Case Study: How InLinks & Waikay Drove a 650% Increase in AI Visibility
Background: A hole found in AI understanding
At the beginning of the year, I spoke with an InLinks client about their long‑term goals. As a major online accounting software provider, they had already established dominance across several niches. One area they identified as a priority, however, was building greater visibility and authority around the entity of “E‑invoicing.”
At that time, Waikay was just emerging, and I wanted to test whether the client had any measurable AI brand visibility for e‑invoicing. The aim was to assess their positioning in this niche and, ideally, encourage them to adopt our new tracking tool.
The starting point was a straightforward prompt: “best e‑invoicing software.” Waikay parsed the results into a competitor list of 15 brands. For each brand, it broke down the associated entities and allowed me to count how often e‑invoicing appeared in their descriptions.
The findings revealed a clear imbalance. Across all competitors, invoicing was consistently emphasized, while e‑invoicing only appeared sporadically. The data was repeatedly reshaped to highlight general invoicing features rather than e‑invoicing, both for Brand A (the anonymized focus of this case study) and for every other competitor.
This highlighted a key insight: even when prompted directly for e‑invoicing, the competitive landscape was framed around traditional invoicing. Entities tied to invoicing, business processes, automation, and customizability dominated, while explicit mentions of “e‑invoicing” remained extremely limited.
Recognizing this gap, I began tracking the trend over time using LLMs. Over the past year, Brand A has leveraged InLinks to execute an entity‑based SEO strategy with strong results — but the question remained: how does this success translate into AI visibility?
The Hypothesis: Can InLinks data improve AI Visibility?
My hypothesis was that an InLinks strategy would help improve an LLM association with a particular entity. I sought evidence that the number of times Brand A was discussed in the context of E-invoicing for a controlled prompt would gradually increase over time, demonstrating that InLinks work and Waikay follows trends accurately.
The InLinks Action Taken
To improve entity relevance for “e‑invoicing,” the Brand A used InLinks to:
- Target the main service page with the correct entity.
- Implement a site‑wide, entity‑based internal linking structure pointing to that page.
- Add schema markup for the e‑invoicing entity to strengthen machine interpretability.
- Added content around the entity of e-invoicing through the content brief side of the tool
InLinks drove the structural change; Waikay provided the measurement layer.
The Waikay Tracking method
Waikay runs a structured, repeatable visibility tracking loop:
- Consistent prompting: Every two days, four LLMs are asked the same question.
- NLP extraction: Waikay parses responses to log brand mentions, competitor mentions, and the descriptors/entities tied to each.
- Time series tracking: Mentions and associated entities are tracked across weeks to observe salience shifts.
This method allows you to see how often your brand is mentioned, what it’s associated with, and how that changes over time — without needing to manually prompt or interpret results.
Here is a screenshot of the way this data is shown on a completely separate project about custom sign prints.
Waikay runs an API call using the same prompt every two days. Over the course of a month, this produces a rolling dataset that shows which entities are consistently associated with each brand, and how often those associations appear.
In this dataset, Smart Sign is strongly linked to the entities of signs and customization. However, it is almost never connected to the entity of printing — a concept that appears regularly for nearly all of its competitors.
This gap is important because the aim of the analysis is to visualize an LLM’s reasoning process: working backwards from the entities it surfaces to identify potential content gaps.
In Smart Sign’s case, the absence of “printing” as a recurring association suggests a risk. Although the brand is well‑positioned around customizable signs, it may lose visibility in future contexts if the printing angle is not reinforced — even though printing is in fact part of their offering.
By tracking these associations over time, Waikay highlights not just where a brand is strong, but also where its narrative may be incomplete. This makes it possible to proactively adjust content strategy to close gaps before they impact AI visibility.
The Results: How Entity Frequency Skyrocketed
The story this tells is fantastic, and a huge learning point for SEOs. Ultimately, Brand A (first row) have dramatically improved an LLMs understanding of E-invoicing. They lead this entity, with 42 mentions of it in the past 30 days , over double their next closest competitor.
Lets take a look at this increase over the last few months:
The number of times the entity “e‑invoicing” appeared in relation to the brand rose from 2 mentions at baseline to 15 at peak — a 650% increase.
Important: What increased was not (necessarily) how often the brand was named, but how often the entity ‘e‑invoicing’ was explicitly tied to the brand in LLM responses.
This shows a deeper and more frequent association between the brand and the entity, hopefully resulting in higher visibility for queries relating to this entity.
Brand A also improved in other entities as well, including Accounting and Automation “Automation”
The heatmap and accompanying graphs illustrates a clear win for Brand A, which dominates in mentions tied directly to the entity of e‑invoicing.
However, the picture is more nuanced when looking across competitors. Brand B and Brand C stand out in other areas — particularly around customizability and features — where they consistently outperform Brand A.
This breakdown begins to reveal how LLMs interpret and prioritize associations over time.
“Rather than simply counting brand mentions, the heatmap surfaces the semantic patterns that shape visibility: which entities are strongly linked to each brand, and how those links evolve. In this way, the data provides an early signal of LLM behavior and understanding, highlighting both strengths and potential gaps in positioning.” – Dixon Jones
What This Means For AI Visibility Strategy
AI visibility is ultimately about positioning your brand closer to the right entities so it surfaces in more queries. You don’t need to track hundreds of prompts — a single, well‑chosen prompt can reveal enough direction to guide strategy for months.
