[01]For Australian mortgage brokers

Your fact-find,
written while you talk

SnapScript listens to the client conversation and fills your Best Interests Duty fields as they are spoken — requirements and objectives, financial situation, products considered, the reason for your recommendation. You finish the appointment with the file already written.

No credit card · works on phone, Zoom or in the room

Recording · fact-find
0/6 captured
Client · live

“…we want to be out of the rental by March, so settlement can't drag.”

  • requirements_objectiveslistening…
  • financial_situation—
  • vulnerability_flags—
  • products_considered—
  • recommendation_reason—
  • exit_strategy—
Structured JSON · webhook or API→ your CRM
[02]Capability

Built for the appointment, not the write-up

Note-takers summarise a meeting. SnapScript returns the specific fields you are obliged to evidence, in the same structure every time.

[01]

Your fields, not someone's template

Name the fields your licensee actually reviews — requirements and objectives, financial situation, products considered, exit strategy — and every appointment comes back in that exact shape. Change the schema once and the whole office follows it.

[02]

Filled by the time you shake hands

Fields resolve during the conversation, not after it. No evening spent reconstructing a 45-minute appointment from three lines of shorthand.

[03]

In the room, on the phone, on Zoom

Capture from your microphone or straight from system audio, so a kitchen-table appointment and a Teams call run down the same path.

[04]

Knows who said what

Speaker separation is on by default. When both applicants are talking over each other, the file still reflects which one set the constraint.

[05]

Their language, your file

Language is detected automatically — you don't declare it up front. The conversation can happen in Mandarin or Arabic and the fields still come back in the structure you defined.

[06]

Ends up in your CRM, not a document

Structured JSON out by webhook or API. The point is the data landing where your process already lives — not another note to copy across by hand.

[03]Provenance

Every field traces back to what the client said

The gap in every other broker AI tool: a reviewer can see the finished note, but not where it came from. SnapScript keeps the conversation and the extraction against the same session, so the answer to “why is this in the file?” is one click, not a memory test.

  1. [01]Source

    What was said

    The full transcript is kept against the session, with speakers separated automatically.

  2. [02]Machine

    What was extracted

    Each field is captured against that same session id, so every value has a line of speech behind it.

  3. [03]Broker

    What you changed

    You review and correct before anything is filed. Your version is the one that goes in — the extraction stays visible underneath it.

  4. [04]File

    What gets filed

    Structured JSON to your CRM by webhook or API, with the conversation it came from still attached.

Field tracerecommendation_reason
Heard at 14:22

“So offset makes sense because the bonus lands in one hit each year, and it just sits there doing nothing until the school fees come out.”

Extracted
{
  "recommendation_reason":
    "Offset suits lumpy annual bonus
     income; funds idle between
     receipt and school fees."
}
Your edit before filing

Added the comparison rate the client was quoted. The original extraction stays attached to the session, so the change is visible rather than silent.

You review and sign off — the file is yours, not the model’s
[04]Before / after

The appointment hasn't changed. The paperwork has

Most brokers already know what belongs in the file. The cost is getting it there twice — once in the room, once again at 9pm.

When the file gets written
Now

That evening, from three lines of shorthand

SnapScript

During the appointment, as it is said

What the fields are
Now

Whatever the template allows

SnapScript

The fields your licensee actually reviews

Where the wording comes from
Now

Your reconstruction, hours later

SnapScript

The client's own words, captured live

If a reviewer asks why
Now

You explain it from memory

SnapScript

Point at the line of speech behind the field

Client's language
Now

English, or you translate as you go

SnapScript

Detected automatically, fields still structured

Where it ends up
Now

Retyped into the CRM

SnapScript

Structured JSON, by webhook or API

[05]Product flow

Three steps, and only one of them is new

You already do step two. SnapScript changes what happens on either side of it.

01Two minutes, once

Set your fields once

Build the fact-find your licensee actually reviews. Keep a separate collection per appointment type — purchase, refinance, construction — so each one returns the right structure.

02Mic or system audio

Have the appointment

Kitchen table, phone, or Teams. Press record and run the conversation you were going to run anyway — no script, no prompting the client, no laptop between you.

03You approve, always

Check it and file it

The fields come back filled, with the transcript behind them. Correct anything the model got wrong, sign off, and push it to your CRM as structured data.

[06]Under the hood

What you actually get back

No black box. Here is the pipeline every appointment runs through, and the exact shape of what lands in your CRM at the other end.

Pipeline06 stages
01 · Transcription
Deepgram nova-2, smart formatting
Punctuation and casing applied as it streams
02 · Speakers
Diarisation on by default
Turns are separated without manual tagging
03 · Languages
Auto-detected, multilingual
No need to declare the language up front
04 · Sources
Microphone or system audio
In-room conversations and calls on the same path
05 · Extraction
Fields you define
Name the fields; every session returns them filled
06 · Delivery
Structured JSON, webhook or API
Lands in your stack, not just in a document
Example schema05 fields

Define the fields once on a collection. Every appointment recorded into it comes back in the same structure — so a file review reads the same whichever broker took the meeting.

  • REQUIREMENTS_OBJECTIVES
    What the client asked for, in their words
  • FINANCIAL_SITUATION
    Income, commitments, dependants, employment
  • PRODUCTS_CONSIDERED
    Options compared, and on what basis
  • RECOMMENDATION_REASON
    Why this product over the alternatives
  • EXIT_STRATEGY
    How the loan ends where term runs past retirement
Returns
{
  "requirements_objectives": "...",
  "financial_situation":     "...",
  "products_considered":     "...",
  "recommendation_reason":   "...",
  "exit_strategy":           "..."
}
[07]Pricing

Less than one hour of your time back

Priced against the evening you currently spend writing up the day’s appointments.

Trial

$0/mo

Run a couple of appointments through it

30 minutes of audio a month
1 collection
5 fields per collection
Email support
Most Popular

Broker

$32/mo

For the broker writing their own files

1,000 minutes a month — roughly 16 hours of appointments
10 collections — one per appointment type
Unlimited fields per collection

Brokerage

$160/mo

One fact-find standard across the office

5,000 minutes a month — around 83 hours
Unlimited collections
Priority support
[08]Start here

Take one appointment
and see what comes back

Set up your fact-find fields, record your next client meeting, and look at the file it hands you. The free tier is enough to find out whether this changes your evenings.

Start free

No credit card required

Powered by Langoedge