NEW! Reimagining the eCommerce Search Experience with SearchAI Get Started.

NEW! Reimagining the eCommerce Search Experience with SearchAI Get Started.

NEW! Reimagining the eCommerce Search Experience with SearchAI Get Started.

understanding the four stages of enterprise search white paper video background image

SearchAI 11.2 — Faster to Deploy, Smarter to Use

SearchAI 11.2 — Faster to Deploy, Smarter to Use

SearchAI 11.2 — Faster to Deploy, Smarter to Use

September 19, 2025

SearchAI 11.2 ships upgrades that make answers cleaner, oversight simpler, and rollout faster.

SearchAI 11.2 ships upgrades that make answers cleaner, oversight simpler, and rollout faster.

SearchAI 11.2 ships upgrades that make answers cleaner, oversight simpler, and rollout faster.

If you’re already using SearchBlox SearchAI, this release enhances control without slowing down your teams. If you’re new, it’s the quickest path to grounded, enterprise-ready AI search.

If you’re already using SearchBlox SearchAI, this release enhances control without slowing down your teams. If you’re new, it’s the quickest path to grounded, enterprise-ready AI search.

If you’re already using SearchBlox SearchAI, this release enhances control without slowing down your teams. If you’re new, it’s the quickest path to grounded, enterprise-ready AI search.

With SearchBlox SearchAI 11.2, AI governance just got easier to control. Now you can approve AI-generated metadata before it goes live, keep conversations on-brand and within company policy with a single master prompt - controlling how answers are generated across SearchAI experiences. 

With SearchBlox SearchAI 11.2, AI governance just got easier to control. Now you can approve AI-generated metadata before it goes live, keep conversations on-brand and within company policy with a single master prompt - controlling how answers are generated across SearchAI experiences. 

With SearchBlox SearchAI 11.2, AI governance just got easier to control. Now you can approve AI-generated metadata before it goes live, keep conversations on-brand and within company policy with a single master prompt - controlling how answers are generated across SearchAI experiences. 

Product discovery gets a major enhancement. Now embedded images inside PDFs and decks instantly searchable across more data sources. Instantly  generate Structured FAQs (JSON-LD) that cites your content on external AI tools (ChatGPT, Gemini, Perplexity, Bing).

Product discovery gets a major enhancement. Now embedded images inside PDFs and decks instantly searchable across more data sources. Instantly  generate Structured FAQs (JSON-LD) that cites your content on external AI tools (ChatGPT, Gemini, Perplexity, Bing).

Product discovery gets a major enhancement. Now embedded images inside PDFs and decks instantly searchable across more data sources. Instantly  generate Structured FAQs (JSON-LD) that cites your content on external AI tools (ChatGPT, Gemini, Perplexity, Bing).

With enhanced governance, control and security, shoppers, employees and customers get consistent answers across all SearchAI products, with the added visibility into what surfaced and why.

With enhanced governance, control and security, shoppers, employees and customers get consistent answers across all SearchAI products, with the added visibility into what surfaced and why.

With enhanced governance, control and security, shoppers, employees and customers get consistent answers across all SearchAI products, with the added visibility into what surfaced and why.

Key Enhancements - Highlights

Key Enhancements - Highlights

Key Enhancements - Highlights

What’s new - at a Glance

What’s new - at a Glance

What’s new - at a Glance

Human-in-the-Loop Metadata

AI proposes; reviewers refine before it goes live. Clean facets, better recall/precision, audit-friendly history.



Human-in-the-Loop Metadata

AI proposes; reviewers refine before it goes live. Clean facets, better recall/precision, audit-friendly history.



Human-in-the-Loop Metadata

AI proposes; reviewers refine before it goes live. Clean facets, better recall/precision, audit-friendly history.



Image-Inclusive Search

Find documents by the images inside—contracts, designs, product catalogs, scanned PDFs—across SharePoint, Google Drive, AEM, S3, Email, and more.


Image-Inclusive Search

Find documents by the images inside—contracts, designs, product catalogs, scanned PDFs—across SharePoint, Google Drive, AEM, S3, Email, and more.


Image-Inclusive Search

Find documents by the images inside—contracts, designs, product catalogs, scanned PDFs—across SharePoint, Google Drive, AEM, S3, Email, and more.


Redesigned Analytics

Redesigned Analytics

Redesigned Analytics

Cleaner views across search, SmartFAQs, chatbot, and recommendations. See what users click, where they stall, and what to fix next. 


Cleaner views across search, SmartFAQs, chatbot, and recommendations. See what users click, where they stall, and what to fix next. 


Cleaner views across search, SmartFAQs, chatbot, and recommendations. See what users click, where they stall, and what to fix next. 


