Case Study — PaperMake Enterprises

The situation

PaperMake has supplied janitorial, washroom, paper and packaging products to businesses across the Lower Mainland since 1985. Established local operation, own delivery fleet, hundreds of active customers. But when the audit ran, the site scored 31/100 — and every one of the nine flagged fixes was a signal AI assistants use to decide whether to name a business in an answer.

The scan surfaced 9 distinct issues. None catastrophic on their own. Together they left PaperMake invisible to ChatGPT, Claude, Perplexity and Google's AI Overview panel — the tools their next customer is using to find suppliers right now.

What was broken

No H1 on the home page. AI extractors treat H1 as "what is this page about" — there was nothing to extract.
44 words of body copy. The home page was almost entirely an animated hero image. LLMs need text to cite; there was almost none.
Meta title 92 characters, description 225. Both truncated in search results; the primary keyword sat past the visible cut-off.
No canonical, no Open Graph, no Twitter Card. Every share and every AI preview looked like a generic link.
JSON-LD limited to a barebones Organization block. No opening hours, no address, no service area, no product catalog — nothing for an AI to answer "who supplies janitorial products in Surrey?"
robots.txt did not name any AI crawler. Implicit access works, but explicit allows are the signal an audit picks up on.
No sitemap.xml. Crawlers had to guess at the URL surface.

What we shipped

Five file changes plus three new crawler files — every change scoped strictly to the client's home directory, deployed the same afternoon.

Home page

Meta title rewritten to Janitorial Supplies Surrey BC | Papermake Enterprises (58 characters, keyword-first).
Meta description rewritten to 152 characters — promise + proof + location.
Canonical URL, Open Graph and Twitter Card added.
Structured data expanded from Organization to a LocalBusiness + Store + WebSite @graph: full address, geo coordinates, phone, opening hours, 8-city area-served list, and an OfferCatalog naming the eight product categories.
Preserved the existing hero. Added a proper content section below it with 1 H1, 5 H2s and ~600 words of substantive copy: what they supply, who they work with, why customers choose them, delivery coverage (19 named cities), and how to order.

Crawler files

New robots.txt with 15 named AI-crawler Allow blocks (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Applebot-Extended, Amazonbot, Bytespider, CCBot, and more).
New sitemap.xml with 17 URLs across home, catalog, contact and product-guide pages — linked from robots.txt.
New llms.txt: a 7.6 KB structured description of the business, its products, its service area and its ordering channels — the emerging convention LLM providers are starting to prefer.

The result

Same audit tool, run again after the deploy. Every check that could be perfected, was.

Check	Before	After	Change
Structured data (JSON-LD)	10/25	25/25	+15
AI crawlers allowed	8/15	15/15	+7
llms.txt file	0/12	12/12	+12
Meta title & description	2/10	10/10	+8
Heading structure (H1/H2)	0/8	8/8	+8
Content depth	0/10	7/10	+7
Open Graph & Twitter	0/6	6/6	+6
Canonical URL	0/3	3/3	+3
HTTPS	4/4	4/4	—
sitemap.xml	0/4	4/4	+4
Robots.txt present	3/3	3/3	—
TOTAL	31/100	97/100	+66

One check still not perfect. Content depth: 609 words is enough for AI extraction but not maximal. Room to grow the site's editorial footprint over time; not a same-day priority.

What this means for PaperMake

Two months from now, when a facility manager in Langley asks Claude or ChatGPT "who supplies commercial cleaning products in Surrey, BC?", PaperMake is now the kind of source those tools are built to cite: named business, resolvable address, service area, opening hours, product catalog — all cross-referenced in structured data.

The technical layer is done. What earns the citations from here is the volume of specific, useful content on the site — product guides, buyer questions answered plainly, application notes. That's a longer arc; the audit re-scores as new content ships.

Read the raw reports

Every claim on this page is reproducible from the audit tool's output. Here are the two full reports plus the case-study PDF.

Report 1

Before — 31/100

Full pre-fix audit. 9 issues flagged with the specific fix for each.

Download PDF →

Report 2

After — 97/100

Full post-fix audit. Same tool, same site, one afternoon later.

Download PDF →

Summary

Case Study

Narrative wrapper: problem, work, result, per-check diff. Print-ready.

Download PDF →

Want the same for your site?

The $199 Full Fix is the tier PaperMake used. Same scope, same delivery pattern, and the audit is included.

See the $199 Full Fix → Not ready yet? Start with the $19 audit and see what your own site scores.

PaperMake Enterprises Ltd.

The situation

What was broken

What we shipped

Home page

Crawler files

The result

What this means for PaperMake

Read the raw reports

Want the same for your site?