How to build influence with Generative AI

This article outlines a general approach to optimizing content strategy and digital assets for generative AI. It sets out how companies can gain visibility in user queries on Chat GPT, Microsoft Copilot and other large language models (LLMs).

Is Chat GPT coming for your website? While AI-powered LLM tools only represent a fraction of overall searches today, they're seeing vertiginous growth. The number of daily Chat GPT prompts more than doubled in the first half of 2025 alone, and a recent study estimated that 5.6% of U.S. desktop search traffic already go to LLMs.

But getting your messages in front of Chat GPT users is not just about optimizing your corporate website. To answer queries about your business, your industry or your chief executive, LLMs draw on a variety of sources to create a summary. If they do it well enough, no-one will need to visit your website at all. That's a nightmare scenario for corporate communications departments, who are used to controlling their company's reputation through a combination of well-rehearsed key messaging and website optimization.

As such, we believe that LLMs will change search optimization radically. While well-written corporate web content that adheres to SEO guidelines remains important, strong visibility on LLMs will depend on a coherent presence across multiple owned and external sites.

Building influence via Chat GPT: what are your goals?

Leveraging Chat GPT to your advantage is more about building an influence footprint online than tweaking existing web content.

Let's say your company is an energy provider with a diversified business spanning renewables and conventional power across several European countries. We'll call it ALPHA ENERGY. Your goals for gaining visibility on LLMs are likely to be twofold.

On the most basic level, you want anyone who types in the question "Who is ALPHA ENERGY?" to get an answer that is a) accurate and b) in line with your messaging. This is comms hygiene.

Now notch it up a level. What about if a user typed in: "Who are the top renewable energy companies in Europe?" or "What big renewable energy projects are being developed in Europe at the moment?" Ideally you would want ALPHA ENERGY to feature in the response.

To develop a Chat GPT-friendly comms strategy, you therefore need to consider what questions your audiences are likely to ask about your industry and your business.

How do LLMs prioritize sources?

To build a Chat GPT-friendly digital footprint, you first need to understand how LLMs source information.

When asked about an industry, ChatGPT and similar LLMs rely heavily on the following types of sources (roughly in order of reliability and frequency of reference):

References and knowledge bases:
Used: to define an industry and provide essential facts.
Examples: Wikipedia, Britannica, Statista, OECD, World Bank.
Media (including trade press)
Used: for current developments, trends, challenges, and expert quotes.
Examples: The Economist, Financial Times, Industry Today, Chemical & Engineering News.
Think tanks and research institutes
Used: as a source of insights, forecasts and conceptual frameworks
Examples: McKinsey, BCG, WRI, IEA.
Government and NGOs
Used: to define a sector and current trends, particularly in regulation-heavy industries like pharmaceuticals or finance.
Examples: Eurostat, INSEE, UNEP, EU Climate Targets, EPA.
Academic articles
Used: for technical points and innovation topics
Examples: The Lancet, Nature, ScienceDirect

When asked specifically about a company, Chat GPT prioritizes sources slightly differently. While Wikipedia and authoritative media outlets still dominate (particularly sites known for reliable corporate information such as Reuters and Bloomberg), a company's website will usually come in as a reliable third. Press releases are also seen as reliable source. Social media content is a lot less visible (although may be considered important if it is frequently cited by other authoritative sources).

How to improve visibility on Chat GPT: a step-by-step guide

Step 1: Carry out an audit and define an overarching LLM strategy

This is about defining what messages you want to be associated with the company and assessing how visible you currently are on the most important LLMs. Start by defining:

the factual information that should appear when users search information on the company (business description, main products / services, leadership team, strategy etc.)
topics on which the company wishes to be referenced and the terms users are likely to use when searching
any specific objectives relating to positioning (e.g. #1 in Europe, more sustainable / innovative than competitors)

Then carry out an audit across the main LLMs (Chat GPT with a share of 60% of the generative AI chatbot market; Microsoft Copilot with around 14% and Google Gemini with around 13%) on the factual information, topics and terms you are looking to rank on.

