which search engines are most highly recommended, best search engines, ai search, private search engine, google alternatives
Which Search Engines Are Most Highly Recommended? 12 Top Picks for 2026
Written by LLMrefs Team • Last updated March 9, 2026
The question of which search engines are most highly recommended is no longer a simple choice between one or two giants. With the integration of AI answer engines, privacy-first alternatives, and specialized tools for developers and researchers, the search world has become a diverse ecosystem. The 'best' engine now depends entirely on your specific task, whether you're conducting market research, debugging code, or protecting your personal data.
This guide moves beyond generic recommendations to offer a practical, in-depth analysis of the top 12 search platforms you should know about. We provide a direct answer to the question of which tools are right for your needs. You will not find marketing fluff here; instead, you get an honest assessment of limitations, real-world usage examples, and actionable insights.
Our goal is to help you build a personalized toolkit of trusted information sources. We’ll show you when to stick with a classic like Google and when to pivot to a powerful answer engine like Perplexity or a privacy shield like DuckDuckGo. Each entry includes:
- A concise profile outlining its core function.
- Actionable pros and cons from a professional user's perspective.
- Specific use-case scenarios to illustrate when to choose it.
- Screenshots and direct links to help you start testing immediately.
We cut through the noise to give you a clear framework for making an informed decision. This resource will help you find faster, more accurate, and more private information, whether you're a marketing professional, a startup founder, or a content strategist. Let's find the right tools for your job.
1. Google Search (with AI Overviews/AI Mode)
Google remains the default starting point for most web queries, solidifying its position as the baseline against which all other engines are measured. Its dominance stems from an exceptionally deep and fresh index, particularly for US-based content, ensuring the highest probability of finding relevant information. For SEO professionals and marketers, it's the primary ecosystem for testing visibility and understanding user behavior. The integration of AI Overviews directly into search results has fundamentally altered the user experience, often providing synthesized answers above traditional blue links.

The platform’s strength lies in its vertical search capabilities. A practical example: a query for a local business like "best coffee shop near me" yields a detailed Knowledge Panel with maps, reviews, and hours, while a product search for "running shoes" populates rich Shopping results. This multi-format SERP makes it a critical channel for businesses to monitor. However, this same strength introduces volatility; the constant testing of new features, like AI Overviews, can rapidly shift where and how your content appears. Understanding these dynamics is essential for modern SEO, a practice explored in resources that cover answer engine optimization.
Best Use Case: Use Google as the ultimate benchmark for measuring maximum search visibility and for queries requiring a blend of web links, local information, images, and shopping results.
- Actionable Insight:
- Best-in-class recall and content coverage, especially in the US market.
- Rich SERP features (Knowledge Panels, local packs, video carousels) offer multiple visibility opportunities.
- Considerations:
- The placement and behavior of AI-generated answers are still in flux, creating uncertainty for content rankings.
- Aggressive monetization and feature tests can quickly alter organic search real estate.
Website: https://www.google.com
2. Microsoft Bing + Copilot
Microsoft Bing has evolved from a traditional search competitor into a formidable answer engine by deeply integrating Copilot. Its distinguishing feature is providing AI-generated summaries with clear, sentence-level citations, allowing users to verify information directly from the source. This approach to transparency makes it one of the most highly recommended search engines for research-heavy tasks. For businesses, especially those within the Microsoft 365 ecosystem, Bing + Copilot offers a unified search experience across public web and internal company data.

