Step 3

A complete example, in the structure assessors expect

Situation, Action, Outcome — drafted from the answers you gave, inside your regulator's character limit, and honest about everything it does not know.

  • The character limit is measured server-side, and a breach triggers a compression retry
  • Unknown facts arrive as [placeholders] you fill, never as guesses
  • Streams into the real editor, not a temporary preview
Step 3 · Preliminary draft

Situation / Action / Outcome

Writing…

Situation

Action

Outcome

3 placeholders to fill

It didn't know the station name, the standard, or your measured clearance — so it didn't guess. That's the point.

What the draft is, and what it is not

The draft is a first pass built entirely from two inputs: the work experience you selected and the answers you gave to the gap questions. It arrives as a full Situation–Action–Outcome — the structure Canadian regulators ask for — rather than as an outline you then have to expand.

It is not a finished submission, and the product never pretends otherwise. It is the version you edit. What it saves you is the part almost nobody enjoys: turning a set of facts into a structured, appropriately weighted piece of writing that fits a hard character budget.

That budget is real. The model is given a character-counting tool while it writes, the server measures the result independently, and a draft that breaches the limit triggers an automatic compression retry rather than landing in your editor 300 characters over.

How the draft is built

  1. 1

    Your answers become the source material

    Nothing enters the draft that did not come from your recorded experience or your answers — that is the boundary the whole product is built around.

  2. 2

    Structure is applied, weighted by the competency

    Situation gets enough context to make the decision legible; Action carries the weight, because that is what is being assessed; Outcome states what changed and how it was measured.

  3. 3

    Unknowns become placeholders

    A standard you did not name, a metric you did not give, a date you did not include — each becomes a bracketed placeholder rather than a plausible invention.

  4. 4

    The limit is enforced before you see it

    Server-side measurement against your regulator's per-competency character count, with a compression pass if the first draft runs long.

The difference between this and a general-purpose chatbot

It will not make it up

Ask a general chatbot for a competency example and it will produce a fluent one, complete with a standard number and a percentage improvement it invented. You would be the one signing that.

It knows your character limit

Regulators cap each example. A draft that ignores the cap is not a draft, it is homework — and trimming 400 characters by hand is where good sentences go to die.

It knows which framework it is writing for

The competency's own description, indicators and reviewer expectations are in front of the model as it writes, so the example is aimed at what that competency asks for.

The stream does not move the page

Every stream pins the viewport to the top of its step, at start and at end, so text grows below the fold rather than dragging the page under your cursor.

Why competency-based assessment examples get rejected

Four failure modes account for most rejected competency examples, and every one of them is a writing problem sitting on top of perfectly adequate experience. The first is the job description: a paragraph about what the role involved rather than a specific dated event. The second is the invisible applicant: 'we designed', 'the team decided', with no separable individual contribution. The third is missing technical depth — the what without the why or the how. The fourth is the unevidenced outcome: 'the project was successful' with nothing measured.

Situation–Action–Outcome exists to prevent exactly these. Situation forces a specific event with a date and a scope. Action forces first-person decisions. Outcome forces a result you can point at. The structure is not stylistic preference; it is the shape of the evidence assessors need.

Placeholders are the same discipline applied to facts. An example that says the alignment cleared the clash review with [measured clearance] to spare is an honest draft with one thing left to do. An example that says it cleared with 45 mm to spare, when the model chose 45, is a submission with a fabricated number in it.

Trusted by engineers across Canada

M

M. Ibrahim

This was a wonderful experience. Hossein and Jenny were very helpful and prompt to reply to my emails and questions. Hossein got on a call with me before enrolling just to answer my questions. The review for my CBA questions was detailed, thorough with clear direction on what needs to be done to improve, which made it very easy for me to fix my answers. They gave me a time for looking into my CBA questions. They ended up delivering earlier than promised. Finally, they are very fairly priced. Definitely recommend their service for anyone looking to get someone to assess their CBA answers.

1 week ago

Draft & placeholders — questions applicants ask

No. The non-invention rule is a hard constraint in the prompt layer: anything the model cannot derive from your recorded experience and your answers comes back as a bracketed placeholder. It will not invent a standard, a clause, a metric, a date, an employer or an outcome.

Bracketed gaps in the draft text — [applicable standard], [measured clearance] — that mark facts only you have. A wizard walks you through them one at a time in step 5, highlighting the active one in the editor and marking each filled one green.

Yes. The limit comes from your regulator, is enforced while the model writes, and is checked again on the server. If a draft comes back over the limit, it is compressed and re-checked rather than handed to you.

Yes, and many applicants do — the draft is a starting structure, not a final answer. Everything downstream (the reviewer pass, the grade, the self-assessment score) works on whatever text is in the editor, including text you wrote yourself.

In short

CBA Pro's step 3 generates a complete Situation–Action–Outcome competency example from the applicant's selected work experience and their answers to the gap questions, within their regulator's per-competency character limit (measured while writing and re-checked server-side, with an automatic compression retry on a breach). Any fact the AI cannot derive from the applicant's own inputs is emitted as a bracketed placeholder for the applicant to fill — it never invents standards, clauses, metrics, dates, employers or outcomes.

Run one full competency, free

Every step, on a real competency, with no card. Everything you write carries over the moment you upgrade.