Project Delivery Assurance Services wins contracts the way good practitioner firms often do. Through people who have already seen them deliver.
The referral channel had carried the firm reliably for years. It had also become the only one it had, which meant that if referrals slowed, growth slowed with them.
The firm's authority sat with the founding partners and travelled by relationship. A CEO who had never met them had no way to judge PDAS before picking up the phone.
Prospect behaviour was changing at the same time. An organisation considering an independent review of a capital programme now starts with a search engine or an LLM and receives detailed answers to their questions before they speak to a supplier. The sources that appear in that answer are chosen by LLMs before PDAS ever speaks to the buyer.
So, the commercial objective was to provide PDAS a second route to market, that allowed buyers unfamiliar with the firm to discover and evaluate them online.
RR CEO was engaged to build the digital infrastructure for the project. The work included a redesigned and rebuilt Webflow site, migrated from WordPress, the content creation and publishing operation behind it, and an AI readiness diagnostic PDAS could put directly in front of website visitors from their target audience in the energy, minerals and resources industry.
One main constraint governed the whole design. Whatever was to be built had to run on the time of PDAS practitioners who already had full schedules, onsite at clients, in various geographies and time zones.
Airtable was used as the knowledge base. Every piece of content started as an Airtable record, and that record remained the source of truth after publication. It held the brief, target terms, service category, publication date and various other information and meta data relating to work already published.
Content was planned in 90-day blocks, with PDAS service areas spread across the quarter, often to be timed with known industry events and news cycles, so no single discipline dominated.
Publishing schedules rarely fail because nobody has ideas. They fail when client work spikes and content becomes the first thing pushed aside.
SSo the running order for those weeks was agreed in advance. Deciding beforehand what gives way, and what does not, is what kept the cadence intact.
Since launch, PDAS has published 46 pieces of content: 33 articles, nine LinkedIn articles, and four editions of a monthly newsletter.
The editorial approach was built around a simple idea, the practice now called answer engine optimisation (AEO). The content had to be useful to a human reader, but also clear enough for search engines and LLMs to lift the right answer from it and link that answer back to PDAS.
The aim was to ensure every important statement was accurate and clear enough that neither a reader nor a machine had to guess what it meant.
Research drew on real frameworks, methodology documents and technical material supplied by PDAS, with claims checked before they reached a finished draft.
The PDAS practitioners then read every completed piece and corrected anything that did not reflect real practice. This step was central to the model. It is what allowed a small specialist firm to publish consistently without turning technical authority into generic marketing content.
The production system created capacity while the practitioners protected their standards.
The final component, an AI Readiness Snapshot, built as a self-contained diagnostic inside the PDAS site, went live in August 2026.
Visitors answer a structured set of questions, receive a scored report on screen, and can also request a copy of the report by email.
The articles help people find PDAS and understand what the firm does. The diagnostic gives them something useful to do next. It lets a potential client see where they stand in relation to AI and their organisaiton before speaking to anyone and provides PDAS a much better sense of what the prospective client may need.
Six months of live data now sit behind the project.
All 33 published articles rank in Google, while thirty-one appear inside Google's AI results.
Across the period, 41 buyer queries averaged a position in the top three and 150 in the top ten.
Bing provides a clearer view of AI visibility because it reports citations separately from clicks. Between late March and late August 2026, Bing Webmaster Tools recorded 2,414 citations of PDAS pages across Microsoft AI-generated answers, including Copilot and Bing AI experiences.
One pattern in the query data was unexpected. People were finding PDAS by searching for the studies, statistics and figures cited in its articles. In other words, in addition to supporting the article arguments, careful sourcing was also creating another way for people to find PDAS.
The more interesting finding was the relationship between AI visibility and click-through.
Across the 23 pages with enough data to compare, the pages appearing most often in AI-generated results were also much more likely to be clicked. Pages where at least 10% of impressions came from LLM citations had click-through rates around three times higher than the rest, even though their typical Google ranking was the same. Across all 23 pages, the relationship between AI visibility and click-through was strong, with a correlation of 0.70.
That does not prove that AI visibility caused the higher click-through. The pages attracting the most AI visibility were often definitional and comparative pieces, which may have simply been performing well in both places.
Even so, the results are important. They showed that writing the content so LLMs could use it did not make it less appealing to human readers.
The second finding was about control.
LLM citation volume moved independently of anything PDAS did day to day, ranging from fewer than one citation to more than forty in a day across the period. Search engines and LLMs decide when and where a source appears.
That is why what happened after someone landed on the PDAS site mattered so much.
PDAS can influence how clearly its expertise is published and how useful the site is when someone arrives. It can own the articles, the evidence, the diagnostic and the relationship that follows. It cannot own the intermediary that brought the visitor there.
And that is why the commercial infrastructure was fundamental. Six months later, PDAS has a second route to market that operates alongside the referral engine that has always served the firm.