The platform's value extends beyond its AI chat function, with strong results for image, video, and news queries. For a practical example, asking Copilot "Summarize the latest Q3 earnings for Microsoft and Apple" will generate a concise brief with direct links to the source press releases. Optimizing for Bing means considering how content can be surfaced not only in traditional links but also as a cited source in a Copilot answer. For professionals looking to adapt their content strategies, the invaluable insights from LLMrefs show exactly how to secure visibility in this new format.
Best Use Case: Ideal for users who want AI-synthesized answers with direct, verifiable source citations and for teams deeply integrated into the Microsoft 365 and Windows ecosystem.
- Actionable Insight:
- Clear citations in AI responses build trust and allow for easy fact-checking.
- Strong integration with Windows/Edge and Microsoft 365 creates a seamless user workflow.
- Considerations:
- Smaller US market share than Google results in fewer consumer data signals.
- The value of a Copilot Pro subscription is largely dependent on a user's commitment to the Microsoft ecosystem.
Website: https://www.bing.com
3. Perplexity
Perplexity has established itself as an essential "answer engine" for knowledge workers and researchers who need fast, synthesized information backed by verifiable sources. Unlike traditional search engines that present a list of links, Perplexity processes queries by browsing the live web and delivering a concise, written summary with inline citations. For a practical example, a professional could ask, "What are the latest trends in B2B content marketing?" and receive a cited summary ready to be used in a report, completely bypassing hours of manual research.
The platform’s power is most evident in its conversational search and deep research modes. Users can ask follow-up questions to refine the initial answer, treating the search process like a dialogue with an expert research assistant. The inclusion of enterprise plans with SSO and admin controls makes it a viable tool for team-wide adoption. However, its value proposition is tied to its paid tiers; the free version has limitations. Actionable insight: monitor which of your company's documents are cited by Perplexity using a platform like LLMrefs, as this provides direct feedback on your content's authority.
Best Use Case: Use Perplexity for generating quick, source-linked research summaries, answering complex questions with citations, and as a starting point for deep-dive discovery tasks.
- Actionable Insight:
- Fast, source-linked syntheses are ideal for initial discovery and creating briefs.
- Enterprise plan options with admin controls support team and business use.
- Considerations:
- Pro plan inclusions and limits have changed; it's wise to verify before purchase.
- Some users report shifting quotas for deep research tasks, affecting predictability.
Website: https://www.perplexity.ai
4. DuckDuckGo (Search + Duck.ai)
DuckDuckGo has carved out a significant niche by positioning itself as the privacy-first alternative to mainstream search engines. Its core promise is simple yet powerful: it doesn't store your personal information or track your search history. This makes it an excellent choice for users concerned with digital surveillance and for SEO professionals seeking a less personalized, baseline view of search results. The platform provides a clean interface and relies on a mix of its own crawler, partner indexes (like Microsoft Bing), and other sources to deliver results.

Recently, the platform has integrated optional AI capabilities through Duck.ai, allowing users to ask questions in a conversational format without creating an account for basic use. This feature lets you interact with different AI models, adding a layer of direct-answer functionality. A powerful, practical feature is "Bangs," which are shortcuts for site-specific searches. For example, typing !a followed by a product name searches Amazon directly, or !w searches Wikipedia, providing an actionable workflow for efficient research. While a subscription unlocks more advanced AI models, the free offering remains a strong contender among which search engines are most highly recommended for private, everyday use.
Best Use Case: Set as the default search engine for daily, privacy-conscious browsing and for conducting research without the influence of a personalized search history.
- Actionable Insight:
- Strong privacy posture with zero search-history storage.
- Optional AI features (Duck.ai) that don't require an account for basic access.
- Considerations:
- Heavier reliance on partners versus a fully independent index can sometimes affect result depth.
- AI/chat usage has limits on the free tier, with premium models behind a subscription.
Website: https://duckduckgo.com
5. Brave Search
Brave Search offers a compelling alternative for those seeking independence from the Google-Bing duopoly. Its primary differentiator is its fully independent web index, which it built from scratch to reduce reliance on third parties and resist the systemic biases or spam that can plague other ecosystems. This makes it a fascinating platform for SEOs to monitor, as rankings can differ significantly from Google, revealing unique content opportunities. The engine integrates an AI-powered "Answer with AI" feature directly in its results, providing synthesized summaries for queries.

The platform’s most distinct feature is "Goggles," which allows users to create or apply community-made filters to re-rank search results. Here's a practical example: a user can apply a Goggle to prioritize results from smaller, independent blogs or filter out major corporate domains, making it a powerful tool for niche research. This offers an actionable way to find non-mainstream viewpoints. Brave Search also surfaces a "Discussions" module, pulling relevant conversations from forums like Reddit. While the core search is free with ads, a Premium subscription removes them entirely.
Best Use Case: Use Brave Search for privacy-conscious queries and for applying custom "Goggles" to filter results for specialized research, offering a view of the web untethered from Google's ranking signals.
- Actionable Insight:
- True index independence provides a hedge against Google/Bing algorithmic monoculture and surfaces different results.
- Customizable SERPs through Goggles create powerful, repeatable workflows for niche content discovery.
- Considerations:
- Its smaller index may not have the depth for extremely obscure, long-tail queries compared to Google.
- Ads are included in the free version; an ad-free experience requires a Brave Premium subscription.
Website: https://search.brave.com
6. Kagi
Kagi represents a different approach to web search, positioning itself as a premium, ad-free service for professionals and power users who value signal over noise. It operates on a subscription model, which completely removes ads and tracking, resulting in clean, signal-dense SERPs. This makes it a strong contender among search engines for anyone conducting serious research, as the experience is predictable and focused purely on delivering high-quality results without commercial interruption. The service is built for users who are willing to pay for a better, more private search experience.

