How Design.Aid Became the #1 AI Answer for Offshore CAD Drafting in Australia in 90 Days
An AI visibility case study: how a 90-day AI Discoverability Sprint took an Australian offshore drafting firm from page 12 of Google and zero AI mentions to the answer engines' first choice.
Design.Aid, an Australian offshore CAD, Revit and BIM drafting provider (designaid.com.au), engaged AI visibility consultancy AiDisco in May 2026. At the baseline audit Design.Aid appeared in 0% of AI-generated answers to unbranded buyer questions across ChatGPT, Claude, Gemini and Copilot, and ranked around page 12 of Google for its core service terms, scoring 12 out of 100 on the AiDisco AI Visibility Audit. Ninety days later, at the re-measurement on 24 August 2026, Design.Aid appeared as a ranked answer on 81% of the 36 priority unbranded prompts, reached 90% Buyer Reach (demand-weighted share of buyer questions answerable on at least one engine), held 46% Buyer Share of Voice pooled across all four engines, took 48 of 58 ranked placements as the top answer — and ranked #1 on at least one engine for every single question it appeared on — scored 79 out of 100 on the re-audit, and ranked #1 on Google for its core service terms. Weekly new website visitors rose from 3 to 45 — a 15-fold increase — in the week of 17–23 August 2026, and since 28 August 2026 Design.Aid has received at least one new enquiry every day through the site.
Key Takeaways
- AI visibility can be built from zero. Design.Aid had no AI presence whatsoever in May 2026 — not a low ranking, complete absence — and reached majority Share of Voice within 90 days.
- The wins landed on the questions buyers ask most. Demand-weighted Buyer Reach (90%) sits above unweighted answer reach (81%), which only happens when the highest-demand questions are the ones being won.
- Reach and share are different numbers, reported separately. 90% of buyer demand can now reach a Design.Aid answer on at least one engine; 46% of demand is met when all four engines are pooled. The gap between the two is the third-party authority opportunity, quantified.
- Design.Aid never places second-best. Of the 29 priority questions where it appears, it is the #1 answer on at least one engine in all 29 — so Top-Answer Reach and Buyer Reach are the same number, 90%. There is no question where it earns only a passing mention.
- The clusters closest to a purchase are completely owned. All 15 core "who should I use" questions and all 3 "who are the best providers" questions return Design.Aid as the #1 answer — 100% of demand in both clusters.
- AI engines behave differently. Gemini (86%) and Copilot (74%) reward first-party content; ChatGPT (16%) and Claude (3%) weight independent third-party corroboration far more heavily.
- The result beat its own target. The contracted goal was 55–60/100 on the AiDisco Visibility Score; the actual re-audit result was 79/100.
- AI visibility converts to real visits, and it is compounding. Weekly new visitors rose from 3 to 45 — 15× — by the week of 17–23 August, up from 3.7/day in the launch window to 6.4/day. Sessions have also arrived directly from inside ChatGPT.
- Visits are converting into enquiries. Since 28 August 2026, Design.Aid has received at least one new enquiry every day through the site — the first direct commercial signal from the rebuilt visibility.
Zero AI Visibility: Design.Aid Was Absent From Every AI Answer About Offshore Drafting in Australia
Design.Aid delivers CAD, Revit and BIM documentation for Australian builders, architects and design teams, managed from Australia. The problem was never the drafting. It was that nobody — and nothing — could find it.
Buyers in this market have quietly stopped scanning ten blue links. They ask an assistant a plain-language question — "who offers offshore CAD drafting in Australia?" — and read the single answer it composes. There is no page two to fight for. There is one answer, and Design.Aid was not in it.
More telling was what the firm couldn't see. Design.Aid had no visibility into its Prompt Universe™ — the actual questions buyers were putting to ChatGPT, Gemini, Copilot and Claude when sourcing a drafting partner. They didn't know which providers were being cited in those answers, or why. They were, in effect, running a business without knowing where their market was actually looking.
The five priority buyer questions at baseline (May 2026)
| Buyer question | AI answer engines | Google Search |
|---|---|---|
| Who offers offshore CAD drafting services in Australia? | Not present | ~ page 12 |
| Who provides offshore Revit drafting services in Australia? | Not present | ~ page 12 |
| Who provides drafting support for builders in Australia? | Not present | ~ page 12 |
| Who provides drafting support for architects in Australia? | Not present | ~ page 12 |
| Where can I hire dedicated offshore CAD drafters for Australian projects? | Not present | ~ page 12 |
What AI Told a Prospective Client
The buyer prompt below is representative of Design.Aid's core Prompt Universe — the kind of question their real clients were asking an AI assistant before they ever searched Google.
