Onboarding a new client is one of the most repetitive, error-prone jobs in a small law firm. Someone opens a matter, drafts an engagement letter, chases photo ID, runs a conflict check, sets up a trust ledger and files everything in the right folder. Do it once and it's fine. Do it forty times a month across three fee earners and it becomes a bottleneck — and a compliance risk when a step gets skipped under pressure.
The good news is that almost every part of this process is rules-based, which makes it a strong candidate for automation. The important caveat: onboarding decisions carry professional obligations, so the goal is not to remove the lawyer. It's to remove the copying, chasing and re-keying, and to put a clean summary in front of a human for approval before the client is activated.
This article walks through an end-to-end onboarding workflow you can build for a small firm, the tools that make it work, and where to keep a person firmly in the loop.
What "end-to-end onboarding" actually covers
Before automating anything, map the full journey. For most Australian firms, onboarding a new client and matter involves six discrete stages:
- Intake — capturing the client's details, the matter type and the other parties involved.
- Conflict check — searching existing clients and matters for any conflict of interest.
- Identity verification — collecting and recording proof of identity to meet your firm's VOI and AML obligations.
- Engagement — generating a costs agreement or engagement letter with the correct scope and fee basis.
- File setup — creating the matter in your practice management system, opening a trust ledger if needed, and building the folder structure.
- Activation — the point at which a fee earner confirms everything is in order and work can begin.
Automation should carry the client smoothly through stages one to five, then stop and wait for a human at stage six. That single design decision — automate the assembly, gate the activation — is what keeps the workflow safe.
Stage 1: Structured intake instead of email tennis
Most onboarding delays start here, with details arriving in scattered emails and phone notes. Replace that with a single structured intake form that captures everything downstream steps will need: full legal names, date of birth, contact details, the matter type, the opposing party, and how the client heard about you.
A form gives you clean, structured data that the rest of the workflow can read automatically. If you'd like a detailed build for this part, we've written a full walkthrough on how to build a client intake form with Notion, ChatGPT and Zapier — it's the natural DIY foundation for everything below.
The key is to make the form the only front door. If new clients can only enter through structured intake, every subsequent step has reliable data to work from.
Stage 2: Conflict checks that run automatically — but decide nothing
This is the stage most firms are nervous about automating, and rightly so. A conflict check is a professional judgement, not a database lookup. So automate the search, not the conclusion.
When a new intake arrives, the workflow can automatically search your existing client and matter records for the new client's name and every other party listed on the form. It then produces a short report: who was matched, on which matters, and how strong the match is. That report goes to the responsible fee earner.
The lawyer still makes the call on whether a conflict exists. What's changed is that they're no longer manually running searches and eyeballing spreadsheets — they're reviewing a pre-assembled summary and clicking approve or flag. The search is automated; the decision stays human.
Stage 3: Identity collection without the chasing
ID collection is where onboarding stalls for days. The client means to send their licence and passport, then forgets. Automating the follow-up alone recovers a surprising amount of time.
Set up the workflow so that once intake is complete, the client automatically receives a secure request to upload their identification. If nothing arrives within a set period — say two business days — a polite reminder goes out automatically, and again after another interval. Your team only gets involved when documents are received or when the reminders have been exhausted.
For firms with VOI or AML obligations, the automation records what was collected and when, creating a timestamped audit trail. It does not replace the verification judgement itself — a person still confirms the documents are genuine and match — but it ensures nothing slips through undocumented.
Stage 4: Engagement letters generated from a template
Drafting costs agreements from scratch, or copying an old one and hunting for every reference to the previous client, is exactly the kind of task that produces embarrassing errors. Automation handles it cleanly.
Build a master template for each matter type with merge fields for the client name, matter description, scope of work, fee basis and key dates. When intake data feeds in, the workflow generates a populated draft. If you want the scope wording tailored to the specific matter rather than boilerplate, a language model can draft that paragraph from the intake details — but always as a draft for review, never a document sent automatically.
The output is a ready-to-review engagement letter sitting in the matter folder, waiting for a fee earner to check the scope and fees before it goes anywhere near the client.
Stage 5: File and matter setup
Once the earlier stages clear, the mechanical setup can happen without anyone touching a keyboard: create the matter in your practice management system, apply the standard folder structure, open a trust ledger where relevant, and populate the file with the intake summary, ID records, conflict report and draft engagement letter.
This is the same category of connective automation we describe in our guide to Zapier for lawyers — using a workflow tool to move data between systems that don't talk to each other, so your team stops doing the copy-paste.
Stage 6: Human approval before activation
Here is the most important part of the whole design. At no point should the automation declare a client "onboarded" on its own. Instead, the workflow assembles everything into a single approval view and pauses.
The responsible fee earner sees one screen: the intake details, the conflict report with any matches flagged, confirmation that ID has been collected and verified, and the draft engagement letter. They review, then either approve — which activates the matter and releases the engagement letter to the client — or send it back with a note.
Automate the assembly of the file. Keep the activation of the client in human hands. That boundary is what makes onboarding automation defensible.
This approval gate is not a formality. It's where professional responsibility lives, and it's what lets you automate confidently everywhere else.
Exception reports: knowing when something's stuck
A well-designed workflow doesn't just handle the happy path — it tells you when something has gone off the rails. Build a daily or weekly exception report that surfaces anything that needs attention:
- Intakes where ID hasn't been received after all reminders
- Conflict checks flagged but not yet cleared by a fee earner
- Engagement letters drafted but not approved for several days
- Matters sitting in the approval queue past a threshold
The exception report is what stops automated processes from silently failing. Instead of assuming everything worked, your practice manager gets a short list of what's stuck and who owns it. That single report often does more for onboarding reliability than any individual automation step.
A realistic build for a small firm
You don't need enterprise software to run this. A workable stack for a three-to-eight person firm looks like:
| Stage | What handles it |
|---|---|
| Intake form | Notion, Typeform or a form tool feeding structured data |
| Orchestration | Zapier or Make connecting the steps |
| Conflict search | Automated lookup against your existing records |
| ID collection | Secure upload request with automated reminders |
| Engagement letter | Template merge, optionally with AI-drafted scope |
| File setup & approval | Practice management system plus an approval step |
Start with one stage, not all six. The intake form and automated ID reminders are the easiest wins and deliver value immediately. Add conflict search and engagement letter generation once the foundation is stable. If you're weighing whether this is worth building versus adding a team member, our piece on what admin work to automate before hiring is a useful gut-check.
Where firms go wrong
The two most common mistakes are opposite extremes. Some firms automate too little — a form here, a template there — and never connect the steps, so the data still gets re-keyed. Others automate too much and let the system activate clients without review, which is where compliance problems begin. The sweet spot is full automation of the mechanical work, with a single, clear human approval before the client goes live.
If you want the broader picture of which onboarding and admin tasks are worth automating first, our free guide, 10 AI Workflows to Save 10+ Hours a Week, walks through the highest-value automations for professional services firms and how to prioritise them — a good next step before you start building.
Want these ideas working in your firm? We build controlled AI workflows for Australian professional services firms — starting with a free automation audit.