Files Slides

DEB JET AI UAT

The evaluation pipeline, step by step

Seven links, run in order. The value of this pipeline is not the list of steps — it is that the order cannot be reversed.

What this pipeline is for

DEB JET has no built-in AI that assesses brokers. So the “AI under test” here is an external LLM (Claude, ChatGPT) acting as a UAT evaluator: it reads the evidence captured from a broker's journey and produces an assessment.

Two consequences follow, and they govern the whole design. First: the AI knows only what we hand it, so packaging the evidence is the core of the work, not a side task. Second: the prompt is the measuring instrument — the same evidence under two different prompts yields two different assessments. The prompt is therefore fixed and stored alongside the result.

Seven links

Research → Standard → Checklist → Scenario → AI Assessment → Comparison → Feedback

The line you do not cross

Step 3 must be finished before step 4. That means writing down what the AI has to say to count as correct before seeing what it says.

Do it the other way round — run first, read the AI's answer, then weigh whether it was right — and you will always find a justification for whatever came out. That is no longer evaluation; it is rationalisation.

Define the evaluation criteria before running the scenarios. Do not decide whether the AI is correct only after seeing its output.

Three verdicts

Each checklist line takes exactly one of three verdicts. The boundaries between them are defined up front, not negotiated while scoring.

PassThe AI reaches the right conclusion and cites at least one specific piece of evidence present in the input.
ReviewRight conclusion, no evidence cited — we cannot tell reasoning from a lucky guess.
FailWrong conclusion, or evidence cited that does not exist in the input. The second is the more serious of the two.

How the pipeline maps to the folder

The brief asked for a folder holding the scenarios and the feedback. Each group of links has its place in it.

FolderLinksContents
00-framework/1 – 3Criteria, checklist, scoring rules, fixed prompts
01-howto/Instructions for others to run it themselves
02-scenarios/4 – 5SC01 strong broker, SC02 weak broker, SC03 system feedback
04-report/6 – 7Comparison table and cross-loan analysis

Why two loans

Running the same persona against both Refinance and Purchase is how the AI's consistency gets tested. But the two loans differ on two levels — as products and by persona — so a rule is needed to avoid the wrong verdict.

The conclusion about the broker must match; the details about the system are allowed to differ. If one run classifies the broker as strong and the other as weak, that is a Fail. If both say strong but surface different friction points, that is correct behaviour — the journeys really are different.

This also strengthens scenario 3: an issue the AI reports on both loans is a system-level problem and a strong P1 candidate; one appearing on a single loan belongs to that flow and ranks lower.

Round results — 17 Sep 2026

What we ran, and what it found

Seven scripted runs, five of them valid — three are withdrawn, created by clicking the wrong button and measuring /loan/ instead of /loanjet/. One AI assessment made in isolation, one audit of that assessment. Full detail in 04-report/comparison.md.

Signalstrong-refiweak-refistrong-jetsmsfweak-jetsmsf
Route/loanjet//loanjet//loanjet//loanjet/
Loan statepopulatedpopulatedblankblank
Next on empty formadvances, no errorsadvances, no errorsdisableddisabled
Field for recommendation reasonpresent, optionalpresent, optionalpresent, optionalpresent, optional
Step marker11/1111/1111/1111/11

Behaviour tracks the state of the loan, not the persona and not the journey. Three populated loans let an empty form through; two newly created ones disable Next. All five runs are on /loanjet/, so the route cannot be the cause.

This conclusion replaces two earlier ones. First we thought weak brokers hit a wall strong ones avoid — wrong, because persona was confounded with route. Then we thought the difference tracked the journey — also wrong, because the two runs added to test that were on a different product. No defect in this package is attributable to broker skill level.

No record of why a product was recommended

P1-1

Rule: leaves the broker without a compliance record. Reproduces on both loans.

input-brokernote exists on all three /loanjet/ runs but required: false; it is absent on both SMSF runs. A broker can pass the product step having recorded no reason at all. Under ASIC RG 273, this is where that evidence would have to live.

Next advances an empty form with no validation

P1-3

Rule: reproduces on both loans.

On all three /loanjet/ runs — both loans, both personas — Next on an untouched product step moved the file forward with errorsShown: [].

SMSF disables Next with no reason shown

P1-4

Rule: blocks completion. Reproduces across both personas.

reasonShownOnScreen: null while the control is disabled. The five errors on screen name other fields entirely.

The current step is never marked

Withdrawn

This was our most strongly evidenced finding — 55 of 55 steps — and it was wrong. The probe looked for the active class on the tab button; the marker lives on the .rc-steps-item ancestor. Re-testing against the live system showed the system marks the step correctly. Kept here rather than deleted, because how it survived every automated check is the lesson.

Validation errors never name the field that was wrong

P2-1

Four successive malformed entries into loan amount — abc!!, 500.000,00, $500k, five hundred thousand — produced an identical five-error list every time, none of which mentions loan amount.

The step marker does not reach assistive technology

P3

aria-current is 0 on every step. The marker is visual only.

Entered data persists correctly

Strength

Pass in every run. Enter an amount, move forward two steps, come back — the value is intact.

The largest lesson of this round

The most serious error was ours, not the AI’s. The AI repeated P1-2 faithfully because the evidence package we handed it said so. No amount of auditing the AI would have caught it — only re-testing the claim against the live system did.

A tool that measures the wrong thing produces consistent, confident, wrong numbers — and consistency reads as proof.

In short

This pipeline exists to enforce a single discipline: write the expected result before you see the answer. Everything else is procedure in service of that.

What it buys is a UAT that is traceable, repeatable and evidence-based — rather than resting on the scorer's subjective impression.