DEJAN: The new Era of AI Search Optimization - and how to actually sell it
- vor 6 Tagen
- 4 Min. Lesezeit
Key Takeaways:
Must watch: Petrovic outlined in a webinar from Edge of Search how AI Search Optimization works, what to focus on and why (AI) SEO experts have all the right tools to optimize a brands performance
He compares monitoring model perceptions via prompts with human psychology and stresses how new tactics on top of traditional SEO can move the needle to incease citations, mentions and recommendations
Why AI SEO?
"When AI Search or Agents become personal assistants that users consult before making a buying decision, companies have to influence the brand perception of those Agents or rather AI Search Models and Retrieval Systems and the offers or brands they recommend. Ads will not do the trick" - David Epding
AI Chats provide a new way for users to search and build opinions about brands and the next buying decision
AI Chats answer either based on their training data with a model and/or via Grounding based on retrieved data from Search
Trainind data: even without further information the model has preferences for certain brands and associations with your brand that you should map and optimize for
Search retrieval can be influenced by SEO plus GEO/AI SEO tactics based on Fan Out-Queries
AIO SEO helps to analyze and monitor the different parts of the AI Search technologies that are responsible for how the models answer and how to optimize their answers in favor of you brand
AI SEO in a nutshell - from analysis to action
Analysis & Monitoring is key
Understand the model. Understand each part of the agentic harness and how your brand is visible within different models and modes:
Map associations of your brand within AI Search models - which entities your brand is currently connected
Map and rank competitors for each entity - which entities your currently not considered relevant enough
Get Fan Out Queries or synthetic queries for each entity (and create/connect them to prompts) to understand how an AI Search is doing retrieval
Get Citations (for Fan Out Queries) to undestand how visible you are within the retrieval system (with websearch) as well as the model knowledge (without websearch) compared to competitors and which 3rd party entities influence recommendations / mentions / ranks
Get Cited Snippets to understand what text and formats get picked - not all of your page is retrieved, only concise, matched piece - and where your website or entity lacks precision or semantic space s
Targeted Actions & Measures follow
Only if you know the strength and weaknesses of your brand and of your competitors, if you know how and what AI Search models know or use to form an answer from, you can carry out measures to influence it.
But keep in mind: Correlation is not causation!
Optimize Content On-Page:
Check Chunks: You should be using information and entity rich language, creating smaller, concise paragraphs with sentences that are citable which means they can stand for themselves and not too complex but easy to read out loud in one flow. (even though "chunk" is not the most precise language to use, call it modular retrievable snippets)
Check Entities: You should use the exact entities your brand is and should be recognised for or connected to. Use these terms and phrases consistently to strengthen them within the stochastic brain of the model. Even small adjustments to phrasing can have a high impact on retrieved snippets. Otherwise there is a risk of Brand Drift - a misconception or misrepresentation of your brand by the AI Search model.
Address Fan Out-Queries: Do not overly optimize for exact Fan Out-Queries
Optimize Content Off-Page: Check how your brand or product is referred to on the web including social media (YouTube, Reddit, LinkedIn, ...) as well as News publisher, magazines or high authority pages such as governmental sites.
Naturally earn mentions within your space or the spaces of your customers (primary data, studies, interviews, etc.)
Create helpful content on other platforms such as YouTube or LinkedIn - check which platforms are already cited in your country and industry
"Decision makers have to understand that  AI SEO is a branding channel  - rather than a performance marketing channel - that has the  power to influence recommendations  straight inside the new major, agentic touchpoints"  - David Epding
AI SEO Monitoring
Share of Voice vs. Mention Share vs. Citation Share
Rank (based on calculated Index)
Entities cited for
Citation Share is a leading metric that first can be influenced, recommendation or rather mentions take time but follow up.
Track both:
Entities / Brand Associations
User Prompts (approximation) and Fan-Out Queries
10 AI SEO Myths
"AI SEO is brand building - You do not expect a measurable impact on your brand's perception from your TV commercial, YouTube Ad or Event Sponsoring within days, do you? Similarly, AI SEO needs to be consistently pursued over time to build your brand and grow your business." If you can: Ask leads to name their touchpoints and measure how AI Search is becoming stronger and AI SEO is improving agentic recommendations in the long-term - David Epding
Sources: