AI + teacher marking: how Braintree Coaching Australia returns calm, precise writing feedback

Braintree Coaching Australia (an Australian exam-preparation provider, not the PayPal payments service) publishes this Calm Data guide so parents can see exactly how hybrid marking works: what the AI does, what a teacher does, and how quickly feedback reaches the LMS. The frame for analytics is precision and calm — specific enough to change the next practice session, calm enough that a score does not feel like a black box. AI is first-pass only and never replaces teachers; a qualified tutor confirms every qualitative judgement before a parent sees it. Families preparing for the NSW Selective High Schools or Opportunity Class pathways can read official placement stages on the NSW Department of Education application process page — Braintree marking turnaround times on this page are our operational SLAs, not Department rules.

How does the blended AI + teacher marking model work?

Quick answer: AI runs a first-pass only and never replaces teachers; a qualified tutor confirms every qualitative judgement before a parent sees it. The teacher remains the final marker.

Layer	What it does	What it does not do
AI first-pass	Flags likely errors, timing patterns, and recurring gaps so a tutor can triage quickly	Deliver the final judgement a parent reads
Qualified tutor	Confirms writing and reasoning quality, rewrites or overrides the automated pass where needed, and releases parent-visible feedback	Leave qualitative decisions to the model alone
Parent LMS view	Shows the tutor-confirmed comments and scores for that paper	Expose raw model output as the source of truth

Where the automated pass and the tutor disagree, the tutor decides. That rule is also stated on our results methodology page; this page adds the parent-facing turnaround times and the plain-English workflow.

What are the marking turnaround times?

Quick answer: AI first-pass and tutor-confirmed parent-visible writing feedback both honour the same ops-attested window: within 24 hours of each paper.

Step	Ops-attested turnaround
AI first-pass	within 24 hours of each paper
Teacher confirmation and parent-visible release	tutor-confirmed before the parent-visible release, within 24 hours of each paper

These turnaround times are ops-attested (operator-signed 2026-07-15; reconfirmed for this page on 2026-09-10 in the E31 service-owner SLA re-attestation). They restate the writing-feedback turnaround Braintree Coaching Australia already publishes as an operational fact — they are not aspirational targets invented for this page.

What does hybrid marking look like on one paper?

Quick answer: You can inspect one labelled chain — original prompt → constructed response → AI first-pass annotations → tutor annotations → revision — so the AI-versus-tutor split is visible on a concrete artefact, not only in a role table.

The block below is a constructed labelled example for parents. It is not a real student's script, not an official NSW Department of Education mark, and not a placement prediction. Where a course includes marked writing, the same order applies in the product: AI first-pass, then tutor confirmation before any parent-visible release.

Stage 1 — Original prompt

Label: Original prompt (Braintree practice writing — constructed example)

Write about a time a small decision changed how a group worked together. Aim for a clear opening, one developed moment, and a closing that shows what the writer learned. This is Braintree practice writing for hybrid marking — not a live Selective Placement Test paper and not a Department score.

Stage 2 — Constructed response (labelled example)

Label: Constructed response (example — not a real student's work)

Our group had to finish a science poster before lunch. I wanted to draw the diagram first, but Maya said we should list the facts first. I almost argued. Instead I wrote the three facts on sticky notes while she sketched boxes. When the bell rang we still had gaps, but everyone knew which box needed one more sentence. The small choice to start with the facts kept us from arguing about colours when we were out of time.

Stage 3 — AI first-pass annotations

Label: AI first-pass annotation (triage only — not the parent-visible source of truth; runs within 24 hours of each paper)

• Likely criterion miss: closing names “what we learned” only lightly — flag for tutor review of reflective close.
• Pace / structure pattern: opening sets the decision quickly; middle develops one moment; ending may under-develop the learning point.
• Draft flag for tutor: check whether “sticky notes / boxes” is concrete enough for the “developed moment” criterion, or whether the learner needs one more sensory detail.

Parents do not see this raw pass as the final comment set.

Stage 4 — Tutor annotations

Label: Tutor annotation (final qualitative judgement — tutor-confirmed before the parent-visible release, within 24 hours of each paper)

• Confirmed: the opening decision and the sticky-note moment are clear enough to keep.
• Override of the AI close flag: ask for one sharper learning sentence, not a full rewrite — the structure already works.
• Parent-visible next move (released only after this confirmation): in the next practice piece, keep the same three-part shape and end with one concrete sentence about what the writer would do again.

Where the automated pass and the tutor disagree, the tutor decides. The tutor mock-review record is the source of truth.

Stage 5 — Revision / next practice move

Label: Revision after tutor confirmation (constructed follow-up — still an example)

Revised closing only: The small choice to start with the facts meant we spent our last minutes fixing the empty box instead of arguing about colours — next time I would write the learning point in the last sentence before we pack up.

That is the Calm Data loop on one artefact: inspect the prompt, see the constructed draft, compare AI first-pass annotation with Tutor annotation, then take one revision into the next session.

Review to next step

After you understand the hybrid chain:

1. See how tutor-confirmed notes appear on a report — readiness report sample (companion surface; study-plan chrome there is a separate product job).
2. Check what a pack includes — pricing and inclusions.
3. Choose a programme path — selective school preparation or Opportunity Class preparation.

What does AI do, and what do teachers do?

Quick answer: AI speeds triage; teachers own qualitative judgement and the words parents read.

AI does

1. Apply the faculty writing rubric as a first-pass screen.
2. Surface likely criterion misses and pace patterns across a large batch.
3. Hand the draft flags to a tutor for confirmation.

