In what seems like only a short period of time, many of us have found ourselves searching on AI to find the answer to a question online, rather than using the more traditional search engines, such as Google. And, often the answers we receive appear more tailored to our particular search.
To look at some figures, according to the site Wix.com, August 2025 saw continued growth for AI search platforms, with combined traffic to ChatGPT, Deepseek, Perplexity, Gemini, Copilot, Grok and Claude, reaching 7.5 billion visits. AI search visits in August grew by 1.9% compared to the month before and are up by 150% year on year, up from 3.5 billion visits in August 2024.
While we write this, standard search is still dominant by a large margin but AI search and AI tools are growing momentum – particularly among younger users.
While we are busy fine tuning how we search on AI and working out what prompts work best for, where does this leave content from thought leaders and good content from companies? How do marketers navigate this new normal and what does this mean for SEO in the future?
Many businesses rely on Google ads to prompt their businesses – when we search online for anything from curry in Nottingham to plumbers in Cardiff. Now, even when searching in a traditional search engine, we often see an AI answer above prompted content and organic content.
The results from an AI agent are arguably less confusing, as you aren’t presented with a number of options to consider. Also, the more you use a particular AI agent – via the prompts/questions you pose it – the better it will get to know you, so the more personalised its answers will become…
One factor that works against AI search engines currently is that they generally have outdated information, don’t always carry real-world data and, as yet, can’t browse the web as efficiently as a traditional search engine, such as Google can. This means the answer it gives you isn’t necessarily that useful or up to date. But, that’s likely to change over time, as AI improves.
At the beginning of October 2025, we asked ChatGPT how up-to-date its data is and received this answer: Hello! I’m running on GPT-5, and my core training goes up to June 2024. But I also have the ability to pull in live web data (like I did for your last question), so if you ask about something current — news, stats, events, product updates, etc. — I can check the web and give you up-to-date answers.”
While ChatGPT doesn’t carry ads like Google currently, the word on the street is that OpenAI (owner of ChatGPT) is considering monetising its platform through paid advertising in the future. This potential change is expected to roll out during 2026, with OpenAI projecting ‘significant’ revenue from this channel by 2029. The type of ad formats this is likely to involve include sponsored content, product recommendations and contextual placements.
Web platforms and technology companies are looking at ways to make websites easier for AI agents to trawl – so their client’s information has more chance of getting in front of somebody doing an AI search. Marketers are considering how best to present content to LLMs (Large Language Models) for them to consume – as LLMs often can’t cope with the structure and context of conventional websites, which has traditionally been optimised for human browsing. This has seen marketers looking to present websites to LLMs in a clean marked down version, making it easier for them to digest.
Alongside this, in 2024, Jeremy Howard, co-founder of Answer.AI, launched the llms.text project, aimed to help LLMs better understand and navigate website content.
On the website, https://llmstxt.org/, Mr Howard explains: “Large language models increasingly rely on website information but face a critical limitation; context windows are too small to handle most websites in their entirety. Converting complex HTML pages with navigation, ads, and JavaScript into LLM-friendly plain text is both difficult and imprecise.
While websites serve both human readers and LLMs, the latter benefit from more concise, expert-level information gathered in a single, accessible location. This is particularly important for use cases like development environments, where LLMs need quick access to programming documentation and APIs.”
He goes on to explain that the llms.text project proposes adding a /llms.txt markdown file to websites to provide LLM-friendly content. This file offers brief background information, guidance and links to detailed markdown files.
Mr Howard adds: “We furthermore propose that pages on websites that have information that might be useful for LLMs to read provide a clean markdown version of those pages at the same URL as the original page, but with .md appended. (URLs without file names should append index.html.md instead.).”
This certainly sounds like the way ahead but there is an issue in that the main LLMs have not confirmed currently that they will use llms.txt.
Some marketing agencies are beginning to offer AI optimisation as a service, in rather the same way as they might offer SEO optimisation.
If you want to talk to us about AI search, then contact Lake Solutions today: 0203 397 3222.