Its standout features are centered on user control. A practical example of its "Lenses" feature is creating a custom search that only includes results from academic journals and trusted news sites, which is an invaluable and actionable tool for research. The integrated Kagi Assistant provides AI-powered summarization and chat, with access to multiple large language models. While it requires a paid subscription for any meaningful use and caps AI features based on fair-use credits, its focus on quality and customization makes it a highly recommended tool for dedicated searchers.
Best Use Case: Ideal for research-intensive tasks where result quality, speed, and the absence of ads are critical. Perfect for professionals, academics, and developers who need to fine-tune their search environment.
- Actionable Insight:
- Completely ad-free and private, providing an uncluttered, high-quality search experience.
- Powerful customization features like "Lenses" allow users to control and rank their search results.
- Considerations:
- Requires a paid subscription, making it a non-starter for casual users accustomed to free search.
- AI features and search quotas are subject to usage caps depending on the chosen pricing plan.
Website: https://kagi.com
7. Startpage
Startpage occupies a unique position for users who want Google-quality search results without Google's data tracking. It acts as a privacy-focused intermediary, submitting your query to Google anonymously and returning the results to you. This makes it an excellent choice for teams and individuals seeking the familiar relevance and deep index of Google but with a strict no-logs policy and robust privacy protections rooted in its EU-based operations. Its core value proposition is delivering results from the world’s largest index without compromising user data.

The platform’s standout feature is the "Anonymous View" proxy. Here's a practical use case: an SEO professional conducting competitor research can click "Anonymous View" next to a result to visit the site without revealing their IP address or leaving a footprint. This is a highly actionable tool for reducing digital trails during sensitive market analysis. While it lacks the rich, integrated features of Google's SERP like Knowledge Panels, its clean interface and dependable performance make it a highly recommended search engine for privacy-conscious users who don't want to sacrifice result quality.
Best Use Case: Use Startpage when you need Google-level search recall for professional or personal queries but want to prevent your search history and personal data from being collected.
- Actionable Insight:
- Easy switch for teams wanting Google-level recall with stronger privacy.
- Anonymous View proxy can reduce tracking when visiting result sites.
- Considerations:
- Dependent on Google’s index, so it isn't a fully independent alternative.
- Fewer bells and whistles compared with major engines; the experience is focused purely on web links.
Website: https://www.startpage.com
8. Ecosia
Ecosia offers a mission-driven alternative in the search engine market, directing its advertising profits toward climate action, primarily tree-planting initiatives. It operates on Bing’s search index, providing a familiar and reliable quality of results while allowing users to support an environmental cause with every query. This makes it an interesting option for teams and individuals looking to align their digital footprint with their values. For marketers, its reliance on the Bing Ads ecosystem means that existing campaigns can often reach this audience with minimal extra effort.
The platform’s unique value is its transparency, with a public counter showing the number of trees planted and regular financial reports. A practical insight for corporate ESG teams is that adopting Ecosia company-wide is a simple, measurable action that contributes to sustainability goals. While its core search functionality is powered by a partner, Ecosia is gradually introducing its own features, including AI summaries and an AI chat in beta for select regions. These emerging AI experiences signal an ambition to add more direct value beyond its partner-based results.
Best Use Case: Ideal for individuals and corporate teams who want to support environmental causes through their daily search activities without sacrificing the result quality of a major search index.
- Actionable Insight:
- Values-aligned brand for eco-minded users, with transparent reporting on its climate action projects.
- Familiar Bing-class result quality and integration with the Microsoft Advertising network.
- Considerations:
- Results and features are primarily partner-powered rather than from a fully independent index.
- New AI features have limited availability and vary significantly by region.
Website: https://www.ecosia.org
9. Mojeek
Mojeek offers a rare alternative in a market dominated by a few major indexes, operating its own crawler (MojeekBot) and independent search results from its UK-based servers. For professionals looking to understand web-wide content discoverability outside the Google/Bing bubble, it serves as an essential, unbiased data point. Its strict no-tracking, no-personalization policy means search results are consistent for all users, providing a stable baseline for assessing raw, non-profiled rankings. This makes it a useful tool when trying to answer the question of which search engines are most highly recommended for pure, unfiltered results.