Enhanced SmartFAQs (JSON-LD)

Enhanced SmartFAQs (JSON-LD)

Enhanced SmartFAQs (JSON-LD)

Machine-readable FAQs give LLMs a trusted source, so answers point back to your site—not someone else’s. Great for consistent answers in search, chat, and external AI.

Machine-readable FAQs give LLMs a trusted source, so answers point back to your site—not someone else’s. Great for consistent answers in search, chat, and external AI.

Machine-readable FAQs give LLMs a trusted source, so answers point back to your site—not someone else’s. Great for consistent answers in search, chat, and external AI.

Prompt Management

Prompt Management

Prompt Management

A customizable master prompt that governs how the LLM responds across SearchAI experiences—ChatBot, PreText NLP, and Assist—with no code.



A customizable master prompt that governs how the LLM responds across SearchAI experiences—ChatBot, PreText NLP, and Assist—with no code.



A customizable master prompt that governs how the LLM responds across SearchAI experiences—ChatBot, PreText NLP, and Assist—with no code.



Flexible LLMs

Flexible LLMs

Flexible LLMs

Choose the right model per job: OpenAI, OpenAI-compatible like vLLM, or local Ollama/LlamaCPP for sensitive workloads. Swap without re-architecting


Choose the right model per job: OpenAI, OpenAI-compatible like vLLM, or local Ollama/LlamaCPP for sensitive workloads. Swap without re-architecting


Choose the right model per job: OpenAI, OpenAI-compatible like vLLM, or local Ollama/LlamaCPP for sensitive workloads. Swap without re-architecting


Prompt Management

Prompt Management

Prompt Management

Set prompt rules to get consistent answers everywhere

Set prompt rules to get consistent answers everywhere

Set prompt rules to get consistent answers everywhere

With 11.2, you can define a single master prompt-—in a clear, conversational tone—to govern how SearchAI responds across ChatBot, PreText NLP, and SearchAI Assist. You set the persona, do/don’t rules, citation expectations, and safety guardrails - ensuring every answers stay on-brand, consistent, and compliant without extra work

With 11.2, you can define a single master prompt-—in a clear, conversational tone—to govern how SearchAI responds across ChatBot, PreText NLP, and SearchAI Assist. You set the persona, do/don’t rules, citation expectations, and safety guardrails - ensuring every answers stay on-brand, consistent, and compliant without extra work

With 11.2, you can define a single master prompt-—in a clear, conversational tone—to govern how SearchAI responds across ChatBot, PreText NLP, and SearchAI Assist. You set the persona, do/don’t rules, citation expectations, and safety guardrails - ensuring every answers stay on-brand, consistent, and compliant without extra work

How it works

How it works

How it works

ChatBot (single prompt): You can set a master prompt to applu tone, style, and rules - so that every conversation follows the same voice, scope, and guardrails. 

PreText NLP (content prompts): Add lightweight prompts for Title, Meta Description, Image Description, and FAQs to keep generated content consistent with your brand.

SearchAI Assist (action prompts): Control SearchAI Assist Actions with their own prompts—e.g., Compare, Summarize, Highlight, Extract. Each action is defined as Action Name + Prompt so Assist delivers the exact table, bullets, or callouts you expect, every time.

ChatBot (single prompt): You can set a master prompt to applu tone, style, and rules - so that every conversation follows the same voice, scope, and guardrails. 

PreText NLP (content prompts): Add lightweight prompts for Title, Meta Description, Image Description, and FAQs to keep generated content consistent with your brand.

SearchAI Assist (action prompts): Control SearchAI Assist Actions with their own prompts—e.g., Compare, Summarize, Highlight, Extract. Each action is defined as Action Name + Prompt so Assist delivers the exact table, bullets, or callouts you expect, every time.

ChatBot (single prompt): You can set a master prompt to applu tone, style, and rules - so that every conversation follows the same voice, scope, and guardrails. 

PreText NLP (content prompts): Add lightweight prompts for Title, Meta Description, Image Description, and FAQs to keep generated content consistent with your brand.

SearchAI Assist (action prompts): Control SearchAI Assist Actions with their own prompts—e.g., Compare, Summarize, Highlight, Extract. Each action is defined as Action Name + Prompt so Assist delivers the exact table, bullets, or callouts you expect, every time.

AI Mode

AI Mode

AI Mode

Keep discovering with conversational search and intelligent follow-ups.

Keep discovering with conversational search and intelligent follow-ups.

Keep discovering with conversational search and intelligent follow-ups.

A conversational search experience inside the search bar that uses AI-generated context to ask clarifying follow-up questions, disambiguate intent, and guide users to a single best next step (the exact answer, page, policy, or form) instead of a long ranked list.

A conversational search experience inside the search bar that uses AI-generated context to ask clarifying follow-up questions, disambiguate intent, and guide users to a single best next step (the exact answer, page, policy, or form) instead of a long ranked list.

