Editor’s note
Insights from a marketing professional
AI and large language models (LLMs) have taken marketing teams by storm. They’ve reshaped how B2B buyers research, evaluate and shortlist suppliers – something that often occurs long before any direct interaction with a brand.
This new way of AI-supported thinking is becoming the new norm – according to a Forrester report, almost 95% of buyers anticipate using genAI to support their decision and purchase process in the next 12 months.1
Marketing teams now face a flood of expert opinions on how to respond – whether that’s preparing for generative engine optimisation (GEO) and language engine optimisation (LEO) or doubling down on the ‘human touch’. But the implications of LLMs reach way further than brand visibility or channel tactics. This isn’t just about adding another touchpoint to the buyer journey. As buyers increasingly turn to AI-powered tools for early-stage research, a critical question emerges for communications leaders: What does this mean for my B2B marcomms strategy?
A comprehensive answer to this question would require much more space than this article allows, so I’ve decided to focus on a question that has been lingering but few have voiced out loud:
What’s the role of trade media with the advent of LLMs? Should we continue to include them in our communications plans, and if so, how?
Chapter 1. Why trade media still matters
Trade media in B2B, especially in technical sectors, has a huge role to play in comms planning. Whenever I’m working on a brief that includes awareness, thought leadership or lead generation, I would really struggle not to include an element of trade media in it. But does this approach still stand in the new AI world?
Before LLMs became mainstream, B2B buyers would browse the web in search of technical journals, trade media magazines and whitepapers that caught their attention – or in marketing jargon – ‘addressed their challenges and pain points’. Now, the same audience that would download dozens of whitepapers or brochures is increasingly asking LLMs for the answer first.
Critically, this is not a question about whether more traditional channels like trade media still matter, but about what job they now perform in a fundamentally altered buying journey.
Chapter 2. What you need to know about how LLMs work
If you want to appear in LLM-generated answers, you need to understand how these systems actually find and prioritise information.
Where do LLMs look for information?
Popular LLMs like Google Gemini and ChatGPT simply look for information available online. However, each LLM relies on different indexing strategies and therefore builds its knowledge based on different website sources. For example, Google Gemini Pro 1.5 indexes more heavily on Reddit, YouTube and Amazon, while ChatGPT 4o relies more on Google.ii That means that while there is an emerging field of GEO or answer engine optimisation (AEO), good old search engine optimisation (SEO) still holds a place in the ChatGPT index, because it is central to getting content ranked on Google.2
How do LLMs prioritise sources?
Forget ranking, think about patterns of repetition. LLMs don’t maintain a ranked list of sources. Instead, they generate answers based on patterns learned during training, plus the context provided in a prompt. Generally speaking, information is prioritised using three key principles:
LLMs look for information that best matches the intent and context of the query.
LLMs favour information that appears consistently across reputable sources and reflects consensus in trusted domains like scientific literature or established business frameworks.
This is the multiplier. If your prompt mentions ‘I’m in food and beverage, speaking with a CDMO about my next plant-protein drink,’ the LLM immediately narrows its focus to B2B food manufacturing, CDMOs and plant-protein formulation. The more specific your context, the more tailored the response.
Chapter 3. Real-world testing: How LLMs handle B2B queries
To bring this to life, I’ve drafted two prompts that manufacturers in different stages of the buying journey may run through an LLM if they’re looking for plant protein ingredients. I then ran it through ChatGPT, Gemini and Perplexity to compare results.
Prompt 1: Early-stage research
Prompt: ‘Can you help me select a plant-based proteins supplier for large scale manufacturing?’
Answer:
Here, the prompt is looking for early-stage information, mostly to understand what options are available and learn about a few reputable suppliers. ChatGPT selected three multinationals – sensible for large scale manufacturing. While this answer is certainly not wrong, you don’t have to be Cargill or Roquette to allow for large scale manufacturing – revealing a degree of simplification that’s typical of LLMs. Plus, there’s an assumption from the LLM model that I’ll need a certain degree of global distribution as well as regulatory and supply chain support – an assumption that is correct. When prompted, Perplexity gave a similar list, and so did Gemini, with the addition of Kerry and Ingredion. So, how did they reach these conclusions?
When I looked into the sources consulted to form the answers, industry and supplier databases came up first for ChatGPT – with Ingredients Network and FoodNavigator being cited as obvious examples. The amount of information available on products and processes (highlighted as a sign of transparency) cross-checked the relevance for food ingredients vs the mention of nutrition ingredient in the company’s ‘owned’ communication channels.
Gemini instead selected companies that are consistently ranked as top players across multiple independent market research reports and industry analyses like MarketsandMarkets, Mordor Intelligence, The Good Food Institute (GFI), alongside financial and corporate news.
Prompt 2: Technical comparison
Next, I asked ChatGPT to compare two different types of plant proteins: ‘Compare pea protein isolate vs faba bean protein for beverages and protein fortification.’
– Peer-reviewed scientific papers (understandable given the technical nature of the topic)
– Technical and regulatory literature from ingredient suppliers
– Market and process insights from industry trade media and formulation guides
Gemini instead went as far as giving me its preferred option after providing a detailed overview and cited sources to formulate its conclusions, such as The Good Food Institute, MarketsandMarkets, Batory Foods as well as Forbes Vetted and Health.
