Introduction: Don’t follow the news, follow the changes in user behaviour
AI is reshaping how people find products. Systems like ChatGPT, Gemini, Bing AI, Claude, they’re no longer just assistants; they’re discovery engines. If your future customers are asking an AI where to buy what you sell, you need to make sure you’re showing up with GEO (Generative Engine Optimisation) or ‘LLMO’ (Large Language Model Optimisation). While we’ve all been used to ‘googling’ product or service information, we’re now having in-depth and nuanced conversations with AI systems to get high-resolution reccomendations for the right products at the right time.
Here are three critical things every business should be focussing on now, if you want to compete in this new world.
1. LLM.txt: What It Does – And Why It’s Suddenly Important
What is LLM.txt?
LLM.txt is a text file placed at the root of your website (e.g. yourdomain.com/llm.txt). Think of it as a signal to large language models (LLMs) that helps them understand:
- Which parts of your site you allow the model to use for training or retrieval.
- What kinds of content you want to be considered when that model is generating output or answering user queries.
- What should not be used (e.g. private content, pay-walled material, or outdated info).
Why it matters in the AI world:
- If an AI is crawling content or retrieving from websites, LLM.txt gives you some control over what content is fair game and how it should be utilised in responses to user queries.
- Without a clear LLM.txt, your best content might not be used, or worse, content you don’t want featured might get surfaced.
- It helps protect your brand messaging, product descriptions, and unique positioning. Helps you avoid being misrepresented or out-ranked by less authoritative or derivative content.
What to do now:
Monitor where your content is showing up in AI results (when possible) to see if it aligns with what you’ve allowed.
Audit your website content: what you want used, what you don’t.
Create or update your LLM.txt to reflect that. Be explicit.
2. Traditional SEO: Why It’s Going to Matter More Than Ever
We often imagine AI discovery replacing Google search. That won’t happen overnight. Traditional SEO (organic search, technical SEO, content quality, backlinking) still matters, and at TDS we have a hunch that it’s going to become more important in a couple of ways:
Search as training / source material
AI systems often rely on large corpora of web content. If you rank well in Google and other search engines, that visibility adds legitimacy. Also, search-optimised content tends to get crawled, indexed, shared and linked, which improves its chances of being used as input to AI systems or referenced by them.
Trust and authority signals
AI systems (or the retrieval layers behind them) will favour content sources with strong signals such as domain authority, clear identity, reviews, regular updates, accuracy. Good SEO helps build those signals. Being correctly structured (schema, fast speed, mobile friendly, good UX) matters a lot.
Keyword intent & topic depth
AI users might ask conversational or complex queries. If your content is shallow or just targeting basic keywords, you’ll miss out. Depth, comprehensiveness, updated content, internal linking, pillars/clusters, all these things still give you leverage. High quality content that answers detailed questions will do better when an AI needs to source or summarise.
3. Brand & Product Delivery: The Silent Game-Changer in AI Rankings
In traditional search, brand matters but in the AI era, it could matter even more – especially when it comes to how users interact with the brand in positive or negative ways.
Why brand matters when AI chooses what to show
- AI retrieval systems often consider reputation, past performance, freshness. If you have a strong brand (good reviews, happy customers, consistent content), those things count in a big way as the companeis digital footprint is studied when AI systems to look to figure out what the best solution to the user query might be.
- If users in AI chats or summarisations are shown options, they’re more likely to choose known brands. That user behaviour feeds back into what the AI learns (or which sources the retrieval layer privileges).
Product delivery & customer experience
- It’s not just what you say anymore but what you do. If your product descriptions, images, return policies, delivery times, customer service are excellent, that’s input into signals outside of SEO: reviews, mentions, social proof.
- A great product experience leads to positive reviews, fewer complaints, more referrals. These in turn reinforce your brand’s credibility.
Aligning brand messaging across channels
- When AI displays or summarises your product, inconsistent messaging or weak product info hurts you. Make sure all touchpoints (site content, product descriptions, FAQ, policies and even social media content that is linked to the site) are consistent, accurate, up to date.
- Invest in things like clean visuals, accurate specs, well-written descriptions. These aren’t “nice to have” in the AI-first future, they form part of your ranking signals.
Conclusion
If you’re a business owner wanting to win in the AI-powered discovery era, here’s a quick checklist:
- Get an LLM.txt in place, and decide what content you allow vs what you restrict.
- Double down on traditional SEO: technical setup, high-quality content, topic depth, strong links.
- Build your brand so it signals trust, quality, and consistency. Aim for product experience that earns the kind of reputation AI systems want to favour.
Do this between now and 2026, and you’ll be in the frame when AI-driven purchasers start treating those models as their first port of call. Miss these, and you risk becoming invisible even if your product is brilliant.