A conversational search experience inside the search bar that uses AI-generated context to ask clarifying follow-up questions, disambiguate intent, and guide users to a single best next step (the exact answer, page, policy, or form) instead of a long ranked list.

How it works

How it works

How it works

Asks the right follow-ups: After the initial question, AI Mode prompts for the missing bits (role, region, product, date) inside the search bar to narrow the task.
Refines to confidence: Each follow-up trims ambiguity; answers are compiled from your content and access controls until there’s a clear, single recommendation.
Takes action and learns: Delivers the exact page/policy/form and preserves the conversation thread; Analytics → Search/Chatbot logs the follow-up path to cut future reformulations.

Asks the right follow-ups: After the initial question, AI Mode prompts for the missing bits (role, region, product, date) inside the search bar to narrow the task.
Refines to confidence: Each follow-up trims ambiguity; answers are compiled from your content and access controls until there’s a clear, single recommendation.
Takes action and learns: Delivers the exact page/policy/form and preserves the conversation thread; Analytics → Search/Chatbot logs the follow-up path to cut future reformulations.

Asks the right follow-ups: After the initial question, AI Mode prompts for the missing bits (role, region, product, date) inside the search bar to narrow the task.
Refines to confidence: Each follow-up trims ambiguity; answers are compiled from your content and access controls until there’s a clear, single recommendation.
Takes action and learns: Delivers the exact page/policy/form and preserves the conversation thread; Analytics → Search/Chatbot logs the follow-up path to cut future reformulations.

Human-in-the-Loop Metadata Validation

Human-in-the-Loop Metadata Validation

Human-in-the-Loop Metadata Validation

Make sure that human-vetted and trusted metadata reaches search

Make sure that human-vetted and trusted metadata reaches search

Make sure that human-vetted and trusted metadata reaches search

A review workflow that gates AI-generated metadata—including titles, descriptions, keywords/tags, and other fields—before it’s saved and indexed. Admins see the AI’s proposed values, edit/approve or reject, and then publish with version history and auditability, ensuring accuracy, compliance, and consistent taxonomy across collections.

A review workflow that gates AI-generated metadata—including titles, descriptions, keywords/tags, and other fields—before it’s saved and indexed. Admins see the AI’s proposed values, edit/approve or reject, and then publish with version history and auditability, ensuring accuracy, compliance, and consistent taxonomy across collections.

A review workflow that gates AI-generated metadata—including titles, descriptions, keywords/tags, and other fields—before it’s saved and indexed. Admins see the AI’s proposed values, edit/approve or reject, and then publish with version history and auditability, ensuring accuracy, compliance, and consistent taxonomy across collections.

How it works

How it works

How it works

AI drafts, you control: Generates titles/descriptions/tags, limited by file-type–specific metadata rules.

Review queue: Editors edit/approve/reject proposed fields (single or batch).
Publish with proof: Only approved metadata is indexed, with version history/audit trail.

AI drafts, you control: Generates titles/descriptions/tags, limited by file-type–specific metadata rules.

Review queue: Editors edit/approve/reject proposed fields (single or batch).
Publish with proof: Only approved metadata is indexed, with version history/audit trail.

AI drafts, you control: Generates titles/descriptions/tags, limited by file-type–specific metadata rules.

Review queue: Editors edit/approve/reject proposed fields (single or batch).
Publish with proof: Only approved metadata is indexed, with version history/audit trail.

Image-Inclusive Search

Image-Inclusive Search

Image-Inclusive Search

Find charts, scans, and specs inside your files and documents

Find charts, scans, and specs inside your files and documents

Find charts, scans, and specs inside your files and documents

Multimodal document search that extracts and indexes visual elements inside files—embedded images, charts, diagrams, and scanned pages—alongside text, now across more sources: Dynamic Auto, File, SharePoint (Online & On-Prem), Google Drive, AEM, Amazon S3, Email, and Product Discovery. Users can find the right item by the images inside a document, not just filenames or surrounding text.

Multimodal document search that extracts and indexes visual elements inside files—embedded images, charts, diagrams, and scanned pages—alongside text, now across more sources: Dynamic Auto, File, SharePoint (Online & On-Prem), Google Drive, AEM, Amazon S3, Email, and Product Discovery. Users can find the right item by the images inside a document, not just filenames or surrounding text.

Multimodal document search that extracts and indexes visual elements inside files—embedded images, charts, diagrams, and scanned pages—alongside text, now across more sources: Dynamic Auto, File, SharePoint (Online & On-Prem), Google Drive, AEM, Amazon S3, Email, and Product Discovery. Users can find the right item by the images inside a document, not just filenames or surrounding text.

How it works

How it works

How it works

Extracts visuals: Indexes images/charts/scans inside PDFs, decks, and scanned files alongside text.
Broader coverage: Works across SharePoint (Online & On-Prem), Google Drive, AEM, S3, Email, Product Discovery, and more.
Unified retrieval: Queries match text and visuals so image-only content becomes findable.