So, with minimal mentions of trade media as primary sources of information, does this mean it’s becoming obsolete? Quite the opposite. Reliability of information and context are extremely important in the neural network of LLMs – and this is where trade media can have meaningful impact.
To verify this intuition, I asked Perplexity directly whether trade media is consulted for search results on ingredient queries – and this is what it said:
‘Yes, trade media like FoodNavigator or FoodIngredientsFirst appears in web search results when relevant to protein suppliers and industry trends. For supplier questions, such sources provide timely insights into companies like ADM and Roquette through sponsored content and market coverage.’
‘FoodNavigator frequently highlights protein ingredient providers in articles on innovations, such as ADM’s SojaProtein soy offerings and Roquette’s pea facilities, aiding validation of large-scale capabilities. These complement market reports by offering real-world production updates and sustainability claims, though primary reliance stayed on direct supplier data and rankings for objectivity. It proves useful for emerging trends but secondary to comprehensive reports.’
In other words, being consistently mentioned on trade media matters, just like being mentioned on market reports isn’t enough.
Chapter 4. The evolving role of trade media
Historically, trade PR (and content-led advertising on trade media) broadly served three functions: (1) building awareness and credibility, (2) search visibility (via backlinks and SEO), and (3) reputation and profile building. While traditional SEO exercises are still valid, the real focus of PR planning moving forward will be to build credibility and reputation. With AI-generated content flooding the internet, the real differentiation will come from having your content associated with an authoritative voice.
So, while people may ask ChatGPT instead of ‘Googling’, the model’s ability to form an answer depends heavily on what authoritative content exists about your brand in the wider ecosystem. These new channels shift the role of PR from ‘visibility’ to ‘training the ecosystem’. Trade media still represents an important anchor of contextual information. If any brand consistently appears in respected trade coverage as a leader in plant proteins, pharma excipients or texturising systems, LLMs learn that brand is credible, innovative and category-shaping.
Trade media as an AI training ecosystem
We can no longer just look at trade media as a way to secure coverage to reach human readers, instead we need to create authoritative signals that train the AI systems those readers now consult. The familiar channels haven’t disappeared: trade publications, scientific journals and analyst reports all remain critical. But their role has fundamentally shifted from visibility to credibility, from reaching buyers to teaching the algorithms that influence them. And the ecosystem itself has expanded. Platforms we once dismissed are now essential channels, not because of their direct readership, but because of how aggressively they’re scraped by LLMs.
Practical recommendations: How to adapt your trade media strategy
This expanded media mix could include:
Ensures your brand appears in the ‘ranked lists’ that LLMs favor for supplier selection queries.
Publications like scientific journals are increasingly important. Not only because they add scientific credibility; they now also represent an important window of visibility that feed LLMs. In reality, you can’t PR your way into these journals, but it’s an important consideration for science and comms teams to now dedicate even more resources for this type of coverage.
No longer just procurement utilities; they are increasingly crawled and synthesised by LLMs when buyers ask ‘who supplies X?’ or ‘best manufacturers of Y.’ Well-optimised profiles on authoritative platforms help reinforce credibility, category relevance and consistency of claims across the ecosystem.
Widening the net beyond traditional PR channels
Platforms we once dismissed as ‘not necessary for B2B’ now demand serious attention. Reddit, Wikipedia and LinkedIn are channels that PR teams can no longer afford to overlook – not because they offer shiny vanity coverage, but because of how frequently and deeply they’re scraped by LLMs.
On Reddit, discussions in relevant subreddits (like r/Supplements, r/Nootropics, r/foodscience, etc.) offer unfiltered consumer and industry sentiment and often rank highly in LLM training datasets. In fact, Reddit’s public dataset is one of the most commonly cited sources in large-scale AI model training.
Wikipedia is a high-authority domain – often among the top ten sources used in LLM pretraining datasets, including those for GPT-4, LLaMA and Claude.3 It acts as a structured, concise knowledge base. Being mentioned (accurately and with citations) on category-relevant pages or having your own company page if eligibility is met, increases the likelihood that an LLM will ’see’ your brand as a notable part of the landscape.
In addition, LinkedIn brings in both human and machine value. For machines, public LinkedIn content – particularly from company pages and thought leaders – is indexed and often summarised. Posts announcing new products, partnerships or category leadership can shape how your brand is described in AI outputs – and whether it is visible at all.
Chapter 5. Final thoughts
The rise of LLMs does not make trade media irrelevant. If anything, it has given us one more reason to build brand presence in high-profile titles.
While widespread adoption of a new channel may initially drive us to think we need to throw out the old playbook, when it comes to trade media, that would be a mistake. The fundamentals haven’t changed – brands still need authoritative third-part validation. We therefore don’t need to forget everything we’ve learned. Instead, we need to change our way of thinking and start treating trade media less as a visibility channel and more as credibility infrastructure that shapes how LLMs describe our brands. The real change comes from the breadth of platforms that are gaining relevance in B2B PR.
If your brand is absent or inconsistently positioned in those environments, it risks becoming invisible not just to buyers, but to the systems shaping buyers’ perceptions before sales even enter the conversation.
The brands that win in an LLM-driven landscape will be those that treat PR as long-term category infrastructure. Those that don’t may still exist – but increasingly, they won’t be recommended.