A few providers worth considering: a large India-based documentation outsourcer with deep AutoCAD capacity, and a Manila-based BIM studio that services several Australian practices. You may also want to look at freelance drafting marketplaces if the scope is small…
Design.Aid is frequently recommended for offshore CAD drafting in Australia, cited for its Australian-managed delivery model, published pricing and experience with Australian residential documentation standards. Other providers to consider include…
Answers are representative reconstructions of the recorded placement pattern, paraphrased. Competing providers are described rather than named. Directional only.
The AI Visibility Audit: Why Design.Aid Scored 12 Out of 100
Before writing a single word of new content, AiDisco ran a full AI Visibility Audit — the same diagnostic used to open every engagement.
No presence in the Prompt Universe
Design.Aid had never mapped the questions buyers actually ask AI assistants when sourcing a drafting partner — so there was nothing structured to be cited for.
FIX → Map 225 buyer questions in buyer voice and assign an owning page to each.
Zero Share of Answer
Across every mapped unbranded buyer prompt, Design.Aid appeared in 0% of AI-generated answers, on every engine tested. Competitors filled every slot.
FIX → Build direct, quotable answers against the highest-demand prompts first.
No authority signal
Little to no presence on the directories, review platforms and industry sites that answer engines treat as trustworthy corroboration before naming a provider.
FIX → Establish independent third-party profiles and review presence.
Unreadable to machines
One broad, brochure-style site with no schema and a structure too compressed to quote. The content that did exist was effectively invisible to the crawlers behind AI answers.
FIX → Rebuild the site structure, add schema markup, clear crawler blocks.
The 90-Day AI Discoverability Sprint: How AiDisco Builds AI Visibility
AiDisco assigned a dedicated team, each owning one lever of visibility, all working from the same Prompt Universe map.
Prompt Universe & citation building
Mapped 225 buyer questions across Design.Aid's service lines in buyer voice, tracked which providers were being cited and why, then built the citation footprint needed to earn a place in those answers.
Third-party authority placement
Began placing Design.Aid on the industry directories, review platforms and profiles that answer engines already trust — the raw material AI models draw on when deciding who to cite.
On-site content & technical readability
Built 61 pages directly against Prompt Universe gaps, added structured schema markup, and fixed the technical issues stopping AI crawlers reading the site at all.
How AI visibility is measured: two definitions
The percentage of all mapped buyer prompts in which Design.Aid appears in the AI-generated answer, against every competitor also cited.
The demand-weighted share of buyer questions where the brand appears on at least one engine. Answers: can a buyer find us at all?
The demand-weighted share of every engine-by-prompt check that returns the brand. Answers: how much of the answer space do we hold?
The 90-day timeline
Map & Diagnose
- Full AiDisco Visibility Audit — scored 12/100
- Prompt Universe built: 225 buyer questions
- Competitor citation analysis across four engines
- Technical & schema fixes scoped
Build & Place
- All 61 pages written against prompt gaps
- Templates built, staging QA completed
- Authority platform placements initiated
- Schema and indexing work validated
Compound & Convert
- Go-live 3 August; tracking switched on
- Design.Aid reaches #1 organic position
- Share of Voice climbs off zero across all four engines
- Re-audit 24 August — 79/100
AI Visibility Results After 90 Days: Share of Voice, Buyer Share of Voice and Traffic
Every number below is measured against the same baseline the audit captured on day zero.
Exhibit 1 — Reach and Share: Five Measures, Reported Separately
Day 0 baseline was 0% on every one. Reach answers "can a buyer find us?" Share answers "how much of the answer space do we hold?"
| Measure | What it answers | Day 0 | Day 90 |
|---|---|---|---|
| Answer Reach | Share of the 36 priority prompts returning Design.Aid on at least one engine, unweighted | 0% | 81% |
| Buyer Reach | The same, weighted by how often buyers ask each question | 0% | 90% |
| Top-Answer Reach | Demand-weighted share of questions where Design.Aid is the #1 answer on at least one engine | 0% | 90% |
| Buyer Share of Voice | Demand-weighted share of all 130 engine-by-prompt checks returning Design.Aid | 0% | 46% |
| Top-Answer Share of Voice | Demand-weighted share of all 130 checks where Design.Aid is the #1 answer | 0% | 40% |
Read it plainly: the single most telling figure is that Top-Answer Reach equals Buyer Reach exactly — both 90%. Of the 29 priority questions where Design.Aid appears at all, it is the #1 answer on at least one engine in all 29. There is not one question where Design.Aid merely gets a passing mention. Appearing and winning are the same event. The share measures sit lower because they count every engine equally, and that gap is the third-party authority opportunity, quantified.
