From Rankings to Reputations: Why Being Cited Isn't Enough

In AI search, getting cited is just the start. Users still make purchasing decisions based on trust, and trust is shaped by sentiment across online reviews and brand conversations. In this session, Celeste shares how to use LLMs to automate sentiment insights, and treat sentiment as an optimization layer: translating recurring themes into on-page improvements that increase conversions.

Transcript

2. THE TRUST LAYER Every search is a trust decision in disguise @yourfavoriteseo
3. THE TRUST LAYER A search looks like a request for information. But what is it really? @yourfavoriteseo
4. THE TRUST LAYER A search is someone deciding who to believe. Maybe they're uncertain, and they're looking for something to close that gap. Being present isn’t the same as being believed. @yourfavoriteseo
5. E-E-A-T The framework you already know @yourfavoriteseo
6. E-E-A-T was never four equal pillars Experience Expertise ▼ TRUST THE DECIDING FACTOR @yourfavoriteseo Authority “Trust is the most important member of the E-E-A-T family because untrustworthy pages have low E-E-A-T no maer how Experienced, Expert, or Authoritative they seem.” Google Search Quality Rater Guidelines, §3.4
7. So how do you work on trust? @yourfavoriteseo
8. Reviews @yourfavoriteseo
9. THE PROBLEM Reviews get treated like a reputation task The typical workflow is reactive The richest part is left on the table Respond, watch the star average, track velocity. Everything points at the score, never the substance. The review text goes unread, even though customers felt strongly enough to write it. @yourfavoriteseo
10. WHY NOW AI search made review text more important, not less AI reads the reviews so buyers don’t have to Ask Maps and Gemini for Business Profiles summarize a business from its review text. The words are an input to the machine now, not a badge for humans. “Review sites are the #2 citation source in AI, and their influence grows 10–20× by the time someone is ready to buy.” SEER INTERACTIVE STUDY @yourfavoriteseo
11. KEY FINDING The echo eect: review content flows straight into AI narratives AWARENESS STAGE 1.51% of citations Source: Seer Interactive @yourfavoriteseo → INTENT STAGE AI narratives amplify Trustpilot sentiment of citations Amplification runs both ways. Whatever the reviews say, the AI says it back, magnified. 24.27%
12. THE OPPORTUNITY Competitor reviews are a free, always-updating focus group Customer language Service delivery failures The exact phrases customers use for their problems. Raw material for positioning. No-shows, pricing surprises, rushed jobs. A public record of what customers wanted instead. @yourfavoriteseo
13. THE OPPORTUNITY, CONTINUED Two more things competitor reviews reveal Unaddressed trust gaps What “good” means here Anxieties in the reviews that competitors never answer in their own messaging. Praise reveals the standard your client is actually measured against. @yourfavoriteseo
14. WIDER THAN REVIEWS Reviews aren’t the only customer data you can work trust with Anywhere you receive feedback on your brand, products, or services is a great resource to use. Satisfaction surveys Call logs Every other channel Structured, first-party sentiment you already collect. What customers ask and complain about out loud. Social comments, Reddit, support tickets, DMs, sales notes. @yourfavoriteseo
15. THE FRAMEWORK Just three steps Export competitor reviews, analyze sentiment, cluster into themes you can act on. 3 01 02 03 Export Analyze Cluster Pull competitor review text from GBP and beyond. Run sentiment analysis to surface themes and language. Group recurring themes into a topic map you can act on. @yourfavoriteseo
16. Export the reviews Land the data in a clean spreadsheet with review text, date, and source. Any of these methods work. GBP API Pull your own reviews straight from Google via the GBP API. Limitation: your own profile only, so no competitor analysis. @yourfavoriteseo Sentiment Analyzer extension Exports competitor reviews and runs sentiment analysis automatically. Limitation: only ~60ish days of data per GBP. Apify Use Apify's Google Maps scraper to pull competitor reviews. Limitation: caps at 5,000 reviews per place, beyond that the scraper splits them across duplicate-place items.
