Every claim, sourced.

AIRA stands for AI Risk Assessment. It's a web app that does two things:

• A free 6-minute assessment that produces your AIRA Score (0-100) and tier (Stabilize / Develop / Intervene / Redline). The score reflects how exposed your specific role is to AI-driven displacement, calibrated against industry benchmarks.

• A paid AIRA Kit — a personalised 30/60/90-day program built from your assessment data. The Kit gives you a structured action plan, role-specific AI tool recommendations, an automation playbook, and a quarterly check-in cadence.

The score is computed deterministically from your answers using a rule-based formula across four dimensions: Automation Exposure, Judgment Protection, AI Fluency Gap, and Adaptation Capacity. No AI is used to compute your free score — the math is transparent and reproducible.

The Kit is generated by Claude Opus 4.7 from your assessment answers, with strict structural constraints to ensure consistency. Tool recommendations come from a curated, monthly-reviewed allowlist. Industry benchmarks are seeded from public authoritative sources (WEF Future of Jobs 2025, Jobs and Skills Australia, US BLS, O*NET).

AIRA is an independent product, built and run in Australia.

AIRA is built for working professionals who want to understand and respond to AI's impact on their career — without sifting through noise.

Best fit:

• Knowledge workers (marketing, ops, analyst, accounting, consulting, HR, legal) feeling pressure from AI-augmented peers
• Mid-career professionals (5-20 years experience) where role compression is a real concern
• Solo operators (freelancers, consultants) wondering how AI changes their offer
• Career-changers wanting a structured read on which target roles are AI-resilient

Less fit:

• AI engineers, ML researchers, or anyone already deep in AI work — your context exceeds what AIRA covers
• Users under 16 (AIRA has an age gate per Australian Privacy Principles + GDPR-K alignment)
• Anyone seeking financial, legal, or medical advice — AIRA covers career direction, not those domains

AIRA works regardless of country, but benchmarks and recommendations are most accurate for Australia, US, UK, EU, Canada, and NZ. Other regions get the global fallback benchmark set.

The product assumes you have at least basic web literacy but does NOT assume you've used AI tools beyond ChatGPT in the browser. Beginners and intermediates are the primary target.

You can ask ChatGPT "will AI replace my job?" right now, for free. Why use AIRA?

Three structural differences

1. Structured, role-specific output instead of free-form chat

ChatGPT gives you whatever you prompt for. AIRA's output is a structured Kit: 30/60/90-day program, specific tools matched to your role, weekly milestones, ROI math, and a measurable progress framework. The structure is the value — open-ended chat doesn't produce action plans you actually execute.

2. Calibrated benchmarks, not assertions

When AIRA says "your industry average is 65 vs your 67," that number comes from public authoritative sources (WEF Future of Jobs 2025, Jobs and Skills Australia, US BLS, O*NET) cited in your Kit. ChatGPT will produce a number when you ask but won't cite a defensible source — and the number changes each time you ask.

3. Curated tool stack instead of "try everything"

AIRA recommends from a curated, monthly-reviewed allowlist of AI tools that actually work for your specific role category. ChatGPT will list 50 tools; AIRA picks 5-7 you should actually use. We accept the curation cost (monthly reviews) so you don't waste time chasing pivoting tools.

What AIRA is NOT

• A replacement for ChatGPT or Claude in your daily work
• A career coach in any traditional sense
• A guarantee that you won't be displaced by AI
• A subscription product (one-time purchase, no recurring billing)
• An AI training course (the Kit recommends learning resources but isn't a curriculum)

When ChatGPT is enough

If you already know exactly what AI tools fit your role, have a clear weekly automation target, and just want a sounding board — ChatGPT is fine. AIRA is for people who DON'T have that clarity yet and want a structured starting point.

Honest answer: trust it for direction, not as absolute prediction.

What the score is

A deterministic, rule-based read of your self-reported answers across four dimensions:

• Automation Exposure
• Judgment Protection
• AI Fluency Gap
• Adaptation Capacity

Same answers always produce the same score. The math is transparent (see [how-the-score-is-calculated.md]). No machine learning, no LLM, no opacity.

