A B2B entity resolution and identity data platform came to us with a specific, measurable problem. Buyers evaluating vendors in this category ask AI engines who to shortlist, and this platform barely appeared in the answer. Over nine weeks, tracked against 11 competitor platforms including Informatica, Quantexa, Reltio, Tamr, Senzing, Semarchy, AWS Entity Resolution, Data Ladder, Splink and Zingg, its AI search visibility moved from 3.9% to 20.9%. That is a 5.4x gain, inside one of the most heavily monetised B2B categories we track. Here is what we found, what we built, and what actually moved the number.
TL;DR
- Visibility up 5.4x in nine weeks: from 3.9% (week of 25 May 2026) to 20.9% (week of 20 July 2026), tracked against 11 competitor platforms.
- Mentions grew from 147 to 666, and average position improved from 4.3 to 2.4, so the platform is named more often and earlier in the answer.
- The single biggest fix was technical, not editorial: the platform's most substantial AI-native content sat behind a login wall no AI crawler can read.
- This happened inside one of the most heavily monetised categories we track: 73.0% of answers carry a sponsored placement, all on unbranded research questions.
| Metric | Value | Caption |
|---|---|---|
| Visibility | 3.9% to 20.9% | |
| Mentions | 147 to 666 | |
| Average position | 4.3 to 2.4 | |
| Platforms tracked in category | 11 |
What was the starting position?
In the week of 25 May 2026, the platform held 3.9% visibility against the 11 tracked platforms, was named 147 times across our sampled prompt set, and when it did appear, it appeared late: an average position of 4.3, well back in the answer.
The number on its own was not the useful part. The pattern behind it was. Where the platform's content was public and indexable, it already won. It held one of the strongest average positions in the whole category on the bottom-funnel prompts about scaling entity resolution to hundreds of millions of records in real time. But its most detailed explanatory content on AI-native identity resolution, the material written to answer exactly the questions a technical buyer researching this category types into an AI engine, sat behind a login gate. AI engines cannot crawl a page behind authentication. The platform was close to invisible on an entire growing subtopic, agentic AI and MCP-era entity resolution, not because the content was thin, but because it was unreachable.
Who else is answering these questions?
| Platform | What it does |
|---|---|
| Informatica | Enterprise data quality and master data management |
| Quantexa | Decision intelligence and entity resolution at scale |
| Reltio | Cloud data unification and MDM |
| Tamr | Machine learning powered data mastering |
| Senzing | Embedded, real-time entity resolution |
| Semarchy | Master data management platform |
| AWS Entity Resolution | Managed cloud entity resolution service |
| Data Ladder | Data quality, matching and deduplication |
| Splink | Open source probabilistic record linkage |
| Zingg | Open source machine learning entity resolution |
Eleven platforms in total, counting the one this case study is about. Some are enterprise incumbents with a decade of category presence. Two are open source projects with a very different citation profile. All eleven were tracked with the same daily prompt set across the AI engines buyers actually use to research this category, including ChatGPT and Gemini.
What did we actually do?
Nothing here was a single lever. It was closing an access defect, rescuing content that was already half-working, and then systematically filling the gaps a straightforward audit surfaced.
Fixed the crawl-access defect first. The platform's flagship identity-resolution explainer content was sitting behind a login wall. We moved it onto public, indexable pages, so AI crawlers could reach it for the first time. This was the single change with the widest blast radius: it is hard to be cited for content an engine has never been able to read.
Rescued content that was already half-working. Seven existing articles were retrieved but under-cited: AI engines had already found these pages and pulled them into the retrieval set, but the citation itself, the line an answer actually links to, went to a competitor instead. We restructured each one with comparison tables, FAQ schema and structured lists, so the citation had somewhere concrete to land.
Built net-new content mapped to specific gaps. Rather than general marketing content, each new piece targeted a prompt where the platform scored zero: an explainer on entity resolution for agentic AI and MCP-era workflows, a buyer's guide for AML and KYC use cases in financial services, a piece on stopping AI agents from confusing customer records across systems, and a comparison of real-time versus batch matching. Each was written to close one identified gap, not to be evergreen marketing copy.