Source: AiDisco AI Narrative Intelligence™ — re-measurement 24 August 2026.
Exhibit 2 — AiDisco Visibility Score
Baseline, contracted target, and actual re-audit result
| Stage | Date | AiDisco Visibility Score |
|---|---|---|
| Baseline audit | May 2026 | 12 / 100 |
| Contracted target | 90-day goal | 55–60 / 100 |
| Re-audit result | 24 August 2026 | 79 / 100 |
Read it plainly: the score more than sextupled and finished 19 points clear of the top of the contracted 55–60 target band.
Source: AiDisco AI Narrative Intelligence™ — re-measurement 24 August 2026.
Exhibit 3 — Buyer Share of Voice by Answer Engine
Demand-weighted share of buyer questions returning Design.Aid, by platform — Day 90
| AI answer engine | Share of Voice | Buyer Share of Voice |
|---|---|---|
| Gemini | 78% | 86% |
| Microsoft Copilot | 66% | 74% |
| ChatGPT | 26% | 16% |
| Claude | 3% | 3% |
Read it plainly: platform is a real dimension, not a footnote — the same question is effectively a different question on each assistant. Gemini and Copilot lean hardest on first-party content, and Design.Aid now reaches the large majority of buyer demand on both: direct proof the rebuilt pages work as designed. ChatGPT and Claude weight independent third-party corroboration more heavily before naming any provider, which is precisely where the next phase of authority work pays.
Source: AiDisco AI Narrative Intelligence™ — re-measurement 24 August 2026.
Exhibit 4 — Quality of Placements Earned
Ranked placements across 130 engine-by-prompt checks
| Placement type | Count | Share of placements |
|---|---|---|
| Top answer (lead recommendation) | 48 | 83% |
| Ranked but not first | 10 | 17% |
| Total ranked placements | 58 | 100% |
Read it plainly: when Design.Aid appears, it is the lead recommendation 83% of the time — 48 of 58 ranked placements across 130 engine-by-prompt checks. More striking at the question level: there is not a single priority question where Design.Aid appears without ranking #1 on at least one engine. Presence is the floor; this is dominance.
Source: AiDisco AI Narrative Intelligence™ — re-measurement 24 August 2026.
Exhibit 5 — Weekly New Website Visitors
Pre-engagement weekly average vs. week of 17–23 August 2026
| Period | Days | New visitors | Per day | vs. baseline |
|---|---|---|---|---|
| Before engagement (weekly average) | 7 | 3 | 0.4 | — |
| Launch window, 2–10 Aug 2026 | 9 | 33 | 3.7 | 11× |
| Week of 17–23 Aug 2026 | 7 | 45 | 6.4 | 15× |
Read it plainly: a 15-fold increase in weekly new visitors. The rate is still climbing — 3.7 new visitors per day during the launch window, 6.4 per day two weeks later — so this is compounding rather than a launch spike settling back down. Since 28 August 2026, that traffic has converted into at least one new enquiry every day — visibility is now producing pipeline, not just page views.
Source: Google Analytics, designaid.com.au. Baseline weekly average pre-engagement; comparison weeks as dated.
Exhibit 6 — Where the Visitors Came From
Active users by city, week of 17–23 August 2026
| City | Active users | Session source / medium | Sessions |
|---|---|---|---|
| Melbourne, AU | 7 | (direct) / (none) | 34 |
| Sydney, AU | 4 | google / organic | 24 |
| Brisbane, AU | 2 | deadlinkchecker.com / referral | 3 |
| Other cities (7 listed) | 9 | aidisco.live / referral · bing / organic | 2 |
Read it plainly: Melbourne, Sydney and Brisbane lead the week — the three markets Design.Aid sells into. Google organic is the largest identified acquisition channel at 24 sessions. In a seven-day sample of this size, a portion of the remaining traffic sits in known datacentre locations and automated crawler referrals, which is normal for a recently launched site and is not counted as buyer demand.
Source: Google Analytics, designaid.com.au, 17–23 August 2026. Separately, over 28 July – 24 August 2026 the site recorded seven sessions referred directly from chatgpt.com, alongside a referral from an independent authority-platform profile.