The platform’s value lies in its genuine independence. A practical application for an SEO is to use Mojeek to check if a new piece of content has been indexed anywhere outside of Google, offering a true second opinion on its discoverability. Its clean, fast interface and available APIs for site search and web search also make it a practical option for developers wanting to integrate a privacy-first search solution. This focus on an independent, clean SERP is its defining characteristic.
Best Use Case: Use Mojeek as an independent control group to check a site’s raw indexation status, de-duplicate content, and get a second opinion on search results free from personalization or tracking.
- Actionable Insight:
- True independence with its own crawler and index, not reliant on Google or Bing.
- Simple, predictable rankings without behavioral profiling or filter bubbles.
- Considerations:
- Smaller index means recall can lag on niche and long-tail queries compared to larger engines.
- Fewer modern SERP modules and interactive features like knowledge panels or rich snippets.
Website: https://www.mojeek.com
10. Qwant
Qwant is a European search engine that champions privacy-by-design, positioning itself as a key alternative for users concerned with data tracking. Based in France, it operates with its own indexing technology but supplements its results with Microsoft Bing to ensure comprehensive coverage. This hybrid model makes it an interesting platform for SEO professionals looking to gauge visibility outside the dominant US-centric search ecosystems and understand how their content performs for an international, privacy-conscious audience.

The platform’s commitment to not using behavioral ad cookies or tracking user history for search personalization offers a "clean slate" search experience. A practical use case is for a brand launching in Europe to use Qwant to see how their website ranks for local-language keywords without any personalization bias. For marketers, this is an actionable way to test content in an environment free from filter bubbles. Qwant provides distinct search verticals, including News, Images, and a family-friendly mode called Qwant Junior, offering specific channels to assess content placement.
Best Use Case: Use Qwant to test content visibility for a European audience and to analyze search performance in a privacy-first environment without personalization bias.
- Actionable Insight:
- Strong privacy protection with no user tracking or search history logging.
- Provides important diversity signals for international SEO, especially for EU-based sources and queries.
- Considerations:
- Partial reliance on Bing means it is not a fully independent index for all result types.
- Relevance and feature depth for US-based queries may not match the leading US search engines.
Website: https://www.qwant.com
11. Phind
Phind has carved out a dedicated niche as an AI search engine built from the ground up for developers and technical users. It excels at answering programming queries, providing not just links but synthesized answers complete with functional code snippets. For tasks like debugging an error, understanding an API, or learning a new library, it often delivers faster, more actionable results than a general-purpose search engine. Its core strength is its ability to process complex technical questions and return code that is immediately usable.
The platform’s design is centered on developer productivity. A practical use case is asking, "How do I implement a secure OAuth 2.0 flow in a Python Flask application?" Phind will generate a step-by-step guide with code examples, saving significant research time. This makes it a highly recommended search engine for anyone whose work involves coding. It offers multiple search modes, including a "Deep research" option for more thorough analysis, and allows users to select from different AI models. While it offers free access, more advanced models and features are part of its paid plans.
Best Use Case: Use Phind as a primary coding assistant for debugging, generating boilerplate code, and getting quick, accurate answers to technical or API-related questions.
- Actionable Insight:
- Highly specialized for programming help, offering faster and more precise code-based answers.
- Synthesizes information from multiple sources into a single, coherent response with citations.
- Considerations:
- General knowledge recall is less comprehensive than broad-based search engines like Google.
- Full pricing details and plan limitations are often presented after account creation.
Website: https://www.phind.com
12. Andi
Andi distinguishes itself as an AI-first conversational search engine that prioritizes user privacy and an ad-free experience. Rather than serving a list of links, it generates synthesized answers in a chat-like interface, aiming to provide direct, accurate information drawn from its proprietary AI index, named 'Trantora'. This approach makes it a strong contender for users looking for quick overviews and direct answers without the clutter of traditional search engine results pages. For SEO professionals, it represents another facet of answer engine optimization.

