited
Client, matter number, practice area, responsible lawyer, opened date, limitation date, and status are set on the document set as shared columns. Anything filed inside inherits them automatically.
This is the step that changes how the firm works. Nobody types metadata, so the metadata is actually complete — which means you can build a view of every open litigation matter with a limitation date in the next ninety days and trust the answer.
6. The workspace and the deadlines
A Teams channel for the matter, a Planner checklist of the opening tasks, and calendar reminders for the limitation date. The matter appears immediately in the responsible lawyer’s filtered view of open files. Nobody has to be told it exists.
7. The closure
When somebody sets the matter status to Closed, a flow applies the retention label, moves the set to the archive library, and drops permissions to read-only. File closure stops depending on whether anyone remembered — which, in most firms, is the difference between a records policy and a records policy on paper.
The part most articles skip
Search for this and you will find a dozen posts that say "then Power Automate creates the document set." It does not, at least not the way you would expect.
Power Automate has no built-in action to create a document set. The Create item and Create new folder actions will not do it. What works is the Send an HTTP request to SharePoint action, which posts to the SharePoint REST API and creates the set with its columns populated in a single call.
The good news: that action is part of the standard SharePoint connector, not the premium HTTP connector, so it requires no additional licensing. The same technique is how you set managed metadata columns drawn from the term store, which the simpler Update file properties action handles badly.
If a vendor quotes this work without mentioning it, they have either done it before and are glossing over the detail, or they have not done it before and are about to find out.
What this actually costs to license
Everything above runs on standard connectors — SharePoint, Outlook, Teams, Approvals, Forms, and Planner — included with Microsoft 365 Business Standard and Business Premium. No premium Power Automate licence is required.
Four things do cost extra, and you should know before somebody demonstrates them to you:
- Automatic Word template population. Generating an engagement letter with the client details already filled in uses the Word Online (Business) connector, which is premium. Worth knowing: a Word template using Quick Parts bound to your library columns does the same job with no flow and no premium licence at all.
- DocuSign or Adobe Sign integration. Premium connectors, on top of the e-signature subscription itself.
- Auto-applying retention and sensitivity labels by policy. A flow can apply a label on a status change with what you already have. Policies that classify content automatically across the tenant require Microsoft 365 E5 or a Purview add-on.
- AI document classification. Reading an incoming document and extracting its metadata requires SharePoint Premium, licensed per user or pay-as-you-go.
Three things that will break this
Automation built on a bad foundation fails in ways that are expensive to unpick. These are the three we get called in to fix most often.
The 5,000-item view threshold
A SharePoint library can hold 30 million items. What it will not do is let a single query scan more than 5,000 of them at once — and in SharePoint Online that ceiling cannot be raised, by you or by anyone else.
This one is worth clearing up, because it is usually explained wrong. A modern view does not break simply because the library holds more than 5,000 items; it will render a library many times that size perfectly well. What breaks is sorting, filtering, or grouping on a column nobody indexed. So the library works, and works, and then one afternoon somebody filters by client name and gets an error instead of a list — and the firm concludes SharePoint does not scale.
The fix is known: index the columns your views actually filter on. The catch is timing. SharePoint Online creates some indexes automatically, but it stops doing so once a list passes 20,000 items. The library that grew for three years while nobody was thinking about it is precisely the one that will not quietly fix itself.
The 400-character path limit
Full paths in SharePoint are capped at 400 characters, and each folder or file name within them at 255. Deep nesting inherited from a file server — client, then year, then matter, then correspondence, then a sub-folder per month — burns through that budget fast, and the result is files that will not sync and will not open.
What catches people out is that the limit applies to the encoded path. Spaces and special characters expand when the URL is encoded, so a path that looks comfortably short in the browser can already be over. Client names with ampersands and accented characters are a common way to discover this.
Metadata is the fix. Filtering a flat library by six columns gives you every view the folder tree gave you without the path length. More folders make it worse, which is counterintuitive enough that most firms do exactly the wrong thing.
The flow that leaves with the employee
The most common failure, and the most avoidable. A workflow built inside a staff member’s personal Power Automate account stops running the day that account is disabled. Nobody gets an alert. You find out when an invoice has sat unapproved for a week, or when a matter opens with no folders in it.
Automation a business depends on belongs to the business: built in a managed Power Platform environment, owned by a service account, with named co-owners and documented connections. Ask whoever builds your flows where they live. If the answer is a person’s name, that is the problem.
Where to start
Not with the flow. Start with the architecture — content types, metadata columns, permission groups, retention — because every one of those decisions constrains what the automation can do afterwards. Building flows first and fixing the structure later means building them twice.
If your files are still on a file server, in Google Drive, or in Dropbox, start further back. A SharePoint migration designed with the automation requirements already known produces a target structure that supports them, rather than one you have to retrofit.
And if your SharePoint is already a mess, you almost certainly do not need to start over. Most of what blocks automation — unindexed libraries, long paths, per-file sharing that should be group permissions — can be remediated in place while everyone keeps working. Our IT guidance for law firms covers the wider picture.
The point of all this
Automating matter intake is not really about saving twenty minutes at a time, although it does that. Across the law firms we have built this for, it returns 20 to 40 hours of administrative time a month — the folder-building, the metadata typing, the chasing, and the numbering.
The bigger change is quieter. A process that runs the same way every time produces data you can trust, and a firm that can answer questions about its own open files is a different firm from one that cannot.
AltaCom designs SharePoint document architecture and builds the Power Automate workflows that run on it for law firms across Calgary, Edmonton, and Alberta. We start with an audit of what you have and the manual processes worth automating.
See our SharePoint Setup & Automation service — or book a workflow assessment and we will map your current intake process end to end.