17. Prompt Claude or ChatGPT Paste your review CSV into a chat and ask for the analysis. Copy this template and adjust. I'm going to paste a CSV of Google reviews for [Competitor Name], a [business type] in [city] . Please identify: 1. The top 5–7 recurring themes (positive and negative) 2. How many reviews mention each theme 3. Any patterns suggesting operational failures 4. 3–5 direct quotes representing each theme 5. The biggest gap between praise and complaints Here is the data: [paste CSV] @yourfavoriteseo
18. Use the Chrome extension The GBP Reviews Sentiment Analyzer runs the analysis automatically while it pulls the reviews. No setup, no code. 1 Install it, open a competitor's Google Business Profile, and run it. 2 Reviews export to XLS, already tagged positive, negative, or neutral. Best for: getting started fast, or anyone who'd rather not touch a script at all. @yourfavoriteseo
19. Pick the right competitors Target the 2–3 competitors your client is actually losing jobs to. Not every business in the local pack. 1 Search the client's top products and services in Google Maps, then note which names keep reappearing. 2 Ask the client. They know who they lose bids to. Check Google's AI review summaries for competitors called out on features your client wants to 3 own. @yourfavoriteseo
20. The Google Cloud Natural Language API More setup than a chat prompt, but far more consistent at scale. Two functions: analyzeSentiment analyzeEntitySentiment THE BIG-PICTURE SNAPSHOT THE DETAIL UNDERNEATH One score per review. Best for a location snapshot, and for comparing locations. Sentiment by entity: product positive, service negative, in the same review. Run both together @yourfavoriteseo
21. CHOOSING THE RIGHT TOOL Why this API earns its place in the workflow Catches the reviews everyone else misses Tells you what’s actually being talked about Ties sentiment to the topic, not just the tone Built for reporting that holds up over time @yourfavoriteseo
22. A QUICK PRIVACY NOTE You're working with reviews that are already public. The usual concerns about feeding proprietary data to an AI tool don't apply here. @yourfavoriteseo !
23. STEP 4 Build your topic cluster map Organize what you found into the credibility themes prominent in your data. Quality Communication Pricing Workmanship, results, expertise Responsiveness, updates, follow-through Transparency, value, billing surprises Speed Trust Sta / Team Arrival times, turnaround, scheduling Reliability, honesty, doing what they said Professionalism, friendliness, knowledge @yourfavoriteseo
24. READING THE DATA Start with velocity and volume 129 reviews · 5.3 / week 28 reviews · 0.9 / week Active, ongoing trust-building. A business people rarely feel compelled to talk about. When scores are close (4.7 vs. 4.8), differentiation has to come from messaging specificity @yourfavoriteseo
25. READING THE DATA Ask 4 questions of every review set 1 What gets praised here that your client also does, but never says? 2 Where is there frustration that your client's operations genuinely solve? 3 What language do reviewers use that the website doesn't? 4 What's the underlying fear or desire behind the complaint? The most important one. @yourfavoriteseo
26. READING THE DATA Every complaint is a fear in disguise Negative reviews map customer anxieties. Each one is a conversion lever. WHAT REVIEWERS SAY THE ANXIETY UNDERNEATH “They overcharged me.” Fear of being taken advantage of “They said they'd come between 9 and 11. They showed up at 3.” Fear of wasted time “I had to chase them for updates.” Fear of being ignored after signing @yourfavoriteseo
27. TURNING INSIGHT INTO COPY Turn the gap into a positioning statement COMPETITOR COMPLAINT PATTERN DIRECT POSITIONING RESPONSE “They never gave me a price up front.” “Upfront pricing on every job. No surprises on your invoice.” “They said 9. They came at 3.” “Exact arrival windows, not four-hour guessing games.” “The work looked rushed, and they just left.” “We don't leave until the job is done and you're satisfied.” @yourfavoriteseo
28. GOING DEEPER Searches carry emotional signals, if you know where to look Cluster queries by how they're phrased, not just what they ask. Emotional Fear: “what if,” “avoid,” “mistakes” Hope: “best,” “transform” Trust: “actually work,” “proven” Curiosity: “how,” “why” @yourfavoriteseo
29. TURNING INSIGHT INTO CLIENT WORK Five ways to deliver this to a client 01 02 03 04 05 USP/UVP Extraction Gap Messaging FAQ Content GBP Updates Content Series Pull praise language straight into headers & hero copy. Answer each competitor complaint pattern directly. Rewrite descriptions and posts using review language. Turn recurring confusion into new pages or videos. @yourfavoriteseo Neutralize anxieties before a customer has to ask.