Matched the format the engines already reuse. A rival's eight-vendor comparison matrix was the only page being cited for a specific shortlist prompt. We rebuilt a comparison page in the same extractable format, with the same structural elements AI engines already pull from that competitor's page, so ours had an equal shot at the citation.
Placed an editorial byline. A guest article was reviewed, accepted and published in an established AI industry publication, adding a third-party citation source outside the platform's own domain.
Kept a social cadence running. Regular LinkedIn posts pointed back to the new material, which does not move an AI engine's index directly but does build the kind of discoverable trail engines increasingly weight.
What actually moved, and by how much?
| Metric | Week of 25 May 2026 | Week of 20 July 2026 |
|---|---|---|
| Visibility share | 3.9% | 20.9% |
| Mentions captured | 147 | 666 |
| Average position | 4.3 | 2.4 |
Visibility is up 5.4x. Mentions are up 4.5x. And the average position improved by nearly two full places, from 4.3 to 2.4. That last figure matters as much as the headline number. A platform can grow its mention count while still being buried at the bottom of every answer that names it. This one did not just get named more. It got named earlier.
Why does the ad context matter here?
Entity resolution and identity data infrastructure is one of the most heavily monetised B2B categories we track. 73.0% of the answers we sampled in this category carried a sponsored placement, every one of them on an unbranded research question, the kind a buyer types before they have a shortlist. LiveRamp, ZoomInfo, BigID, OneTrust, Arovy and Armor are all buying that inventory.
| Category | Value |
|---|---|
| Waste, dumpster and hauling software | 77.8% |
| Entity resolution and identity data infrastructure (highlighted) | 73% |
| Finance automation, treasury, AP and payroll | 7.1% |
| UK trades job management and gas compliance software | 3.6% |
| London hospitality, wine bars and restaurants | 1.9% |
Growing organic visibility 5.4x while several competitors are paying to sit in the same answers is a different kind of result than growing visibility in a quiet category. Nobody handed this platform its citations. It earned them while rivals were renting the ad slot next to them.
What does this mean if you compete in a monetised B2B category?
Two things worth taking away, in order. First, check crawl access before anything else. A login gate, an SSO wall, a form fill required before content loads: any of these can make genuinely strong content invisible to an AI engine, and no amount of new writing fixes that until the access problem is fixed. Second, paid placements and organic citations are not competing for the same budget line, but they are competing for the same reader's attention inside the same answer. A category with heavy ad spend is not a category to avoid. It is a signal that buyers are actively researching there, which is exactly where a genuine information-gain advantage compounds fastest.
FAQ
What is a B2B entity resolution and identity data platform, and why does its AI search visibility matter? Entity resolution platforms match and merge records that refer to the same real-world person, company or account across systems that were never designed to talk to each other. Buyers researching this category increasingly ask AI engines for a shortlist before they ever fill out a demo form, so whether a platform gets named in that answer, and where in the answer it gets named, now shapes which vendors even make it to a sales conversation.
How much did this platform's AI search visibility grow, and over what period? Visibility rose from 3.9% to 20.9% over nine weeks, a 5.4x gain, tracked against 11 competitor platforms across the AI engines buyers use to research this category.
What actually moved this platform's AI search visibility in this window? The single biggest fix was removing a login gate that had made its best content unreadable to AI crawlers. Alongside that: restructuring seven existing articles that were already being retrieved but not cited, publishing new content mapped to specific zero-visibility prompts, matching a competitor's winning comparison format, and placing an editorial byline in an AI industry publication.
Why is the entity resolution and identity data category one of the most heavily monetised categories tracked? 73.0% of the answers sampled in this category carry a sponsored placement, all on unbranded research questions, the exact moment a buyer is forming their shortlist. LiveRamp, ZoomInfo, BigID, OneTrust, Arovy and Armor are among the companies buying that inventory.
Does growing organic AI search visibility still matter if competitors are buying ads in the same category? Yes. An organic citation and a paid placement are answering the same question for the same reader, but only one of them can be earned rather than bought. Growing organic visibility inside a heavily monetised category is a stronger signal than growing it inside a quiet one, because it means the content itself, not a media budget, is winning the citation.
If you want to see where your own category stands, the free AI search audit shows your current visibility across the AI engines your buyers actually ask.
