Research and development

Published work from the team behind AnswerRank

AnswerRank is built on research the team has published under its own name, software it has released under an open licence, and analysis it has put on a dated page. Each item below links to the copy held by an independent host, so the claim can be checked without taking our word for it.

Peer-distributed preprint

The Death of a Search Engine

The Progression of Content Discovery, the Mechanics of Intent, and What the Answer-First Web Asks of Businesses

Author
Himanshu Bisht
Year
2026

Abstract

The first point of contact between a person and the internet has changed shape five times in three decades: from human-curated directories to crawler-based keyword engines, to link-authority ranking, to enriched result pages, and now to answer-first interfaces built on transformer language models. This paper traces that progression and examines what it does to user behavior and to the economics of being found. Drawing on clickstream panels, clickthrough studies, and the technical lineage that runs from the attention mechanism of Vaswani et al. (2017) through retrieval-augmented generation, the paper formalizes discovery as an expected-visit model in which traffic decomposes into a shrinking ranked-list term and a growing citation term. The evidence reviewed shows a consistent pattern: sixty-eight percent of US Google searches now end without a click, click-through to top-ranked pages falls by more than half when an AI summary is present, and yet visitors who do arrive from AI systems convert at materially higher rates. The paper then revisits Broder's intent taxonomy under synthesis, showing which intent classes the answer layer absorbs and which it leaves intact, and derives five imperatives for businesses: be retrievable, be citable, be corroborated, be measurable, and be worth the click. The strategic variable has moved from position on a page to probability of inclusion in an answer.

What it establishes

The paper models discovery as an expected number of visits, and splits that number into a ranked-list term and a citation term. The first term is shrinking and the second is growing, which is the formal version of what AnswerRank measures: the probability that an assistant includes a brand in its answer, rather than the position a page holds on a results list. The five imperatives it derives, retrievable, citable, corroborated, measurable and worth the click, are the five things the product reports on.

Open-source software

Rank Monkey, a WordPress plugin under the GPL

Research-driven articles written for answer engines, published from a WordPress site with the site owner's own model keys.

Licence
GPL-2.0, source public
Language
PHP
Current release
1.1.1 on the WordPress.org directory

Why it is open

The plugin puts the editorial method the paper argues for into code: research first, a plan, a draft, a revision pass, structured data that engines can lift. Publishing the source under the GPL means the method can be read, audited and modified by anyone, and the WordPress.org review process applies its own checks before each release. The plugin bills nothing through us; it runs on the site owner's own OpenRouter and fal.ai keys.

Long-form analysis

Generative Engine Optimization for Business Owners

Most business owners still chase Google rankings, yet the visibility that now decides who gets quoted by ChatGPT or Gemini comes from making your content the single clearest answer a model can lift and cite.

Author
Himanshu Bisht
Published
2026-07-03
Format
Field guide, on this site

What it covers

The applied companion to the paper. Where the preprint establishes that inclusion in an answer has replaced position on a page as the variable that matters, this article works through what a business owner changes as a result: which pages get rewritten, which third-party surfaces need attention, and how to tell whether any of it worked when the traffic report no longer records the visit.

The author

Himanshu Bisht

Founder, AnswerRank. Former Forbes Councils member and growth operator with eight years building distribution for startups across ecommerce, SaaS, wellness and local business. The paper, the plugin and the blog are written under that name, and the product is the instrument built to measure what the paper describes.

How to cite

Bisht, H. (2026). The Death of a Search Engine: The Progression of Content Discovery, the Mechanics of Intent, and What the Answer-First Web Asks of Businesses. SSRN. https://doi.org/10.2139/ssrn.7136778