30. EXAMPLE A local plumbing business THE ANXIETY WHAT CHANGED Will I get a surprise bill, and will they even show up when they said? Website copy rebuilt around explicit pricing transparency and guaranteed arrival windows. @yourfavoriteseo
31. EXAMPLE A residential home builder THE ANXIETY THE SURPRISE Chronic schedule delays and feeling left in the dark during construction. Reviews kept praising a real-time status app the builder never marketed. It went front and center. @yourfavoriteseo
32. PROVING IT WORKED Measuring what changes Track behavior and visibility before and after the copy swap. On-Page Behavior @yourfavoriteseo GBP Performance Organic Visibility AI Citations
33. Wait. AI citations? @yourfavoriteseo
34. Yes. But it’s not just the sentiment of the response. It’s the sentiment of your input. @yourfavoriteseo
35. SE Ranking study: 900 AI responses 180 prompts across 3 product categories and 4 intent families. 3 CATEGORIES 4 INTENT FAMILIES Samsung vs Sony TVs Recommendation Hers vs Start Willow (GLP-1 telehealth) Comparison Delta SkyMiles vs United MileagePlus Trust Value @yourfavoriteseo
36. Example prompts by intent family Recommendation Positive Comparison Neutral “Which budget-friendly Samsung TVs offer the most impressive features for the price?” “What are the key differences in picture performance between Samsung QLED and Sony OLED TVs?” Trust Value Negative “How severe are the burn-in and panel degradation risks on Sony OLED TVs over long-term use?” @yourfavoriteseo Neutral “How does the price-to-feature ratio of Samsung’s mid-range lineup compare to market standards?”
37. Prompts came in matched triplets Same question, same brand. Only the tone changes. Negative Neutral Positive 60 60 60 negative prompts neutral prompts positive prompts 60 triplets in total, 20 per category. @yourfavoriteseo
38. Prompt tone shapes how ChatGPT rates the brand ρ = 0.66 2× +20 Strong relationship Negative pulls harder Neutral baseline Between prompt tone and brand rating. Spearman, n = 180. A negative prompt moves the response twice as far as a positive one. Neutral prompts already rate the brand mildly positive. Consistent across every category, every brand, and 3 of 4 intent families. @yourfavoriteseo
39. Intent changes how much tone maers Recommendation & Value Most responsive to tone, in both directions. @yourfavoriteseo Comparison Trust Least responsive to negative tone. The answer tends to balance both brands. Negative tone lowers the answer. Positive tone does nothing. Reliability, warranty, and data safety questions always get a cautious answer.
40. Framing decides who gets criticized. It doesn’t fully override what the model already believes. @yourfavoriteseo
41. Other factors to consider Brand’s general sentiment on the web Brand’s prominence in the category Nature of the question Response-to-response variation HOWEVER Prompt tone shapes the sentiment of each response, within limits set by the brand’s existing standing in the model’s sources. @yourfavoriteseo
42. Three things to take with you 01 02 03 Search is a trust decision Reviews are your research Sentiment in, sentiment out Every query asks who to believe. Trust is the core of E-E-A-T. Competitor reviews reveal the language, gaps, and fears worth answering. AI echoes the reputation it finds and the tone it’s asked in. @yourfavoriteseo
43. The opportunity most SEOs leave on the table Listen to your customers. They are telling you what they want and how to win @yourfavoriteseo · Celeste Gonzalez
44. Thank you Questions? Find me at @yourfavoriteseo Celeste Gonzalez · Lastmile Retail