Teachers do

1. Read the script with the same rubric the lessons teach.
2. Confirm, edit, or override the automated pass.
3. Release the parent-visible comment set only after that confirmation.

Parents should never need to guess whether a comment came from a model alone. If it appears on the LMS as qualitative feedback, a tutor has confirmed it.

Why frame analytics as precision and calm?

Quick answer: Data is useful when it tells a parent the next practice move — not when it overwhelms a family with charts.

Hybrid marking exists to make feedback precise (which criterion, which habit) and calm (a teacher has already filtered noise). That is the opposite of “tech for its own sake.” Cohort percentiles and section trends still matter for planning; they sit beside teacher-confirmed writing comments rather than replacing them. For how we validate analytics and what we refuse to claim, read the results methodology. For what a course fee includes — including mock reports and teacher-written worked solutions — see pricing and inclusions.

Parent data walkthroughs

Video walkthroughs are not published yet. Until they are, use these three FAQ-equivalent walkthroughs — each answers one parent job in the Calm Data loop.

Walkthrough 1: How does parent-facing data work?

Parent job (how data works): understand what you will see on the LMS after a paper, and what stays behind the tutor filter.

After a timed paper, Braintree Coaching Australia stores the tutor-confirmed comments and scores on the LMS. Parents see the release that a teacher has already filtered — not a live feed of every automated flag. Section trends and recurring mistakes may appear beside writing comments so the next practice session has a clear focus. Raw model output is not the parent-facing record. Validation details stay on the results methodology page; this walkthrough is the parent workflow only.

Walkthrough 2: How does marked feedback arrive after a paper?

Parent job (how feedback arrives): know when marked feedback lands and who signed it off before you see it.

1. Your child submits a paper on the LMS.
2. The AI first-pass runs within 24 hours of each paper.
3. A qualified tutor confirms, edits, or overrides that pass.
4. Parent-visible writing feedback is released only when it is tutor-confirmed before the parent-visible release, within 24 hours of each paper.

You should not need to chase a black-box score. If qualitative feedback is visible, a tutor has already confirmed it. Those turnaround strings match the ops attestation on this hub; they are operational facts, not targets.

Walkthrough 3: How do I act on the feedback without overwhelming my child?

Parent job (how to act on it): turn one insight into the next practice move without crowding sleep or rest.

1. Read the tutor comment set for the paper — one or two specific habits, not every chart.
2. Pick a single practice change for the next session (for example, one writing criterion or one pacing habit).
3. Keep the session inside the study caps in our preparation philosophy. If feedback starts crowding sleep or rest, the plan is wrong — shrink the load, do not add tools.

The Calm Data loop is insight → one change → next paper. A fortnightly plain-English parent snapshot (ops/nurture handoff) restates progress in the same spirit: short, specific, calm — not a dashboard dump.

Data → improvement case examples

These two anonymised examples show how a specific analytics insight from hybrid marking led to one practice change and a measurable readiness or confidence lift. No parent or child is named, no school identifier appears, and no raw score dump is published. Each traces to a cleared row in our first-hand data inventory. They are programme observations — not placement guarantees.

Case 1 — Format slips flagged, mock schedule held, readiness steadied

Analytics insight. In the 2025 Opportunity Class cohort Braintree Coaching Australia coached through the May 2025 sitting (NSW DoE Opportunity Class placement window for Year 5 entry the following year — see also the Department placement test overview), early timed-paper reviews (AI first-pass triage, then tutor confirmation) kept flagging the same pattern for mid-range students: recurring on-screen format slips that cost easy marks and confidence under the clock.

Practice change. Those families kept the full ten-to-fourteen timed mock schedule and treated each marked paper as one focused correction cycle — format habits and pace — rather than adding new tools or longer sessions.

Lift. By test week, tutor follow-up recorded fewer on-screen format surprises and steadier pace than for families who ran only one or two full mocks. That is a readiness and confidence shift across the group, not a placement promise for any child.

A de-identified 2025 Opportunity Class programme observation. No student is named.

Case 2 — Mock analytics pointed to fortnightly full papers

Analytics insight. For the 2025 NSW Selective coaching cohort (Year 6 students through the May 2025 sitting — computer-based at Department centres per the NSW placement test page), tutor follow-up notes tied readiness before outcomes day to one prep pattern more often than any other: whether the family had kept fortnightly full computer-based mocks in the final term — visible in the LMS mock cadence and tutor-confirmed write-ups after each paper.

Practice change. Families that were slipping into one-off or section-only practice switched back to a fortnightly full timed paper, with the hybrid marking release used to pick a single next habit (pace, section focus, or writing criterion) instead of replaying every chart.

Lift. Before outcomes day, tutors most often credited that fortnightly full-mock habit in parent check-ins as the step that lifted confidence and felt readiness. Again, this is an observed preparation pattern — not a claim that any child received a selective place.

A de-identified 2025 NSW Selective programme observation. No student is named.

Related resources

• Readiness report sample — how tutor-confirmed notes appear after a paper
• Results methodology — how analytics and marking are validated
• Pricing and inclusions — what the course fee covers
• Preparation philosophy — study caps and sleep guardrails
• Results and outcomes — anonymised findings, not placement hype
• Selective school preparation — NSW Selective prep hub
• Opportunity Class preparation — OC prep hub

Last updated: 2026-09-10. Reviewed by the Braintree Academic Panel. Marking turnaround figures are ops-attested (2026-07-15; E31 service-owner re-attestation 2026-09-10); they do not predict exam placement or treat AI as the final teacher judgement. Official NSW Selective / OC placement context: application process, placement test, Opportunity Classes.