In this case, Brand A demonstrated strong foresight by aligning with the entity of e‑invoicing. However, the analysis also exposed areas where they fell short. For example, they are not positioned around customizability, yet they are likely to dominate in compliance and accounting contexts simply because of their strong associations there.
This is a powerful illustration of how entity‑based SEO directly shapes LLM understanding, and how Waikay can measure those shifts with precision. The right way to assess AI visibility isn’t by counting brand mentions, but by analyzing which entities consistently appear alongside brands — and then turning those associations into actionable data.
With this approach, prompts can be reverse‑engineered. The heatmap shows where Brand A excels (e‑invoicing), but also where gaps exist. If a user asks, “Which e‑invoicing software is the most customizable?” the data suggests Brand C would dominate that context.
This isn’t about search volume. It’s about testing, retesting, and observing LLM behavior — then adjusting your narrative to fill the gaps. Entity‑based strategy worked here, and Waikay’s tracking method surfaces that success without bias or guesswork.
What should Brand A do now?
Now that the gaps around e‑invoicing are clear, Waikay can step in to shape the next stage. The data highlights where expansion is needed, and from this we can build an actionable topic report. For Brand A, the recommendations are organized into three areas: new content that should be added, enhancements to existing material, and broader structural improvements informed by the insights we’ve collected.
Below is an action plan made by Waikay (modified for anonymity)
Brand A Topic Report for E-invoicing
Content additions
High priority
Content Type: Pillar page plus interlinked regional guides (blog/guide format, not duplicating existing regional product pages)
Create a central, SEO‑optimized pillar page (e.g., /global/e-invoicing or /e-invoicing/guide) that:
- Explains what e‑invoicing is
- Compares it to traditional invoicing
- Outlines benefits (automation, faster payment, fewer errors)
- Explains PEPPOL/InvoiceNow and cross‑border networks
- Links out to existing regional e‑invoicing product pages (/au/e-invoicing, /nz/e-invoicing, /sg/e-invoicing, /my/accounting-software/e-invoicing, etc.)
- Summarizes mandatory e‑invoicing timelines by country with links into more detailed local guides or campaigns
- Surfaces existing media releases and blog posts as a ‘latest updates’ module
Developer‑Facing e‑Invoicing API and Integration Overview
High Priority
Content Type: Developer documentation overview page + supporting technical article series
Produce a public, indexable developer overview (on the developer portal, but SEO‑discoverable from the main site) titled along the lines of “Integrate e‑Invoicing with APIs.” It should:
- Explain supported e‑invoicing flows (send, receive, status updates)
- Outline how the platform works with PEPPOL/InvoiceNow and other networks at the API level
- Describe integration patterns with ERP or procurement systems
- Link into existing API reference docs
- Include code examples or diagrams
Follow up with 1–2 technical blog/guide articles per key use case (e.g., integrating e‑invoicing into an ERP, automating AP with e‑invoices).
Content Enhancement
High Priority
Existing Content: regional e‑invoicing product pages (AU, NZ, SG, MY, UK, US, etc.) and campaign pages
On each e‑invoicing product/initiative page, add sections that:
- Explicitly connect e‑invoicing to automated reminders, scheduled/recurring invoices, and invoice tracking (reusing language and features from existing invoicing pages)
- Highlight integration with online payments (e.g., Stripe) and how e‑invoicing accelerates cash collection and improves cash‑flow visibility
- Include concrete, quantified benefit statements where possible (e.g., reduced days to payment) and link to cash‑flow guides and calculators
- Briefly mention personalization/branding options available for e‑invoices and link to free invoice template resources where appropriate
Advisor and Industry‑Specific e‑Invoicing Positioning
Medium Priority
Existing Content: advisor portals, industry pages (construction, self‑employed, etc.), and accountant/bookkeeper guides
Add advisor‑ and industry‑specific e‑invoicing sections or sidebars that:
- Explain how accountants and bookkeepers can use e‑invoicing to streamline AR/AP for clients, improve accuracy, and prepare for mandatory regimes
- Position e‑invoicing as an advisory opportunity, with links back to the global e‑invoicing hub and “what is e‑invoicing” guide
- On industry pages (e.g., construction, self‑employed), include short scenarios showing how e‑invoicing works for project‑based billing, retainers, and B2G contracts, with CTAs to relevant e‑invoicing regional pages
Structural Improvements
High Priority
Create a clearly linked e‑invoicing topic cluster and navigation entry
- From main invoicing pages (e.g., //accounting-software/send-invoices/ and ‘explore invoicing software’ pages) and from payments/Stripe pricing pages, add consistent navigation and in‑content links to the e‑invoicing hub and regional e‑invoicing pages
- Implement a small e‑invoicing ‘module’ or banner within invoicing pages that explains in one sentence what e‑invoicing is and routes users to the hub
- This will consolidate topical authority and improve discovery
Medium Priority
Standardize regional e‑invoicing page structures and schema
- Align page structure across AU, NZ, SG, MY, US, and UK e‑invoicing pages: consistent H1/H2 patterns (What is e‑invoicing, How it works in [region], Benefits, Compliance, How to get started)
- Add FAQ sections addressing mandatory rules, PEPPOL/InvoiceNow, registration steps, and integration with existing invoicing features
- Implement appropriate FAQ schema to capture more SERP real estate
- Ensure each regional page links back to the global hub and relevant support/learning resources (e.g., ‘Register to receive e‑invoices’, ‘send and receive PEPPOL e‑invoices’, learning course for Malaysia)