What the score is NOT

• A predictive forecast of whether you'll keep your job
• A measure of your skill or competence
• A verdict on your career
• Calibrated against actual displacement outcomes (no AIRA user has been displaced because of an AIRA score)

Why we still trust it for direction

• The four dimensions map to risk factors documented in academic + industry research (Frey-Osborne, McKinsey, WEF Future of Jobs)
• Industry modifiers come from public authoritative data (WEF 2025, JSA 2025, BLS, O*NET, PwC)
• Tier cutoffs are equal quartiles of the 0-100 scale — there is no weighting toward scarier tiers
• A fully neutral set of answers lands in Develop ("you have a window to adapt"), not in an alarm tier

Why your score might feel "off"

Common reasons:

• You answered conservatively — Likerts at the middle of the scale don't capture nuance well
• You haven't filled Step 4 — free-text doesn't change the score but enriches the Kit; without it, the picture is partial
• Your role is non-standard — the score uses role-category averages; an unusual hybrid role may sit between two categories
• The benchmark is off for your specific situation — public data is industry-level, not company-level

What if you disagree

• Retake with more honest Likerts — first retake is free
• Add Step 4 free-text for a richer Kit
• Email support@useairanow.com if you think the methodology missed something specific — we read every email

The bottom line

The score is a structured snapshot, not a forecast. It exists to prompt a considered response, not to alarm you. If your score feels wrong in either direction, retake with more honest answers — the math follows your inputs.

AIRA shows industry averages and safe-threshold comparisons alongside your AIRA Score. Those numbers are AIRA model estimates, informed by public automation data — not from AIRA user data, and not figures extracted directly from any one source below (none of them measure a bespoke 0-100 "AIRA Score").

Our sources

• WEF — World Economic Forum *Future of Jobs Report 2025*: technology-adoption and role-displacement outlooks
• JSA — Jobs and Skills Australia *Gen AI Capacity Study 2025*: occupation-level AI exposure data for the Australian labour market
• BLS — US Bureau of Labor Statistics: occupational employment projections and task-level data
• **O*NET** — the US Department of Labor's occupational database: task composition, skills, and work-activity profiles per role
• PwC, Frey & Osborne, McKinsey — supporting automation-risk research (full citations on request)

How we use them

• Industry averages are AIRA model estimates, informed by role- and industry-level automation exposure findings in these datasets, mapped to our role taxonomy — a person built the estimate from the pattern in the research, it was not read off any single table
• Safe thresholds are estimated the same way, reflecting where roles with durable, judgment-heavy task profiles tend to cluster in the same research

These benchmarks contextualise your score — they answer "compared to what?" They do not change your AIRA Score, which is computed only from your own assessment answers.

What we don't use

• Not AIRA user data. We don't show "cohort averages" or pool other users' scores into your benchmarks.
• Not AI-invented numbers. Benchmark figures are estimated by AIRA from the datasets above, not generated by a language model.

Limitations, honestly

These are v1 estimates, not empirically validated against real outcomes yet — that validation is scheduled before launch. Public datasets also lag reality — occupational data is typically 1-3 years behind the frontier of AI capability. We treat benchmarks as directional context, not precision instruments, and we update our estimates as new releases of these datasets ship and as real user-outcome data becomes available.

Questions about a specific benchmark on your score page? Email support@useairanow.com.

No. The four tier bands are equal quarters of the 0-100 score range, with no thumb on the scale toward any particular tier.

The tier cutoffs

| Score | Tier |
| 0-25 | Stabilize |
| 26-50 | Develop |
| 51-75 | Intervene |
| 76-100 | Redline |

Each band is exactly 25 points wide. There's no adjustment that pushes more people into the higher-risk tiers — it's a straightforward, symmetric split of the 0-100 scale.

Where does a typical person land?

A fully neutral respondent — middle-of-the-scale answers across every Likert, mid-range resources — scores 48, which places them in Develop: "you have a window to adapt, use it deliberately." That's the honest answer for a profile that reports nothing unusual. See [how-the-score-is-calculated.md] for the full formula.

So why does AIRA talk about urgency at all?

