Introduction
What are the best practices in LLMO? Work your way through the list below to see how you’re doing and where you have opportunities to do better.
P.S. Not yet sure what LLMO even is? Check out The Beginner’s Guide to LLMO first.
1. Make sure your website is accessible to bots
Make sure you’re not blocking search engine or AI bots and that you’re not using technology that’s making it hard for bots to retrieve the content from your pages.
2. Publish testimonials
Get quotes from your most enthusiastic and/or well-known customers and publish them prominently on your website.
3. Build your reputation on review sites
If there are review sites that cover your industry, try to build up a good profile on them. This may mean encouraging more of your customers to leave reviews (but be careful not to break any of the rules that the site has around this.)
4. Invest in your customer experience
With AI helping people to make buying decisions there’s going to be less room for brands with poor reputations to hide. Make sure you’re delivering the best possible customer experience so that customers speak highly of you and don’t complain online. If customers have common complaints, consider putting together a plan to address them.
5. Include citations in your content
When your research to create a piece of content has involved reading other pieces of content, mention those sources (especially if they’re well-regarded). This will boost the authoritativeness of your work and has been found to give as much as 40% more visibility to your content in the output of LLM-based search engines.
6. Include quotations from relevant sources
Including quotations is another tactic that has been shown to make it more likely that your content is cited by AI-powered systems. (Not to mention that a well-chosen quotation can add interest to an article that makes it more appealing to humans as well.)
7. Include statistics
Including relevant statistics can also lend more authority to your content and, perhaps as a result, seems to make it more likely to be referenced in the output of LLMs.
8. Continue to do traditional SEO
The better your content ranks in traditional human-focused search engines, the better it’s likely to rank in searches performed behind-the-scenes for LLMs and, consequently, the more likely it is to influence (or be cited in) the output of these LLMs.
9. Brush up on Bing-specific SEO
Bing is a popular search engine for LLM-based solutions to be using behind the scenes so, if your SEO has traditionally been very focussed on Google, now’s a good time to brush up on the differences between the ranking factors for Google and Bing and adjust your tactics if appropriate.
10. Continue to do PR
The higher the online profile of your brand, the greater the awareness that generative AI systems will have of it and the more likely it will be that they refer to it in their output (even without looking up information online.)
11. Learn about LLMs, RAG, etc.
The better you understand how generative AI-based systems work, the easier it will be for you to see how best to influence them most usefully. Some good places to start are:
- Andrej Karpathy’s “Intro to Large Language Models”, 2023 (1-hour video)
- “A Simple Guide To Retrieval Augmented Generation Language Models” by Joas Pambou, 2024 (article)
12. Build a system for keeping up-to-date with developments in LLMO
LLMO is an evolving space. Make sure you have a way to keep up-to-date with new thinking around how best to do it. For example, follow me on X.
Final Thoughts
I hope you’ve found these best practices helpful. There are certainly more things that you could be doing but I think the list here is a good start to get you going on your LLMO journey.
Let me know if there’s an LLMO best practice that you think I should be including.
Until then, good luck!