How Digital PR and Affiliates Won AI Recommendations

PR has always been a visibility game. We compete in inboxes, newsfeeds, and SERPs. Now there’s a new arena: the AI chat window.

With 58% of consumers using GenAI for product research, the answers these models give are shaping what people buy, and if your brand isn’t in those answers, you’re invisible at a key point of consideration.

Fortunately for PR teams, when someone asks an AI for a recommendation, it usually doesn’t just invent one. It synthesises an answer from sources it considers trustworthy, and it overwhelmingly leans on publisher content to do that.

In fact, research shows that up to 90% of LLM citations reference publisher content. Over a quarter of all AI citations for product-led queries come from listicles and “best of” guides. Why? Because they’re updated frequently and are easy for the models to summarise. 

So how do you improve your odds of being recommended by AI?

You stop working in silos.

Breaking down the old silo problem

Let’s be honest: most brands still treat Digital PR and Affiliates as separate channels. Both teams speak to publishers, but they operate with different goals:

• Digital PR: earns coverage, builds authority, supports SEO

• Affiliates: secures paid placements, builds commercial relationships

Usually these teams barely speak, and historically you could “get away with it”. However, the rise of LLMs as a product discovery engine has created a critical reason to merge these workflows.

Our integrated approach

We built a three-step process to bridge PR and Affiliates and increase the chances of being included in the types of articles AI commonly summarises.

Step 1: Map the AI’s most influential articles 

We treated LLMs like a discovery engine.

Using tools like Ahrefs alongside recommendation-intent prompts (e.g. “best [product] UK”, “[product] for [use case]”, “[product] vs [product]”), we built a list of the exact URLs that mattered - not just the publisher domains.

We prioritised pages based on:

• Topical relevance and overall category fit

• Authority, including whether the page consistently ranks or gets referenced

• Update frequency, such as “last updated” signals, seasonal refreshes, and living guides

• Are easy for AI models to summarise, such as listicles and comparisons

The output was a live target list of URLs both teams could work from.

Step 2: Equip Affiliates with PR assets that add credibility

PR didn’t just hand the target list to Affiliates and hope for a commercial win.

Instead, we packaged the kinds of assets that help publishers upgrade their pages (and make them harder for AI to ignore):

• Expert commentary backed by clear credentials

• Original data with a transparent methodology

• Timely insight, including search trend context and seasonal angles

The goal wasn’t “get mentioned”. It was: improve the quality of the page while getting included in the update.

Step 3: Secure placements that improve the content 

Once Affiliates had the targets and the assets, the conversation with publishers changed. They weren’t just negotiating a placement. They were enabling journalists to:

• Refresh outdated content

• Add expert-led detail

• Strengthen comparison criteria

• Improve sourcing and clarity

This meant our brand was being featured alongside authoritative content, improving credibility signals to Google, LLMs, and consumers.

The Results

It wasn’t just about links, although link growth came as a by-product. It was about aligning two teams to hit KPIs across visibility and commercial performance.

What we saw from the integrated approach:

• AI visibility: measurable uplift in brand mentions across a wide set of keyword prompts where we appeared in 45% of AI responses

• Commercial impact: affiliate traffic and conversion performance increased (% uplift, rather than raw totals)

• Organic uplift: search visibility improved by 137% YoY

We also witnessed a standout moment when we were explicitly labelled as a “best” option in an AI answer

How we measured “AI visibility” 

AI answers aren’t static. Ask the same question twice and you won’t always get the same response. That’s exactly why “AI visibility” can’t be measured casually.

So we didn’t.

Instead of chasing screenshots or cherry-picking examples, we used a method we could explain, export, and repeat.

We started by defining a set of high-value, recommendation-intent keywords — the kinds of searches real people use when they’re close to choosing (think “best X,” “X for Y,” “X vs Y”). These weren’t vanity terms. They were commercial, competitive, and tied directly to product discovery.

Those keywords were then fed into Ahrefs Brand Radar, which monitors brand mentions across AI-generated responses. From there, we exported the full dataset of AI outputs tied to that keyword set and calculated our client’s share in those results. 

Why it worked

Three reasons:

1. We prioritised the sources AI is most likely to use. These were third‑party pages that already exist to answer, “what’s best?”

2. We raised credibility on those pages with expert quotes, primary data, and clear sourcing.

3. We turned two separate channels into one workflow, so research, assets, and publisher access moved together.

This post was written by Alex Lovesy, Digital PR Account Manager at EssenceMediacom North

Enjoy learning more about Digital PR? Subscribe to The Digital PR Observer Newsletter to stay up to date with all of the latest Digital PR news and tips!