WIP - Restructuring vector search docs and adding vector landing page and conceptual page

[kosabogi]

👀 💬 Draft PR - seeking your feedback. This is a proposal for collaboration. The changes are open to discussion and revision.

Vector search documentation restructure

Summary of changes

What was done

• Added a vector search landing page (solutions/search/vector.md): A dedicated entry point for vector search (use cases, workflows, collection of resources)

• Added a concepts page (solutions/search/vector/how-vector-search-works.md): A terminology and concepts page covering important terminology (like vector embeddings, embedding models, dimensions), quantization, querying.

• Restructured the information architecture:

  • Removed the "AI-powered search" naming layer.
  • Vector search is now the top-level concept under Search approaches; semantic search is presented as a type of vector search workflow.

• Updated cross-links across search docs to reflect the new structure (e.g., search-approaches.md, semantic-search.md, hybrid-search.md, get-started/semantic-search.md).

---

Main decisions and reasoning

I did research to inform these decisions. Here you can check the research and the key takeaways: search term analysis and semantic/vector relationship.

• Remove the AI-powered search umbrella category. "AI-powered search" is nowadays more associated with LLMs, chat interfaces, and generative AI features. "Vector search" and "semantic search" better describe the underlying technical capability and align with user expectations.

• Treat semantic search as part of vector search. Logically clearer: vector search is the foundation; semantic search is a workflow built on it.

---

Reviewer focus: first round

For this first review, please focus on structure and naming, not detailed content. We can do deeper content review in a follow-up round once the structure is agreed.

What to review

1. Overall structure

   • What do you think about the the proposed hierarchy?
   • Missing or misplaced topics in the TOC or navigation?

2. Naming changes

   • Is removing the "AI-powered search" layer appropriate?

3. Vector search landing page (vector.md)

   • Topics covered: Are the sections and topics right? Anything else or different?
   • Grouping resources: Does this grouping (Start here, Managed workflows, Manual implementation, etc.) seem useful? Any alternative groupings to suggest?

4. How vector search works (vector/how-vector-search-works.md)

   • Terminology: Are the terms covered (vector, embedding, dimensions, dense vs. sparse, quantization, query) the right set for this page?
   • Gaps: What other terms or concepts should be included or moved elsewhere?

Out of scope for this round

• Detailed line-by-line content review
• Technical accuracy deep-dive
• Copy-editing and style refinements

Once structure and naming are confirmed, we can proceed with a content-focused review.

[github-actions]

Vale Linting Results

Summary: 3 warnings, 1 suggestion found

⚠️ Warnings (3)

File	Line	Rule	Message
solutions/search/vector.md	72	Elastic.Latinisms	Latin terms and abbreviations are a common source of confusion. Use 'versus' instead of 'vs'.
solutions/search/vector.md	109	Elastic.Spelling	'Cohere's' is a possible misspelling.
solutions/search/vector/how-vector-search-works.md	47	Elastic.Latinisms	Latin terms and abbreviations are a common source of confusion. Use 'versus' instead of 'vs'.

💡 Suggestions (1)

File	Line	Rule	Message
solutions/search/vector.md	21	Elastic.WordChoice	Consider using 'efficient' instead of 'easy', unless the term is in the UI.

---

The Vale linter checks documentation changes against the Elastic Docs style guide.

To use Vale locally or report issues, refer to Elastic style guide for Vale.

[github-actions]

🔍 Preview links for changed docs

• solutions/search/vector/how-vector-search-works.md
• explore-analyze/ai-features.md
• solutions/search/get-started/semantic-search.md
• solutions/search/hybrid-search.md
• solutions/search/search-approaches.md
• solutions/search/semantic-search.md
• solutions/search/vector.md
• solutions/search/vector/bring-own-vectors.md