Where you land is a direct readout of your own answers, not a design choice about how many people should end up in each tier. If your score places you in Intervene or Redline, that reflects the automation exposure, judgment protection, AI fluency gap, and adaptation capacity you actually reported — see [can-i-trust-the-score.md] for how those four dimensions are weighted.

Will the bands change?

They're locked for v1 as equal quartiles. Post-launch, AIRA monitors the real distribution of scores and may recalibrate the cutoffs to better fit what the actual user population looks like. There's no pre-set target share for any tier — any future change would be a response to real data, not an adjustment toward a desired outcome.

Your AIRA Score reflects your answers at the time you took the assessment. If your answers changed, your score changes with them.

Common reasons your score moves

Your work pattern actually shifted. More AI-assisted tasks, a role change, or new tools in your workflow all change your inputs — and the score follows.

You answered more accurately the second time. First-time takers sometimes under- or over-estimate their exposure. A retake with more honest self-assessment produces a more accurate score, not a "wrong" one.

Your industry or role classification changed. If you changed jobs or reclassified your industry between assessments, the benchmark comparison changes too.

Small answer changes near a tier boundary. If your score sits close to a cutoff (24/44/64), a modest shift in one or two answers can tip you into a different tier label even though your underlying risk barely moved.

What does NOT change your score

• Time passing alone, with no change in your answers
• Retaking without changing anything (you'll get the same score)
• Random variation — AIRA's scoring is rule-based and deterministic, not probabilistic. See [can-i-trust-the-score.md].

Is the scoring stable?

Yes. Given identical answers, AIRA always produces the identical score. There's no randomness in the calculation — the same inputs always map to the same output.

Tracking your trajectory

Your account keeps every past assessment. You can see your score history over time, which is more useful than any single score — it shows whether your AI exposure is trending up or down as your habits and role evolve.

See [retake-policy.md] for retake costs and cooldown, and [when-to-retake.md] for guidance on timing.

A retake is a fresh submission of the AIRA Assessment. It produces a new AIRA Score so you can track your trajectory. (Your first retake is free; see the retake policy for pricing after that.)

Good times to retake

• After completing a Kit phase. Finished the 30-day block of your 30/60/90 Roadmap? Retake to see whether the work moved your score.
• Quarterly. The Kit's own cadence is a quarterly reassess — pairing a retake with each Full Kit regeneration keeps your Kit built on current data.
• After a job change. New role, new industry, new company size — your Automation Exposure and Judgment Protection can shift materially. Retake so your score reflects your actual situation.
• After a significant skills milestone. Shipped a real AI-augmented workflow at work, completed serious upskilling — anything that changes your honest answers to the AI Fluency Gap questions.

When a retake is wasted

• Days after your last one. Nothing real has changed; you'll mostly measure noise in your own self-reports.
• To "game" a better number. Answering aspirationally instead of honestly gets you a prettier score and a worse Kit — the Kit is generated from your answers.
• Right before a regeneration, out of habit. Only retake first if something actually changed; otherwise regenerate from your existing data.

Retakes and your Kit

A retake does not automatically regenerate your Kit. Your Kit stays as-is until you regenerate it (Full Kit buyers: 4 regens per year). The recommended rhythm: retake → then regenerate, so the new Kit uses your latest Assessment Result.

Rule of thumb

Retake when your situation or skills have genuinely changed, or on your quarterly cycle — whichever comes first.

Your Kit is an interactive web page in your account, generated specifically from your assessment data. Sections vary slightly between Starter and Full — see [starter-vs-full-kit.md] for the comparison.

Common sections (both tiers)

Personal Profile — your role, industry, employment status, assessment date, and AIRA Score with tier label.

Score Brief — 2-3 paragraphs explaining why your score landed where it did, with reference to the dimensions and benchmarks that drove it. Includes time horizon, urgency level, market signal for your industry, adaptation window, and the cost of inaction.

Immediate Actions — 3 ranked actions you can take this week. Specific, role-relevant, and short.

Closing — short closing motto + your next concrete action.