The results can be used to create a gap analysis and identify opportunities for optimization - then develop a company-specific strategy.

Step 2: Optimize your corporate website

Your corporate site may not be the most important source for LLMs- but it's the easiest place to start.

When it comes to corporate websites, LLMs prioritize structure, clarity, and crawlability. They do not assess visual design. Many of the features that appeal to LLMs also appeal to human readers, so a solid Generative Engine Optimization (GEO) web strategy should not conflict with your core audiences' needs.

Begin with a thorough content audit of your website to identify areas for optimization, and quick wins. Specific pages (e.g. About Us, Strategy, Sustainability or ESG landing pages, and financial results overviews) tend to perform well, so it can be helpful to prioritize them. Chat GPT and other LLMs also like specific formats, like glossaries or knowledge pages that provide easy answers to questions (for example: "What is sustainable aviation fuel?")

What are you aiming for when optimizing for GEO?

A page that provides enough information to be considered relevant and important- which in practice means upwards of 600 words
A clear structure, with a well-defined hierarchy of titles and subtitles
A crisp 1-2 sentence summary as an opening paragraph (this is often extracted for responses)
Use of key phrases aligned with how users might ask AI bots about the topic. (e.g., Instead of Our environmental impact, write Our approach to reducing CO₂ emissions from food packaging)
Sentence structure in line with editorial and standard SEO best practice (simple, short and active)

LLMs will also look for signals that demonstrate reliability and accuracy. These include outbound links to other reputable sources, and timestamps that show when the page was last updated.

Finally, consider developing some content formats that particularly appeal to LLMs. These include CEO messages or open letters like this oped by the CEO of Unilever or Salesforce's Chair and CEO Letter to Stakeholders.

Step 3. Invest in PR

Because LLMs rank media coverage highly, traditional media relations are starting to increase in importance again.

While the FT, Economist, Reuters etc. are clearly the holy grail for large companies on business and IR topics, we recommend also focusing on trade media. Not only does this approach increase the likelihood of business and product topics being picked up directly by Chat GPT and the like, it also provides sources for Wikipedia pages to reference.

To support PR efforts, create press packs on products and partnerships that go beyond a short media release. You may also consider carrying out surveys or research with a view to generating statistics and insights that can be quoted by media, and published on your own website.

What about Wikipedia?

A Wikipedia page containing the company’s messages, key products and innovations, with references to major media coverage is one of the most useful assets to improve rankings.

Wikipedia asks that contributors be neutral, and flags any suspected editing by the company or a paid advisor at the top of pages. Any edits to a company's Wikipedia can therefore only be made by contributors with a track record of editing pages- although using the "Edit Request" feature, companies can suggest edits that the Wikipedia community can choose to adopt.

Step 4: Triangulate

Triangulation refers to the process by which large language models and AI search systems validate and elevate information based on its presence in multiple independent sources. This means ensuring the same information appears on the company website and on authoritative external sites such as major media or Wikipedia.

There is high value for partnering with external organizations, notably think tanks or large consulting firms. These partnerships can focus on sharing data and use cases with a view to gaining visibility in external research reports.

Securing your place in Chat GPT answers

The new era of AI search inevitably means some loss of control for corporate communicators, who risk seeing their carefully-crafted messages paraphrased by AI bots. But by developing a comprehensive strategy to enhance visibility on Chat GPT and other platforms, they can potentially extend their influence.

AI search logically favors large companies: they are the ones with the giant digital footprints, extensive media coverage and a detailed Wikipedia page. The trick is to carefully cultivate that footprint so that messaging is up-to-date and replicated across multiple platforms.

HOLLIS & BEAN works with clients to shape AI-ready content strategies and optimize corporate websites for the new era of AI search. Reach out to us to discuss your needs.

Last updated: September 2025