The platform’s clean user experience is a major draw, offering modes to summarize, explain, or generate content based on a query. A practical example is using Andi to ask "Explain quantum computing like I'm five," to get a simple, jargon-free summary. Its free consumer search requires no account, making it highly accessible. For developers and teams, Andi provides an API with paid tiers, enabling integration of its AI search capabilities into other applications. This dual model makes it a useful tool for both personal research and professional product development.
Best Use Case: Use Andi for quick, ad-free, synthesized answers to straightforward questions, or for exploring its API for custom AI search integrations.
- Actionable Insight:
- Ad-free, no-account consumer search with a clean, conversational interface.
- Generates direct answers with source links, which is ideal for fast overviews.
- Considerations:
- As a younger product, its index depth and the nuance of its answers are still maturing.
- Premium features and the future roadmap are subject to change, introducing some uncertainty.
Website: https://andisearch.com
Top 12 Search Engines: Recommendations & Comparison
| Engine | Core strengths | AI answer behavior | Target audience | Unique selling point | Price / Access |
|---|---|---|---|---|---|
| Google Search (AI Overviews/AI Mode) | Massive index, rich SERP features, strong local & vertical coverage | AI Overviews with source links; behavior evolving | Agencies, brands, SEO teams needing max recall | Best US recall and diverse SERP features | Free; enterprise/experimental features via Search Labs |
| Microsoft Bing + Copilot | AI summaries with sentence‑level citations, strong news/media index | Clear source‑linked answers, Copilot context across M365 | Microsoft ecosystem teams, enterprises | Copilot + Microsoft 365 integration, inline citations | Free; Copilot paid tiers for advanced features |
| Perplexity | Research‑focused browsing, live web access, enterprise controls | Fast, source‑linked syntheses and briefs | Knowledge workers, researchers, enterprise teams | Live web browsing and deep research modes | Free + Pro/enterprise plans |
| DuckDuckGo (Search + Duck.ai) | Privacy‑first, bangs shortcuts, browsers/extensions | Optional Duck.ai chat with model choice; limits on free tier | Privacy‑conscious users and teams | Zero search‑history storage and privacy bundles | Free; subscription for Duck.ai premium & VPN |
| Brave Search | Independent index, Goggles community filters, dev APIs | 'Answer with AI' blended into SERP, community re‑ranking | Privacy-minded researchers, developers | Independent index + customizable SERPs (Goggles) | Free; Premium ad‑free and API paid plans |
| Kagi | Ad‑free, customizable ranking (Lenses), team plans | Kagi Assistant with multi‑LLM access on paid tiers | Power users, teams, heavy searchers | Signal‑dense, uncluttered results and Lenses | Paid subscription (tiered plans) |
| Startpage | Google‑style relevancy with EU privacy stance, Anonymous View | Google‑quality results; proxy browsing to hide IP | Teams wanting Google recall with stronger privacy | Anonymous View proxy to reduce tracking | Free; privacy‑focused service |
| Ecosia | Bing‑powered results, climate donations, privacy policies | Emerging AI summaries/chat in select regions | Eco‑minded users and organizations | Profits fund tree‑planting (values alignment) | Free; AI features region‑dependent |
| Mojeek | Own crawler and independent index, strict no‑tracking | Predictable, non‑personalized rankings | Researchers needing independent second‑opinion | True independence from Google/Bing | Free; site search & web APIs available |
| Qwant | EU privacy‑by‑design, selective partner use, verticals | Privacy‑focused results; feature depth varies by region | European researchers, privacy‑oriented teams | European privacy signals and diversity | Free; uses partners for some results |
| Phind | Developer‑tuned search, code syntheses, model selection | Fast, actionable code answers and docs synthesis | Developers, engineers, dev teams | Optimized for coding, debugging, API lookups | Free; business plans available via sign‑up |
| Andi | AI‑first ad‑free index (Trantora), summaries & modes | Synthesized high‑accuracy answers with source links | Consumers wanting clean, fast overviews | Ad‑free UX with claimed large AI index | Free; paid developer/team tiers |
Building Your Modern Search Strategy
The journey through today's top search engines reveals a clear and compelling truth: the era of relying on a single, one-size-fits-all engine is over. While Google remains a formidable force for general discovery, the most effective information-seekers, marketers, and product managers are building a personalized toolkit. Answering the question of which search engines are most highly recommended no longer has a single answer; it has a strategic one, tailored to the specific task at hand.