AIRA Starter additional sections

Top 5 AI Tools for your role — curated picks from our monthly-reviewed allowlist, matched to your role category. Each with: what it does, why it fits your role, one specific weekly use case to start with.

30-Day Plan (Lite) — 4 weeks of high-level goals, with estimated weekly hours saved.

Basic ROI Projection — single-tier estimate based on your stated weekly time commitment.

AIRA Kit (Full) additional sections

Key Vulnerabilities + Strengths/Resilience — 3-4 specific risk drivers + 3-4 protective factors for your role.

Risk Breakdown by Task Category — 5 categories (Admin / Analysis / Communication / Creativity / Strategy) with risk percentage and per-category explanation.

30-Day Plan (Detailed) — week-by-week guidance: tools to use, acceptance criteria, artifacts produced. Includes the BRIDGE prioritisation framework for picking your first automation pilot.

Curated AI Resources — 5-7 authoritative resources from sources like OpenAI, Anthropic, Microsoft, DeepMind, Hugging Face. Each with role-specific application notes.

ROI Projection (3-tier) — projections at 2hr/week, 5hr/week, and 10hr/week commitment levels, with personalised note based on your stated budget and time.

Daily Discipline — daily, weekly, monthly, and quarterly habit tracker with measurable metrics and a closing motto.

90-Day Roadmap — Days 31-60 (Depth & Specialisation) and Days 61-90 (Strategic Application & Leadership), plus a 90-day milestone.

Skills to Master — 3-5 named skills relevant to your role.

Format

The Kit lives in your account as an interactive web page. PDF export is not currently supported. You can use your browser's print-to-PDF if you want a hard copy.

Two paid options. Both are one-time purchases — no subscription.

AIRA Starter — $14 AUD

A focused 30-day starter program for users who want to act on their score without committing to the full program.

Includes:

• Your AIRA Score + tier with personalised rationale
• 3 immediate actions ranked for impact
• Top 5 AI tools recommended for your specific role
• 30-day action plan (4 weeks, light structure)
• Basic ROI projection (estimated hours saved + annual value)
• Interactive web Kit accessible from your account

Does NOT include:

• 60/90-day roadmap
• Task-category risk breakdown
• Daily Discipline tracker
• Curated authoritative resources (OpenAI, Anthropic, MS, etc.)
• Quarterly Kit regenerations
• First retake included

AIRA Kit (Full) — $39 AUD ⭐ Recommended

The complete program. What every assessment is designed to deliver.

Includes everything in Starter, plus:

• Full 30/60/90-day program with detailed weekly guidance
• Task-category risk breakdown (Admin / Analysis / Communication / Creativity / Strategy)
• Key vulnerabilities + strengths analysis
• Daily Discipline tracker (daily / weekly / monthly / quarterly habits)
• Curated AI resources from authoritative sources (OpenAI, Anthropic, Microsoft, DeepMind, Hugging Face)
• Full ROI projection across 3 commitment scenarios
• 90-day milestone roadmap with skills-to-master list
• Quarterly Kit regenerations for 1 year (4 fresh Kits as your situation evolves)
• First retake included free
• BRIDGE scoring framework for prioritising automation projects

Upgrade path

Already bought Starter? Upgrade to Full anytime for $25 (total = same as buying Full direct, $39). No penalty for starting smaller.

Which should you pick?

• Lower-tier scores (Stabilize / Develop): AIRA Starter is enough to act on; you have time
• Higher-tier scores (Intervene / Redline): AIRA Kit (Full) — you need the complete playbook
• Unsure / want to try first: Starter, upgrade later if it lands

The Full Kit is what AIRA was built around. Starter is a downsell, not the hero product.

Yes — and we're upfront about it.

What generates your Kit

Your AIRA Kit is generated by Claude Opus 4.7 (Anthropic's frontier model), working from:

• Your Assessment Result — the normalized answers from your AIRA Assessment
• Your AIRA Score and Tier, computed by our deterministic scoring function (not AI)
• A human-authored system of templates, rubrics, and a curated tool allowlist

The AI writes the personalised content; the structure, quality bars, and recommendations framework are human-designed.