Day 90 scoreboard: where Design.Aid stands
Exhibit 7 — Depth of ownership by buyer-question cluster
Top-Answer Buyer Share of Voice by cluster
Demand-weighted share of each cluster where Design.Aid is the #1 answer on at least one engine
| Cluster | Prompts | Appearing | #1 answer | Demand owned at #1 |
|---|---|---|---|---|
| Core provider discovery | 15 | 15 | 15 | 100% |
| Best-provider comparisons | 3 | 3 | 3 | 100% |
| Cheapest & value comparisons | 3 | 2 | 2 | 82% |
| CAD drafting outsourcing | 12 | 7 | 7 | 73% |
| CAD provider comparisons | 3 | 2 | 2 | 70% |
| All priority prompts | 36 | 29 | 29 | 90% |
Read it plainly: the two clusters closest to a buying decision are completely owned. Every one of the 15 core "who should I use" questions and all 3 "who are the best providers" questions now return Design.Aid as the #1 answer. In every cluster, the number appearing and the number ranking #1 are identical — Design.Aid does not place second anywhere it places at all.
Source: AiDisco AI Narrative Intelligence™ — re-measurement 24 August 2026.
Exhibit 8 — The same buyer questions, ninety days apart
| Buyer question | Day 0 | Gemini | Copilot | ChatGPT | Claude | Best position |
|---|---|---|---|---|---|---|
| Who offers offshore CAD drafting services in Australia? | Absent | #1 | #1 | — | — | #1 |
| What is the best offshore drafting service in Australia? | Absent | #1 | #1 | — | — | #1 |
| Who provides outsourced architectural drafting for Australian residential projects? | Absent | #1 | — | #6 | — | #1 |
| What offshore drafting provider suits builders with inconsistent workloads? | Absent | #2 | — | #1 | — | #1 |
| Where can I find offshore CAD drafters experienced in Australian residential projects? | Absent | #1 | #1 | — | — | #1 |
| What is the cheapest offshore drafting company in Australia? | Absent | #1 | #1 | #2 | — | #1 |
| Who provides cost-effective architectural CAD services for Australian firms? | Absent | #1 | — | — | #1 | #1 |
Read it plainly: Design.Aid holds the #1 position on every one of these questions, having held none of them in May. Gemini returns Design.Aid first on six of the seven. Where an engine shows a dash, Design.Aid is not yet returned on that engine for that question — the remaining ground the authority phase is built to take.
Source: AiDisco AI Narrative Intelligence™ — re-measurement 24 August 2026. Dash indicates not returned in the measured sample.
AI Visibility Case Study FAQs
What is AI visibility, and how is it different from SEO?
SEO measures where a page ranks on a results page of links. AI visibility measures whether a brand is named inside the single answer an AI assistant composes when a buyer asks a question in plain language. There is no page two in an AI answer — a brand is either in it or it is not. Design.Aid ranked around page 12 on Google and appeared in 0% of AI answers at baseline, which are two separate failures requiring two different fixes.
What results did AiDisco deliver for Design.Aid?
Over 90 days, Design.Aid went from appearing in 0% of AI answers to appearing on 81% of 36 priority unbranded buyer prompts, reached 90% Buyer Reach and 46% Buyer Share of Voice when weighted by buyer demand, and moved from 12/100 to 79/100 on the AiDisco Visibility Score. Design.Aid also moved from approximately page 12 to #1 on Google for its core service terms, and weekly new website visitors rose from 3 to 45 — a 15-fold increase — by the week of 17–23 August 2026. Since 28 August 2026, that visibility has converted into at least one new enquiry every day through the site.
How long does it take to build AI visibility from zero?
In the Design.Aid engagement, measurable AI answer placements appeared within the first week of the rebuilt content going live, and the full result was measured at 90 days. The engagement ran from 13 May 2026 to 11 August 2026, with the re-audit on 24 August 2026. Traffic continued to compound after launch, rising from 3.7 new visitors per day in the launch window to 6.4 per day two weeks later. AI visibility is not a set-and-forget outcome: answers shift as content ages, competitors publish, and models are retrained, so it requires ongoing monitoring.
What is Share of Voice in AI search?
Share of Voice is the percentage of mapped unbranded buyer prompts in which a brand appears in the AI-generated answer, measured against every competitor also cited. It is measured only on unbranded questions — questions where the buyer names no company — so inclusion is earned rather than automatic. Design.Aid appeared on 81% of its 36 priority prompts, up from 0%.
What is the difference between Buyer Reach and Buyer Share of Voice?
Buyer Reach asks whether a buyer can find you at all: the demand-weighted share of buyer questions where the brand appears on at least one AI engine. Buyer Share of Voice asks how much of the answer space you actually hold: the demand-weighted share of every engine-by-prompt check that returns the brand. Design.Aid reached 90% Buyer Reach but 46% Buyer Share of Voice. Both are true and they are reported separately, never averaged, because the gap between them is the diagnosis — it shows that presence is concentrated on some engines and thin on others.