We've explored the powerhouse incumbents like Google and Bing, which are now integrating AI Overviews and Copilot to provide instant answers. We've also delved into the world of dedicated "answer engines" like Perplexity and Phind, which are purpose-built for deep, cited research and technical problem-solving. Finally, we've highlighted the crucial role of privacy-first alternatives such as DuckDuckGo, Brave Search, and Startpage, which offer a necessary shield in a data-hungry world.
The most significant takeaway is that choosing a search engine is now an act of intention. Your selection should be a conscious decision based on your immediate goal, whether that is privacy, research depth, or broad-based discovery.
From Passive Searching to Active Information Curation
Adopting a multi-engine approach is the first practical step. This isn't about abandoning your default; it's about augmenting it. Think of your search process as a workshop, not a single hammer. Here's an actionable plan:
- For Everyday Privacy: Make a privacy-centric engine like DuckDuckGo or Brave Search your browser's default. This handles the bulk of your daily, low-stakes queries without building a long-term data profile.
- For Deep Research & Factual Accuracy: When you need authoritative, well-cited answers for a report, project, or technical question, pivot directly to a specialist like Perplexity, Kagi, or Phind. Their focus on sourcing and synthesis saves immense time and improves the quality of your output.
- For Broad Discovery & Local Queries: For finding local businesses, exploring broad topics, or accessing the sheer volume of the web, Google and Bing remain incredibly powerful. Use them when you need the widest possible net.
By segmenting your search activity this way, you gain more control over your data, get to high-quality information faster, and build a more complete picture of the web.
The New Imperative for Brands and SEO Professionals
For businesses, founders, and marketing teams, this fragmentation presents a critical new challenge and a major opportunity. Your audience's attention is no longer concentrated in one place. They are asking questions and seeking solutions across a spectrum of traditional search engines, private alternatives, and AI answer engines.
A modern visibility strategy is not about just ranking on Google. It's about being the cited authority wherever your customers are asking questions. This means your content must be valuable enough to be surfaced by Perplexity, mentioned in a Copilot summary, and discovered through a privacy-focused Brave search.
Your content needs to be optimized for a new kind of "search": the answer. This requires a shift in focus from pure keyword targeting to creating clear, authoritative, and well-structured information that AI models can easily parse and present. The goal is to become a trusted source that these new gatekeepers cite. An actionable insight is to use an excellent service like LLMrefs to monitor your brand's presence and source citations across these platforms. This is no longer optional; it's a core component of modern brand management and SEO, and it's how you discover which search engines are most highly recommended by your actual target audience.
Ultimately, the power has shifted. We now have an incredible array of tools at our disposal, each with distinct strengths. The smart user will create a personal search stack, and the smart brand will learn to be visible across all of them. This is not just a change in technology; it's a fundamental change in how we access and verify knowledge. Embracing it gives you a distinct advantage in finding clarity and building authority.
Are your brand's key messages and data being cited correctly by AI answer engines like Perplexity and Copilot? LLMrefs is the first and most effective monitoring platform built to track your source citations and brand mentions across the leading AI search tools. Visit LLMrefs to see where your content is powering the answers of tomorrow and ensure your authority is recognized.
Related Posts

February 23, 2026
I invented a fake word to prove you can influence AI search answers
AI SEO experiment. I made up the word "glimmergraftorium". Days later, ChatGPT confidently cited my definition as fact. Here is how to influence AI answers.

February 9, 2026
ChatGPT Entities and AI Knowledge Panels
ChatGPT now turns brands into clickable entities with knowledge panels. Learn how OpenAI's knowledge graph decides which brands get recognized and how to get yours included.

February 5, 2026
What are zero-click searches? How AI stole your traffic
Over 80% of searches in 2026 end without a click. Users get answers from AI Overviews or skip Google for ChatGPT. Learn what zero-click means and why CTR metrics no longer work.

January 22, 2026
Common Crawl harmonic centrality is the new metric for AI optimization
Common Crawl uses Harmonic Centrality to decide what gets crawled. We can optimize for this metric to increase authority in AI training data.