Search Engine Journal covered this week a Victorious study that analyzed 107,011 AI responses across eight platforms (ChatGPT, Perplexity, Gemini, Google AI Overview, AI Mode, Microsoft Copilot, Claude, Meta AI) on 177 brands across five verticals. The headline number is brutal: 89.8% of brands had zero mentions. It's good news and bad news for your blog depending on how you read it.
What the data means for your blog
The usual read is "the shift to AI search hasn't taken off yet". The useful read is different: AI search already dominates millions of queries a day, but most brands haven't done anything to appear in them. If your blog tweaks four specific things now, you arrive before 90%.
The 4 fixes that move the needle
1. Structure your blog like a database, not a timeline
AI parses sources organized by topic and answer, not by chronological order. Each article should have a clear answer to a concrete question in the first 200-300 words, before details. That dramatically increases the probability of being cited.
Concrete practice: add a summary block at the beginning ("Summary: X works because Y, Z, W") that the AI can extract literally.
2. Visible authorship with real credentials
When an AI cites a source, it looks for authority signals: byline with photo, About page with verifiable experience, links to LinkedIn or other public profiles. A blog without visible author or with generic author ("Editorial Team") is rarely cited. A blog with identifiable author and short bio is.
3. Own data, not summaries of others' data
AI can summarize generic content better than you. What it CAN'T do is invent an own data point. When your blog produces its own research (survey of your audience, internal metrics you share, documented experiment), you become a primary source. Primary sources get cited; secondary ones get skipped.
4. Article + Person schema markup for the author
JSON-LD with Article type that includes a Person as author, with sameAs pointing to verifiable profiles. It's the technical signal that most moves the AIs today. If your platform doesn't generate it by default, add it manually.
| Fix | Cost | Impact on AI citations |
|---|---|---|
| Summary at article start | 10 min/article | High |
| About page and author with credentials | 1 hour (one-time) | High |
| Produce 1 own data point every 5 posts | Variable, depends on data | Very high |
| Article + Person schema | 30 min setup | Medium-high |
What the study revealed about verticals
AI mentions don't distribute evenly across sectors. Healthcare, SaaS, and Financial Services showed consistent patterns between mentions and citations; other verticals have gaps:
- Sectors with dominant incumbents: one or two brands capture 80% of mentions. Breaking in there requires specific angle, not general content.
- Sectors with high dispersion: many mentions spread out, clear opportunity to build constant presence.
- New sectors: AI doesn't yet have settled sources. Being first to publish authoritative content in an emerging sub-topic is the cheapest 2026 advantage.
The study revealed AIs show different preferences by platform. What ChatGPT cites isn't what Perplexity or Google AI Mode cite. Source diversity (publishing the same article on your blog + LinkedIn article + dev.to) increases cross-coverage probability.
Common mistakes that reduce your AI mentions
- Generic headlines. "10 marketing tips" doesn't get cited; "Why we lowered cost per lead 30% in Q1" does.
- Weak attribution. "Studies show..." without link or author loses credibility for the AI.
- Repetition of secondary sources. If all data comes from HubSpot blog, AI cites HubSpot, not you.
- Lack of temporal review. An article with 2022 data is still valid in stable categories, but AI prefers recent sources for anything that can change.
How to measure if the fixes work
Measuring AI citations is still effortful. Three practical paths:
- Monthly manual queries: prepare 10 representative queries from your niche, run them on ChatGPT, Perplexity, Gemini, and Google AI Mode once a month. Note appearances.
- Dedicated tools (Otterly, BrandRank, Am I Cited): monitor automatically. Paid, but saves time.
- Search Console + referrer analysis: although AI doesn't send clear referrer, you do see direct traffic spikes correlated with recent chat mentions. Not proof but signal.
"With only 18 of the 177 brands we measured earning mentions in AI search, there's still white space in your vertical waiting to be claimed" (summary of the Victorious study via Search Engine Journal).
Conclusion
Appearing in AI responses in 2026 isn't magic: it's discipline applied to four concrete things (clear structure, visible authorship, own data, correct schema). 90% of the competition isn't doing it, so the opportunity window is real and cheap. The investment isn't money, it's thirty extra minutes per article.
If your blog platform already includes schema markup, visible byline, and multimedia support to insert own data without fighting plugins, try Vlogerly: technical plumbing is done, your work is producing the content the AI wants to cite.