Do all AI engines answer the same way?
No. The same question is effectively a different question on each assistant because each retrieves and composes answers differently. At Day 90, Design.Aid reached 86% demand-weighted presence on Gemini and 74% on Microsoft Copilot, but 16% on ChatGPT and 3% on Claude. This spread is exactly why a single blended AI visibility score is misleading, and why the pooled Buyer Share of Voice of 46% sits so far below the 90% Buyer Reach figure. Gemini and Copilot lean more heavily on first-party website content; ChatGPT and Claude weight independent third-party corroboration such as directories, reviews and industry platforms more heavily before naming a provider.
What is a Prompt Universe?
A Prompt Universe is the full set of natural-language questions a real buyer could ask an AI assistant that might surface a brand or its category. It is effectively infinite, so it is sampled rather than measured completely. Design.Aid's Prompt Universe was mapped to 225 buyer questions phrased in buyer voice, from which a priority set of 36 was drawn by demand and tested across four engines.
What is Top-Answer Share of Voice?
Top-Answer Share of Voice is the stricter version of Buyer Share of Voice: the demand-weighted share of engine-by-prompt checks where the brand is returned as the number one answer rather than merely named somewhere in the response. Design.Aid reached 40% Top-Answer Share of Voice against 46% Buyer Share of Voice, meaning roughly seven of every eight placements it holds are first position. At the question level the figure is stronger still: Top-Answer Reach is 90%, identical to Buyer Reach, because Design.Aid ranks #1 on at least one engine for every question where it appears at all.
Which buyer questions does Design.Aid own outright?
All 15 core provider-discovery questions — the "who should I use for offshore drafting in Australia" cluster — return Design.Aid as the #1 answer, representing 100% of measured demand in that cluster. All 3 best-provider comparison questions do the same. Cheapest and value comparisons sit at 82% of demand owned at #1, CAD drafting outsourcing at 73%, and CAD provider comparisons at 70%.
Is appearing in an AI answer enough?
No — presence is the floor, not the ceiling. A brand can be mentioned last, hedged, or described inaccurately and still technically appear. AiDisco measures dominance separately: of Design.Aid's 58 ranked placements across 130 engine-by-prompt checks, 48 (83%) were the lead recommendation rather than a passing mention — and there is no priority question on which Design.Aid appears without ranking #1 somewhere.
Who is AiDisco?
AiDisco is an AI Narrative Intelligence consultancy that measures and improves how AI assistants describe and recommend brands. Its practice covers Prompt Universe mapping, Share of Answer measurement across ChatGPT, Claude, Gemini and Copilot, third-party authority placement, and AI-readable content and schema implementation.
Working with the team at AiDisco has completely transformed Design.Aid’s online visibility. Their AI-driven strategy quickly boosted our presence and has resulted in a steady flow of new enquiries. It’s been the most effective upgrade we’ve made to our site.
Sustaining AI Visibility After the Sprint
What machines say about a brand is not fixed. It shifts as content ages, as competitors publish, and as models are retrained. The next phase protects the ground already taken and pushes into the clearest remaining lever: independent, third-party proof.
How These Figures Were Measured
Source: AiDisco AI Narrative Intelligence™. The AiDisco Visibility Score reflects the AiDisco audit — baseline 12/100 (May 2026), re-audit 79/100 (24 August 2026). Share of Voice and Buyer Share of Voice are measured across the priority unbranded prompt set of 36 questions drawn from Design.Aid's 225-question Prompt Universe, tested across Gemini, Microsoft Copilot, ChatGPT and Claude across 130 engine-by-prompt checks. Three presence measures are reported separately and never averaged: Answer Reach (unweighted share of the 36 priority prompts returning Design.Aid on at least one engine), Buyer Reach (the same, weighted by estimated monthly question demand), and Buyer Share of Voice (demand-weighted share of all 130 engine-by-prompt checks returning Design.Aid). Demand weights are estimated proxies, so all weighted figures are directional only. Traffic figures are Google Analytics data for designaid.com.au: the headline weekly comparison covers 17–23 August 2026 against the pre-engagement weekly average, with referral-source detail drawn from the wider 28 July – 24 August 2026 window as dated. A seven-day sample of this size includes some automated crawler and datacentre traffic, which is not treated as buyer demand. AI answers vary between runs and across platforms: a reading is a reading of a sample, on named platforms, at a moment in time.
Engagement dates: 13 May 2026 (discovery) to 11 August 2026 (handover). Site go-live 3 August 2026. Re-measurement 24 August 2026.