What "prompt caching" means for you

We use prompt caching — an Anthropic API feature that keeps the system prompt, scoring rubric, and role taxonomy identical and cached across generations. Practically, this means:

• Consistency: every Kit is generated against the same locked instructions, not a drifting prompt
• Quality control: when we improve the templates, the improvement applies uniformly to all future generations

Human editorial oversight

The templates, tier language, tool allowlist, and roadmap structures are written and maintained by a human. The AI fills in your specifics; it doesn't invent the methodology. Named tool recommendations come from a curated allowlist that we update — not from the model's memory.

Is the AIRA Score AI-generated too?

No. Your AIRA Score is computed by a rule-based, deterministic scoring function — same inputs, same score, every time. Only the Kit's written content is AI-generated. See "Can I trust the score?" for details.

What about variability?

AI-generated content has reasonable variability — two generations from the same data may phrase things differently. The substance (your Dimensions, tier, recommendations) stays anchored to your data. If a generation contains broken or materially mismatched content, contact support@useairanow.com and we'll regenerate it at no cost.

Only you.

How this is enforced

Your Kit is protected by Row-Level Security (RLS) at the database level, not just by app logic. Every query for Kit data is scoped to your authenticated account — there's no URL or link that exposes your Kit to anyone else, even if they guessed it.

Can I share my Kit with someone (manager, coach, friend)?

Not currently. There's no share link or export-to-share feature at launch. If you want someone else to see your results, your options are:

• Show them your screen
• Use browser print-to-PDF (see [kit-format-faq.md]) and send the file yourself

Does AIRA staff see my Kit?

AIRA can access records for support and debugging purposes — for example, to investigate a failed Kit generation you've reported. Routine access to individual Kits for any other reason doesn't happen.

Does my data train AI models?

No. Your assessment data is sent to Anthropic to generate your Kit, but Anthropic's policy is that API data is not used to train models. It's retained by Anthropic for 30 days for abuse monitoring, then deleted. See [privacy-policy-summary.md] and [data-usage.md] for the full data flow.

What about anonymous assessments?

If you took the assessment without an account, your data is tied to an anonymous session token, not an identity — until you sign in and it migrates to your account. See [anonymous-to-account.md].

All prices in AUD (Australian dollars). LemonSqueezy displays your local currency at checkout.

Products

| Product | Price | One-time? |
| AIRA Assessment + Score | **Free** | Yes |
| First retake | **Free** | Yes (one per user) |
| AIRA Starter | **$14 AUD** | One-time |
| AIRA Kit (Full) | **$39 AUD** | One-time |
| Starter → Full upgrade | **$25 AUD** | One-time, brings total to $39 |
| Additional retake | **$5 AUD** each | Per retake after the free ones |

No subscriptions. No recurring billing. Pay once for a Kit, regenerate quarterly for 12 months (Full only), then optionally rebuy.

What's free vs what costs money

Free:

• Take the AIRA Assessment (6 minutes)
• See your AIRA Score and tier
• See abbreviated rationale for your score
• See the immediate-actions list (3 items)
• One free retake to track score changes over time

Paid:

• The full Kit content (action plan, tools, ROI projection, etc.)
• All retakes after the first free one ($5 each)
• Kit regenerations (Full Kit buyers get 4/year free for 12 months)

Refunds

We don't advertise a refund policy because we don't want to invite refund-seeking. But you have statutory consumer rights under Australian Consumer Law that we honour. If your Kit is faulty, doesn't match the description, or isn't fit for purpose, contact support@useairanow.com — we'll work it out.

Tax + currency

LemonSqueezy is the Merchant of Record. They collect and remit GST, VAT, and US sales tax on our behalf. Your local taxes are handled at checkout — no surprise bills.

We don't advertise a refund policy because we don't want to invite refund-seeking — but you have statutory consumer rights and we honour them.

Australian Consumer Law (ACL)

Under the Competition and Consumer Act 2010 (Schedule 2), digital products purchased in Australia carry statutory guarantees:

• The product must be fit for the purpose described
• The product must match the description provided
• The product must be of acceptable quality (recognising that AI-generated content has reasonable variability)

If your AIRA Kit fails any of these — for example, the Kit doesn't generate, generates broken content, or contains content materially mismatched with your assessment data — contact support@useairanow.com and we'll work it out.

