Every AI session starts knowing nothing about you.

You already know AI can help with drafting, summarising, and research. But it keeps wasting your time before it can be useful.

01

You re-explain everything, every time

Client name, matter type, jurisdiction, which solicitor is handling it, the history — every session starts from zero. That setup time adds up across a team.

02

Generic AI doesn't know your practice

Queensland property law, your firm's standard clauses, your preferred precedents, your client relationships — none of that exists in ChatGPT or Claude by default.

03

Checking the output takes as long as writing it

When AI doesn't know the matter, it guesses — and you spend more time fixing generic output than you saved generating it.

04

Sensitive context gets typed into third-party systems

Pasting client details into ChatGPT every session creates a data trail you don't control and can't audit. That's a professional conduct risk.

"It takes longer to check what the AI wrote than to just write it myself — because it doesn't know anything about the client."

— Common feedback from legal practitioners using generic AI tools

It gives your AI a permanent memory of your practice.

Engram sits quietly behind the scenes. When you open a conversation with Claude or ChatGPT, Engram has already told it who the client is, what the matter involves, and what your firm needs to know — before you type a single word.

📋

Client briefings

Before a client call or meeting, ask your AI to brief you. It already knows the matter history, open issues, and last correspondence.

"Brief me on the Morrison estate before my 2pm call" → full summary, no setup required.
✍️

First drafts that fit

When the AI already knows the parties, jurisdiction, and matter type, first drafts come out correctly scoped — less time fixing generic output.

"Draft a cover letter for the Harrison conveyance" → correct parties, address, matter reference, your firm's style.
🔍

Conflict checks

Cross-reference a new client against existing matters and relationships automatically, with your own client register loaded in.

"Run a conflict check for Chen Family Trust" → checks against your loaded matter register.
📨

Correspondence drafts

Letters, emails, and file notes drafted in your firm's voice — already addressed to the right parties with the right matter reference.

"Draft a follow-up to the Nguyen contract negotiation" → knows the parties, sticking points, your preferred tone.
📂

Matter summaries

When a partner needs to pick up a matter from a colleague, ask Engram to summarise the file to date — status, key dates, open items.

"Where are we up to on the Riverside development?" → structured summary from loaded matter notes.
📅

Deadline & task tracking

Load your key matter dates. Ask what's due this week or flag anything urgent across active files — without opening each one manually.

"What deadlines do we have in the next 14 days?" → scans across loaded matters and surfaces urgent items.

Same AI. Same question. Completely different result.

Here's what actually changes when Engram is connected — using a scenario from a Queensland property practice.

Without Engram
You
Draft a contract review letter for the Henderson purchase.
AI response
To draft a contract review letter I'll need some information:

• What is the buyer's full name and address?
• What is the property address?
• What is the purchase price?
• What jurisdiction applies?
• Are there any special conditions?
• What is your firm's letterhead format?
⏱ You spend 5–10 minutes providing context before any useful output.
With Engram
You
Draft a contract review letter for the Henderson purchase.
AI response — Engram context loaded
Henderson Family Trust — 14 Riverview Tce, Fig Tree Pocket QLD
Purchase price: $1.24M · Standard REIQ contract · Settlement 45 days

Drafting your contract review letter now. Note I've flagged the pool safety certificate clause as non-standard — do you want me to include a requisition on that?
✓ Useful output in seconds. Your review focuses on substance, not setup.

What Engram's AI does — and doesn't — do.

We understand why lawyers are cautious about AI. The concerns are legitimate. Here is exactly where the line is drawn.

✓ What it does well
  • Drafting first versions of standard correspondence and file notes
  • Summarising matter histories and flagging open items
  • Generating structured briefings before client meetings
  • Cross-referencing client names against your matter register
  • Reformatting and proofreading documents in your firm's style
  • Answering "where are we up to" questions across loaded files
  • Saving repetitive admin time on routine communications
— What it does not do
  • Give legal advice or apply legal judgment to a matter
  • Determine the correct legal framework or strategy
  • Replace your review — everything it produces needs sign-off
  • Access court systems, PEXA, or external databases
  • Make decisions about client matters autonomously
  • Guarantee accuracy on complex or novel legal questions
Think of it as a very well-briefed paralegal who drafts quickly but whose work you always review. The legal judgment stays with you. Engram just removes the setup friction so your time goes to work that actually requires a lawyer.