Extracts visuals: Indexes images/charts/scans inside PDFs, decks, and scanned files alongside text.
Broader coverage: Works across SharePoint (Online & On-Prem), Google Drive, AEM, S3, Email, Product Discovery, and more.
Unified retrieval: Queries match text and visuals so image-only content becomes findable.

Extracts visuals: Indexes images/charts/scans inside PDFs, decks, and scanned files alongside text.
Broader coverage: Works across SharePoint (Online & On-Prem), Google Drive, AEM, S3, Email, Product Discovery, and more.
Unified retrieval: Queries match text and visuals so image-only content becomes findable.

Enhanced SmartFAQs (JSON-LD)

Enhanced SmartFAQs (JSON-LD)

Enhanced SmartFAQs (JSON-LD)

Make external AI's cite your site—not someone else’s

Make external AI's cite your site—not someone else’s

Make external AI's cite your site—not someone else’s

An FAQ publishing format that outputs your Q&A as Google-friendly JSON-LD structured data, creating a machine-readable, canonical source that external LLMs like ChatGPT, Gemini, Perplexity, and Copilot can reference a trusted source—your site. FAQs stay synchronized with your content, so answers remain consistent and attributable to your site.

An FAQ publishing format that outputs your Q&A as Google-friendly JSON-LD structured data, creating a machine-readable, canonical source that external LLMs like ChatGPT, Gemini, Perplexity, and Copilot can reference a trusted source—your site. FAQs stay synchronized with your content, so answers remain consistent and attributable to your site.

An FAQ publishing format that outputs your Q&A as Google-friendly JSON-LD structured data, creating a machine-readable, canonical source that external LLMs like ChatGPT, Gemini, Perplexity, and Copilot can reference a trusted source—your site. FAQs stay synchronized with your content, so answers remain consistent and attributable to your site.

How it works

How it works

How it works

Canonical, machine-readable source: Publish FAQs as Google-friendly JSON-LD that summarises key pages in clear Q&A, giving ChatGPT, Gemini, Perplexity, and Copilot a reliable dataset so answers cite your site.
Fast creation with oversight: Pull Q&As from any URL or content collection in one click, then approve/tweak/delete each item before it goes live (human-in-the-loop).
Always current, built for impact: Updates stay in sync without heavy dev work, turning AI answers into qualified traffic and reinforcing brand authority and safety.

Canonical, machine-readable source: Publish FAQs as Google-friendly JSON-LD that summarises key pages in clear Q&A, giving ChatGPT, Gemini, Perplexity, and Copilot a reliable dataset so answers cite your site.
Fast creation with oversight: Pull Q&As from any URL or content collection in one click, then approve/tweak/delete each item before it goes live (human-in-the-loop).
Always current, built for impact: Updates stay in sync without heavy dev work, turning AI answers into qualified traffic and reinforcing brand authority and safety.

Canonical, machine-readable source: Publish FAQs as Google-friendly JSON-LD that summarises key pages in clear Q&A, giving ChatGPT, Gemini, Perplexity, and Copilot a reliable dataset so answers cite your site.
Fast creation with oversight: Pull Q&As from any URL or content collection in one click, then approve/tweak/delete each item before it goes live (human-in-the-loop).
Always current, built for impact: Updates stay in sync without heavy dev work, turning AI answers into qualified traffic and reinforcing brand authority and safety.

The Future of Work Today

The Future of Work Today

The Future of Work Today

We constantly listen to the frictions that teams feel every day. From too many steps, unclear results, or fixes that take too long to land. SearcBlox SearchAI 11.2 tackles these head-on with practical changes that shorten paths, clarify sources, and speed up improvements.

We constantly listen to the frictions that teams feel every day. From too many steps, unclear results, or fixes that take too long to land. SearcBlox SearchAI 11.2 tackles these head-on with practical changes that shorten paths, clarify sources, and speed up improvements.

We constantly listen to the frictions that teams feel every day. From too many steps, unclear results, or fixes that take too long to land. SearcBlox SearchAI 11.2 tackles these head-on with practical changes that shorten paths, clarify sources, and speed up improvements.

“The win isn’t another control; it’s finishing the task on the first try. 11.2 focuses on that last mile—cutting dead ends, showing where answers come from, and giving admins a faster way to tune behavior when needs change."

“The win isn’t another control; it’s finishing the task on the first try. 11.2 focuses on that last mile—cutting dead ends, showing where answers come from, and giving admins a faster way to tune behavior when needs change."

“The win isn’t another control; it’s finishing the task on the first try. 11.2 focuses on that last mile—cutting dead ends, showing where answers come from, and giving admins a faster way to tune behavior when needs change."

Timo Selvaraj, Chief Product Officer, SearchBlox