Presented: February 16, 2026 Claude Code London 12 slides · 376 KB PDF Keynote

About this presentation

In 2026, you're either the source AI cites — or you are invisible. This keynote, delivered at Claude Code London, defines the inflection point where AI interfaces became the primary discovery layer.

The presentation introduces three foundational concepts for modern AI visibility:

  • The Citation Probability Model: P(citation) ≈ Authority + Relevance + Information Gain
  • Information Gain (G): Why generic content scores zero and proprietary data wins citations
  • Technical AIO: Treating your website as an API with JSON-LD, llms.txt, and chain-of-thought structure

Includes proprietary data on the shift from keywords to vectors, the zero-click reality, and actionable frameworks for AEO and GEO implementation.

Key frameworks

Citation Probability Model

P(citation) ≈ Authority + Relevance + Information Gain

Each factor weighted differently by AI models (ChatGPT, Gemini, Claude, Perplexity). Introduces the first predictive model for AI citation.

Information Gain (G)

G = 0 (Invisible) vs. G = High (Cited)

Generic content = zero. Proprietary data, original case studies, and unique expert analysis drive citation velocity.

Technical AIO Stack

JSON-LD + llms.txt + Chain-of-Thought

Treat your site as an API for AI crawlers. Structured data, clean markdown, and reasoning-based content architecture.

From Keywords to Vectors

2010-2023: Keyword Era → 2026: Vector Era

AI calculates semantic proximity, not keyword matches. Content closest to intent centroid wins.

Preview insights

INFLECTION POINT

~59% of consumers believe AI will become their primary method for finding information

Eight Oh Two, February 2026

The zero-click reality: 25% predicted reduction in traditional search volume (Gartner 2026). Users trust synthesized answers over link lists.

+40%

more likely to be cited when content provides direct, verifiable answers

G=0 (INVISIBLE)
  • Generic "What is X?" content
  • Rewritten Wikipedia summaries
  • Same stats everyone quotes
  • No original perspective
G=HIGH (CITED)
  • Proprietary data & benchmarks
  • Original case studies
  • Primary research & surveys
  • Unique expert analysis
YH

Yuliya Halavachova

Founder & Principal Data Scientist, UltraScout AI

Principal Data Scientist with 16+ years in AI architecture, NLP, and production AI systems. Cited in 10TH GLOBAL TECHMINING CONFERENCE (GTM2020). Yuliya leads the strategy and remains hands-on with code, models, and deployment.

Get the full presentation

12 slides covering the Citation Probability Model, Information Gain framework, and technical AIO stack.

Download PDF (376 KB)

Includes data visualizations, proprietary frameworks, and implementation checklists.