What we'll typically do

In order of preference:

• Regenerate the Kit — fastest fix; you get a fresh Kit at no cost
• Manual override — AIRA writes your Kit content directly if regeneration also fails (rare)
• Refund — full purchase amount back to your original payment method via LemonSqueezy

We aim to resolve within 24 business hours.

What we generally won't refund

• Buyer's remorse alone ("I changed my mind") — Kit is delivered digital good
• Disagreement with the score's tone or framing — the urgency tone is intentional and disclosed
• Content the Kit didn't include because you skipped Step 4 (Additional Details) of the assessment

How to request

Email support@useairanow.com with:

• Your AIRA account email
• Order ID from your LemonSqueezy receipt
• A short description of the issue

We respond within 24 business hours during AU work week (Mon-Fri 9am-5pm AEST). Outside those hours, expect next-business-day response.

Chargebacks

If you initiate a chargeback through your bank without contacting us first, we lose the LS dispute fee plus the original sale. Please email us first — we're a solo team and chargebacks are disproportionately costly.

What we collect

• Email address — when you save your score or purchase
• Assessment answers — your role, industry, employment status, work-pattern responses, and any optional free-text answers
• Purchase data — order ID, amount, and SKU from LemonSqueezy; LemonSqueezy itself holds your card details, not AIRA
• IP address — temporarily, for rate limiting and bot protection; not retained past 30 days
• Usage events — anonymised product analytics (page views, funnel steps) via PostHog

What we don't collect

• Your card number (LemonSqueezy handles this entirely)
• Browsing activity outside AIRA
• Social media account data
• Government ID numbers
• Your phone number, unless you choose to provide one for account recovery

Who your data is shared with

• Anthropic — your assessment data is sent to generate your Kit. Not used to train models; retained 30 days for abuse monitoring, then deleted.
• Supabase — hosts your data, Sydney region (ap-southeast-2)
• LemonSqueezy — payment processing and tax handling
• PostHog — anonymised usage analytics, EU Cloud region

We have Data Processing Agreements in place with each.

How your assessment answers are used

Your answers feed two things: your AIRA Score (calculated locally, rule-based, no AI involved — see [how-the-score-is-calculated.md]) and, if you purchase, your Kit content (generated by Claude based on your answers and score).

How long data is kept

• Assessment answers, Kits, and chat history: until you delete your account
• Purchase records: 7 years post-deletion for AU tax law, with personal details stripped
• Logs (errors, cron jobs, webhooks): 30-90 days

Full detail: [privacy-policy-summary.md]. To remove your data entirely: [delete-my-account.md].

AIRA is sold to Australian consumers, so Australian Consumer Law (ACL) statutory guarantees apply to your purchase regardless of AIRA's own refund policy wording.

What ACL guarantees

When you buy a digital product like the AIRA Kit, ACL guarantees it will:

• Be provided with acceptable care and skill
• Be fit for the purpose it's sold for
• Match its description
• Be delivered within a reasonable time

What this means in practice

If your Kit fails to generate, is significantly different from what was described, or otherwise doesn't do what AIRA says it does, you're entitled to a remedy under ACL — this exists independently of anything in AIRA's own refund policy (see [refund-policy.md]).

If something goes wrong

AIRA's locked process: if Kit generation fails after an automatic retry, you're notified automatically, the founder is alerted, and your Kit is delivered manually within 24 business hours. You keep full purchase entitlement (your retake and regen quotas start counting from delivery, not from the original payment) — you're never short-changed by a delivery delay.

These rights can't be excluded

ACL guarantees can't be excluded, restricted, or modified by any terms of service or refund policy — no business, including AIRA, can contract out of them.

Where to raise an ACL issue

Email support@useairanow.com. If you believe your ACL rights haven't been honoured after contacting us directly, the ACCC (Australian Competition and Consumer Commission) is the relevant regulator for further escalation.

See also [terms-of-service-summary.md] and [refund-policy.md].