How will the relationship between public relations and artificial intelligence play out in 2026? In 2025, we saw the potential of earned and owned in the future of large language model-driven search.
As we head into 2026, with AI search and Generative Engine Optimisation (GEO) on top of mind, Telum Media spoke with Matt Collette, CEO and Founder of Sequencr AI, to hear about what this future of PR holds and more importantly, how industry professionals can best prepare for the journey ahead.
GEO has emerged as the latest frontier in PR. What does it mean to create for large language models and optimise for GEO?
What’s changed is how differently the average person can now manipulate search results as compared to before. To understand GEO, communicators must first understand the changes in search.
Search used to be basic. Google “pet food” with a few keywords, and product images and links would pop up, along with search engine optimised results. You’d go through a journey of opening the links, consuming that content, and synthesising the information.
With AI search, you can end up with results from 10 different websites at once, the information already synthesised by the LLM and returned as digestible information.
The kinds of questions people ask have also changed dramatically. A 2024 study found that 70 per cent of ChatGPT search queries were completely unique. Users are also not only asking very different questions but query stacking as well. Moreover, AI models now serve as filters that suggest next courses of action.
We’ve gone from a world where we key in a search, get a variety of results, and go through the information journey, to having AI synthesising those results and guiding the search journey.
In this reality, companies and communicators need to start considering: how have queries changed? How are they being stacked? How will they evolve? What are people asking about your company? What owned or earned content will prompt LLMs to trust, retrieve, and provide to their users? Search is no longer static, because the ways people are searching have changed dramatically.
As AI search continues to shape corporate reputation, managing GEO means managing the way that people think about your company.
When it comes to PR in the age of GEO, it's more than uploading a press release. How do storytelling, reputation, and thought leadership shape GEO’s performance and strengthen brand visibility in AI-driven search?
There’s a lot of debate inside companies: who owns GEO? Marketing? The search engine team? Or is it comms?
The answer is all. Different models surface different issues, and different teams have responsibility for the different types of content and communication. Generally, what LLMs say about you is almost a representative sample of what the Internet thinks about you, and PR is essential in shaping this.
To get a sense of what models are currently surfacing about their company, communicators should conduct an audit and use that to inform their strategies, pitches, and media engagement, especially with publications that have authority within AI search.
In addition to earned and news, owned content is an important source for generative AI, which comms teams are also responsible for shaping. Comms will need to review dated blog entries and old corporate newsrooms and whether such information should be crawled by LLMs.
Another way to strengthen brand visibility is something that communications already does on a day-to-day basis. What LLMs love about earned and owned content is communications’ logical way of writing. Take a press release as example: the key news in the first paragraph, followed by a funnel of information. LLMs - and the way that they learn - prefer this structured approach.
In this new world, comms leaders will need to constantly consider how information is being discovered, how corporate reputation is being evaluated, and how purchase decisions are being made.
What factors go into how AI-led search perceives brand reputation and trust?
Different factors go into which sources end up being referenced. Communicators need to consider what might affect the questions asked about your organisation, especially those queried most.
For example, people turn to AI search before making B2B purchases. The LLM will quickly surface and synthesise both positive and negative details about a provider, including partnerships and product quality.
Another factor is the model you’re optimising for. Each has rules on what’s considered authoritative content. ChatGPT sometimes references Reddit; Gemini will reference social media content - specifically YouTube, X, and Instagram. Of course, reviews across websites have impact as well.
Generative AI is non-deterministic, and one query is not always representative. With the same prompt, you’re going to get a different response each time - different series of words, combinations of facts, and sites surfaced.
Are there ways to measure AI share of voice and monitor brand health effectively? For teams adapting to GEO, what steps can they take?
The best thing is to get an audit to see how your organisation is represented. Then, you can plan a GEO-informed comms strategy to address the findings.
Each model has a different way of presenting positive or negative information. Sometimes, Gemini is more negative than ChatGPT or vice versa, depending on the question and company. Setting up dashboards can help with monitoring reach, visibility, and sentiment across different LLMs.
Another thing is to revisit your website’s structure. Schema markups make a big difference in how LLMs surf and identify information on a webpage. Even technical tweaks to facilitate information retrieval and machine readability can improve how models represent information about your company or products.
As GEO continues to evolve, what's a suggestion you have for communicators concerned about the ethics or perception of AI usage in PR?
One issue that might arise is one of copyright infringement. There are ongoing cases looking into whether companies might have violated or circumvented them as part of training their models.
For communicators, it’s always good to keep in mind how you’re prompting models. That’s where the ethical lines start to come into play: is there intent to violate copyright when you are prompting? Or is the intent to create original content based on what’s been released by your company?
As practical guidance, write prompts that don’t infringe copyright and steer away from the ethical challenges that others have already gone through.
How would you suggest agencies and in-house teams go about implementing, or even just considering, AI programs to get a head start in 2026?
To start, the corporate policy environment is very important.
Back in 2022, many companies had knee-jerk reactions to generative AI. Policies tended to be restrictive and negative - “don’t use ChatGPT at work” or “it’s going to produce outputs that may infringe copyright.” These policies have had negative outcomes, with employees feeling uncomfortable discussing or sharing how they use AI at work and feeling undervalued if they do.
But there are various AI policies that you can write, ones that are functional, strategic, and in part, mission statements. For example, there are some very compelling ones from organisations concerned about the environmental impact of generative AI and how to manage that. Through constructive AI policies, you can incentivise value-aligned behaviours so that increasingly, you become an AI-native organisation.
Proficiency is another big issue. Statistics show that 70 per cent of generative AI users are novices, and only 10 per cent of all users are considered proficient. So, training remains an important area. I’ve seen many companies have an expectation that people will use AI more, without investing in helping people understand how to do so.
Most people think that generative AI is best used to produce content: “help me write an email,” “help me write a blog post.” But generative AI goes beyond that. For example, with simulation and prediction, you can run messaging by it to see potential audience reacts.
If there’s one thing I would recommend, it’s to understand the technology so that you can see the full spectrum of possibilities with it, because it is quite exciting when you start to apply it to these different things.
After generative AI, you can move onto the implications of agentic AI. Most people, when talking about Generative Engine Optimisation, stop at ChatGPT. But beyond that, there’s a world of agentic processes where models can execute tasks for users, and it will have a big impact on the information economy and search.
This is the world that we’re heading into, and that’s where comms leaders need to think towards: how different the world will become and prepare for that, versus how the world currently is and reacting to the present. As an industry, we need to become more future-oriented.
Hear more from Matt Collette in Telum Media's upcoming webinar (5th February) looking at the Top AI tools for 2026. Register for free here.
You might also enjoy
-2.png)