The challenge
BuildPass wanted to show up in AI answers when Australian builders ask which construction software to use. Procore (US$11B market cap), Autodesk and Visibuild were already cited in those conversations. BuildPass was not.
When Peec AI tracking went live in November 2025, BuildPass sat at near-zero visibility across every prompt in the category. The competition had content teams and budgets BuildPass could not match.
The brief was specific: become a cited brand in AI answers about construction software, in Australia, on a Series-A budget BuildPass could sustain quarter after quarter.
The approach
Every piece of work started with a tracked prompt in Peec AI, not a keyword in a search tool. We called it prompt-led content, and it ran in three layers.
-
Prompt research. We used Peec AI to identify the 11 questions BuildPass needed to be the cited answer to: defect tracking, site sign-in, drawing organisation, punch lists, JSAs, ITPs, incident reporting, scheduling, compliance, quality audits and subcontractor management.
-
Long-form content production. One guide per prompt, structured for direct extraction by AI engines: scannable headings, definitional intros, numbered steps, comparison tables. Each guide was rewritten and re-tested until it earned a citation.
-
Technical layer. We added FAQ and HowTo schema to every guide in JSON-LD, then rewrote eight product pages for plain-language clarity so AI engines could quote them verbatim.
The cadence was two-week sprints, each one targeting one prompt end-to-end: pull the latest Peec visibility data, read the top five cited sources to see what was getting cited and why, draft a better answer, ship the schema, rewrite the linked product page, monitor week by week.
The defect tracking guide was rewritten three times before BuildPass hit 52% visibility on that prompt. The site sign-in guide hit 40% after one rewrite, putting BuildPass at #1 for that prompt.
We capped the campaign in May 2026 with a full site rebuild. BuildPass migrated off Webflow onto Next.js on Vercel for server-side rendering, faster pages and clean schema at build time. The new site launched 23 May 2026.
Picking the right battles
Two prompts never moved: subcontractor management and Australian compliance. AI engines route those questions to legal and government sources, not software vendors. We retired them from the content plan and reinvested the effort in prompts BuildPass could actually own.
That reallocation matters. Most agencies would have kept pushing on a flat prompt because it was in the original brief. Watching Peec visibility week by week meant we could cut what was not working and put the time into what was.
The results
Measured across April 2025 to April 2026, with the AI search optimisation phase running mid-October 2025 to May 2026.
AI visibility (Peec AI)
- #2 brand share of voice on Google AI Overview at 21%, behind Procore (49%) and ahead of Autodesk (15%) and Visibuild (15%). Procore has a US$11B market cap.
- Outright #1 on "efficient construction site sign-in management": BuildPass 40% visibility, Procore 27%, Autodesk 2%.
- Strong #2 on "best software for construction defect tracking": BuildPass 52% visibility, Procore 89%, Autodesk 48%.
- Strong #2 on "storing and accessing construction incident reports digitally": BuildPass 24% visibility, Procore 47%.
- Highest sentiment of any tracked brand at 64 (Procore 62, Autodesk 61, Visibuild 60).
- buildpass.ai cited as a source in 21% of relevant AI chats, with a citation rate of 1.2.
AI-attributed traffic (GA4)
- 1,023 AI-sourced sessions across the 13-month window.
- Monthly AI sessions grew from 21 (April 2025) to a 100+ run-rate from October 2025, peaking at 109 in February 2026, a 5× lift from start to peak.
Conversions from AI traffic
- 81 key events: 50 demo bookings, 16 demo form submissions, 14 "Let's Chat" engagements, 1 newsletter signup.
- 4.5× lift in monthly AI-attributed conversions versus the pre-LLM baseline. The pre-LLM phase (April to September 2025) delivered 13 conversions across six months, 2.2 per month. The LLM phase (October 2025 to April 2026) delivered 68 conversions across seven months, 9.7 per month.
- November 2025 delivered 16 AI-source conversions, the highest single month.
- Conversion rate from AI traffic: 7.9%, against a typical B2B SaaS demo-form conversion of 1 to 3%.
What we'd point to
A small Australian agency outranked an US$11B competitor on a high-intent commercial prompt. BuildPass beats Procore 40% to 27% on "construction site sign-in management".
The methodology is repeatable. Every piece was briefed against a tracked Peec AI prompt, so visibility moved week by week instead of quarter by quarter.
And the work tied AI search visibility to revenue. Most AI search work is brand-only. This one moved demo bookings and form submissions for a B2B SaaS, which is the question the rest of the industry is still trying to answer.