Simple in practice.
No new software to learn.

There are no new apps to install and no change to how you currently use Claude or ChatGPT. Engram works behind the scenes.

Most firms are up and running within a week. The onboarding session is one hour — we handle the rest.

1

We build your firm's context file

We sit with you for about an hour and collect the information your AI should always know: practice areas, key clients, matter types, jurisdiction, preferred precedents and style. We structure this into a secure context document.

You tell us what to include. You own it entirely.
2

We load specific matter folders for the pilot

For the trial, we take two or three of your existing client folders — anonymised or selected by you — and structure that information into Engram. Nothing is shared with AI companies. Engram stores it on Australian servers.

You choose which matters. Anonymised data works fine for the pilot.
3

Your AI tools connect to Engram automatically

When you open Claude or ChatGPT through your normal workflow, Engram silently provides the relevant context before your session starts. You don't configure anything each time — it just works.

No new interface to learn. Same AI tools you already use.
4

We refine based on your feedback

Over the 4-week pilot we check in weekly to understand what's working and what the AI is getting wrong. We update the context file based on your feedback.

The pilot is a learning exercise — not a finished product commitment.

Your client data. Your control. Always.

For a law firm, data handling is not optional — it's a professional obligation. Here is exactly how Engram handles your information.

🇦🇺

Australian servers only

Your context data is stored exclusively on Australian infrastructure, operated by FormFill Group Pty Ltd, South East Queensland. It does not leave the country.

Australian hosted
🔒

Encrypted at rest and in transit

All data stored in Engram is encrypted. All connections use TLS 1.3 — the same encryption standard as online banking.

Bank-grade encryption
🚫

Never used to train AI models

Your client data is never shared with OpenAI, Anthropic, or any other AI company. It is never used to train AI models. It exists only to serve your queries.

Zero training use
👤

You control access entirely

Only authorised people at your firm can access your context data. Access is logged. You can revoke it at any time and request deletion of all stored data.

Full access control
📋

Separate from AI company servers

Your data lives in Engram, not in ChatGPT's or Claude's servers. The AI receives context for your session only — it does not retain or store it.

Isolated from AI platforms
🗑️

You can delete everything, anytime

At any point — including at the end of the pilot — you can request complete deletion of all data held in Engram. No retention, no archiving without consent.

Right to deletion

Four weeks. No cost. No obligation.

We're offering three practices a structured 4-week trial at no charge. You get a real working setup with your own matter data. We get feedback on what works in a real legal environment. At the end of four weeks, you decide if it's worth continuing.

1
Onboarding session
1 hour with your team to build the context file
2
Load 2–3 matters
You select which files we use
3
4-week live trial
Weekly check-ins to refine and improve
4
Debrief & decision
Continue, adjust, or walk away — your call
🔒  No cost for the pilot  ·  Your data deleted on request  ·  No lock-in contract

Apply for one of the three pilot spots.

We're limiting the first round to three legal practices so we can give each one proper attention. Tell us about your firm and we'll be in touch within one business day.

What you get

  • Full 4-week working pilot at no cost
  • Firm context file built by our team
  • 2–3 real matters loaded and live
  • Weekly refinement check-ins
  • Complete data deletion if you walk away

Good fit if you…

  • Already use Claude or ChatGPT
  • Have 2–20 practitioners
  • Do property, commercial or family law
  • Feel frustrated re-explaining context
  • Are in Queensland or happy with remote onboarding
✓ Application received! We'll be in touch within one business day.

🔒 Confidential. No spam. No obligation. Pilot limited to 3 practices.