Case study
A structural engineering consultancy of roughly sixty staff won work through referrals until two quiet quarters made that dependence uncomfortable. We rebuilt the content library around the questions municipal engineers actually ask.
Named only with client approval Figures marked are illustrative Reply within one business day
Vancouver, BC
custom software development vancouver
3
1
2
5
Impressions
13 weeks
+186%
Impressions
4.2
Average position
27
Page-one keywords
OlDevs rebuilt the organic search presence of a structural engineering consultancy serving municipalities across Ontario, replacing a referral led pipeline with a content library built around real procurement questions.Each page answers one question in its opening paragraph and carries Service and FAQPage schema, while an llms.txt file points AI assistants to the canonical sources. Twenty four question pages are published, non branded organic sessions have reached 3.4x, and enquiries from search are up 52 percent. The firm's own engineers write the first drafts, so the expertise on the page stays theirs.
Key facts
The challenge
Almost every project arrived through a referral or a past client, so the pipeline moved with the phone rather than with a plan. When municipal procurement staff searched for the firm's specialisms, competitors came up instead, and the firm was absent from the answers people increasingly get from AI assistants. Two quiet quarters made that dependence uncomfortable enough to act on.
Work arrived because a past client or a colleague made an introduction, not because a buyer found the firm. That made forecasting guesswork, and when two quarters went quiet there was no second channel to turn on, because none had ever been built.
When municipal engineers and procurement staff searched the firm's specialisms, other consultancies held the results. The site described the practice in broad terms and never answered a specific technical question a buyer would actually type into a search box.
Buyers were already asking ChatGPT, Perplexity and Google AI Overviews for shortlists and guidance. Nothing on the site was structured to be quoted back accurately: no direct answers up front, no schema, and no llms.txt pointing assistants at canonical sources.
Sixty staff held detailed municipal engineering knowledge and almost none of it existed in public. Marketing had no repeatable way to turn a conversation with a principal into a page, so the website said far less than the firm actually knew.
What we built
We rebuilt the site's content around the questions municipal engineers actually ask, one page per problem, with the direct answer in the opening paragraph and the technical detail underneath. Every page carries Service and FAQPage schema, an llms.txt file points to the canonical sources, and the library is tracked in Google Search Console alongside prompt level monitoring in ChatGPT, Perplexity and Google AI Overviews. The firm's own engineers write the first drafts and we edit, so the expertise stays theirs.
Twenty four pages, each built around one question a municipal engineer would genuinely type. The direct answer sits in the opening paragraph, and the working, the standards and the caveats sit underneath for the reader who needs the detail.
Structured data on every page in the library, validated and kept consistent, so search engines and AI assistants can read what the firm does and which question each page settles, instead of inferring it from prose that may be summarised badly.
An llms.txt file points assistants to the canonical version of each answer, supported by clean heading hierarchy, stable URLs and self contained paragraphs that still make sense when they are quoted away from the page around them.
Google Search Console tracks the library page by page, and prompt level monitoring records whether ChatGPT, Perplexity and Google AI Overviews name the firm for the questions its buyers ask, so both channels are read side by side each month.
Process
Buyer question audit
We interviewed principals and reviewed enquiry email, RFP language and Search Console queries to build a ranked list of the questions municipal buyers actually bring to the firm.
Page map and priorities
Each question became one page with a defined answer, an owner and internal links, ordered by how close that question sits to an actual procurement decision.
Engineers draft, we edit
The firm's engineers write first drafts in their own words. We edit for structure, clarity and Canadian spelling, then check every technical claim back with the author before publishing.
Structure and schema build
We publish with Service and FAQPage schema, an llms.txt file, clean heading hierarchy and stable URLs, validating the markup and the accessibility of each page before it goes live.
Monitor, then commission the next questions
Each month we review Search Console alongside prompt level checks in ChatGPT, Perplexity and Google AI Overviews, and the gaps we find decide what gets written next.
Stack
Content and structure
Search measurement
AI answer monitoring
Editorial workflow
Page standards
3.4x
Non branded organic sessions
+52%
Enquiries from search
24
Question pages published
Outcomes
Municipal buyers now find the firm before anyone introduces it, so quiet referral quarters no longer decide the year.
Search traffic from people who did not already know the firm's name grew to 3.4x, which is the number that matters when the goal is reaching strangers.
More of that traffic turns into a conversation, because each page answers a real question rather than describing the practice in general terms.
The firm's own knowledge is finally written down in public, owned by the firm, and the publishing rhythm continues rather than ending with a launch.
In their words
We now get calls from municipalities where nobody has ever heard our name from a colleague.
Managing principal, regional engineering consultancy
Professional services, engineering
FAQ
Structured pages usually start collecting impressions within weeks, but enquiries follow the publishing rate and the competitiveness of each question. This engagement has been running since 2024, and the compounding came from steady publishing rather than one launch. While the pipeline builds we report on leading signals such as impressions, rankings and question coverage.
It is the part that makes the work defensible. Your engineers write a first draft in their own words, usually faster than a long interview would take, and we edit for structure, clarity and Canadian spelling before checking every technical claim back with the author. The expertise on the page stays yours, and so does the credibility that comes with it.
It is a plain text file at the root of a site that points AI assistants to the canonical version of your key content. It is not a guarantee of citation and we do not sell it as one. It works alongside clean headings, self contained paragraphs and Service and FAQPage schema, which together make a page easy for an assistant to quote correctly.
We run prompt level monitoring. A fixed set of the questions your buyers ask is checked on a schedule in ChatGPT, Perplexity and Google AI Overviews, and the results are recorded so movement is visible over time. That sits next to Google Search Console data for the same questions, so classic search and AI answers can be read together.
Yes. The method is not sector specific: find the questions a buyer types, answer each one directly on its own page, structure it so machines can read it, then measure both search and AI answers. Our studio is in Vancouver and we work with clients across Canada remotely, with video calls in your time zone and on-site visits when the work calls for it.
Let’s connect
Tell us what is not working. We’ll reply within one business day with how we would approach it and a tailored quote.