When did generative engine optimisation and answer engine optimisation start?

By Yamm Labs · Updated September 2026 · 6 min read

A head of digital at a life insurer is building next year’s plan. One line reads GEO, the next reads AEO, and the finance head asks a fair question: are GEO and AEO new, or old SEO with new names? The answer changes the budget. One of the two ideas grew out of search features from the 2010s. The other has a birth certificate: a research paper posted in November 2023.

Where did answer engine optimisation come from?

AEO grew out of Google’s move from ten blue links to direct answers. In May 2012 Google introduced the Knowledge Graph, which let Search understand people, places and things rather than match strings of words. In January 2014 it introduced featured snippets, the answer box at the top of a results page that quotes one source.

By January 2018 Google was describing how, with the growth of voice assistants, a spoken answer on Google Home cited its source page, in a post about featured snippets. That is the environment in which search marketers began to talk about optimising for an answer rather than a ranking. Who first used the phrase answer engine optimisation is still disputed, so treat any exact birth year you see quoted with caution.

When did generative engine optimisation start?

GEO has a precise starting point. OpenAI released ChatGPT on 30 November 2022, and generated answers moved from a lab curiosity to a daily habit. On 16 November 2023 a team of academic researchers posted “GEO: Generative Engine Optimization” to arXiv. The paper named the practice, built a benchmark for it and reported that, in its tests, some content changes raised a source’s visibility in generated answers by up to 40%.

The difference from AEO is the engine. A featured snippet quotes one page. A generative engine reads many pages, writes its own answer and may or may not cite you. GEO is the work of being one of the sources it trusts and names.

AEO taught brands to be the answer; GEO asks them to be the source an engine trusts enough to name.

When did GEO and AEO start to matter for Indian financial brands?

For India, the practical start was August 2023, when Google let people in India opt into its generative Search experiment in English and Hindi, with a toggle between the two and the option to listen to responses. A year later, on 15 August 2024, Google announced AI Overviews for India and five other countries. From then on, a question about term insurance or a SIP could be answered by Google before a buyer reached any bank, insurer or fund house’s page.

Which regulator rules shape GEO and AEO work for a bank, insurer or fund?

The rules did not change when answers became generated. What changed is that an engine can lift one paragraph out of a page and show it without the footer. Whether a given page counts as an advertisement is a call for your compliance team. We design as if it does, and that forces specific decisions.

• Insurers: the registered name travels with the answer. IRDAI’s 2024 policyholder-protection regulations require an insurer’s advertisements to show its registered name along with its trade name or logo prominently, and define advertisement as a communication through any mode or medium. The design decision: every quotable answer block names the insurer by its registered name as well as the short brand name.
• Fund houses: the warning sits next to the claim. SEBI requires the mutual fund standard warning in audio-visual advertisements to be displayed and voiced for five seconds. The design decision: a fund explainer cut for YouTube keeps the five-second end frame, and on the page the risk line sits inside any section that discusses returns as well as in the footer.
• Lenders: the APR line is part of the answer. RBI requires a Key Facts Statement, including the annual percentage rate, before the loan contract is executed, for retail and MSME term loans sanctioned from 1 October 2024. The design decision: a page that answers “what will this loan cost” points to the APR in the Key Facts Statement inside the answer paragraph, not in a tooltip.
• Timeline: compliance reviews each answer block. Because every block can be quoted alone, compliance signs off block by block. Plan one review round per batch of answer pages from the first week.

SIGNALNewsletter by Yamm Labs

The next piece, straight to your inbox.

New writing on BFSI brands, B2B and influencer marketing, and AI search.

Email *

Subscribe

Subscription form is not available at the moment

One click to unsubscribe. Privacy policy

Should a financial brand buy GEO and AEO as two separate services?

No. Treat them as one programme with two layers. AEO is the base: question headings, a direct first sentence, clean structure and schema. GEO is the layer on top: being present, accurate and cited in the answers ChatGPT, Perplexity, Gemini and Google’s AI Overviews write.

The decision to make is which buyer questions you must own, which engines you will track and how you will measure being named. Yamm Labs works from Gurugram as an AI visibility agency for BFSI brands, serving clients in India, the Gulf and elsewhere remotely. Our BFSI clients have included TATA AIA Life Insurance, HDFC Life, Axis Bank, Paytm Money and Aditya Birla Sun Life Mutual Fund. The short version of this answer sits in our AI Search Optimisation: 100 Questions Answered hub, and definitions live in our GEO and AEO FAQ for financial brands.

Common myths about when GEO and AEO started

The dates matter because each myth leads to a different budget decision. These are the ones we hear most often from marketing and compliance teams.

• “AEO started with ChatGPT.” It did not. The work of shaping a page so an engine quotes it began with featured snippets and voice answers, years before generated answers became common.
• “GEO is just a new name for SEO.” The two share foundations, such as crawlable pages and clear structure. But the 2023 paper described a different target: being one of the sources a generative engine reads and cites, not a position in a list of links.
• “AEO has a known founder and birth year.” It has neither on record. Any single year you see quoted is someone’s estimate, so treat it as one.
• “GEO arrived in India with AI Overviews.” Indian searchers could opt into Google’s generative Search experiment from August 2023. AI Overviews, announced for India in August 2024, made generated answers the default.
• “The 40% figure is a promise.” The paper reported that some content changes raised visibility by up to 40% in its own tests. That is a research result under set conditions, not a forecast for any brand’s pages.

A worked example: turning the timeline into a plan

Go back to the head of digital at the life insurer, with GEO and AEO on two lines of next year’s plan. The timeline answers the finance head’s question in three parts.

First, the AEO line is not new spend on a new idea. It is the tidy-up the site should have had since featured snippets: question headings, a direct first sentence under each, clean structure and schema. The team can audit its top product and claims pages against that standard and fix them in order.

Second, the GEO line is newer and needs its own measure. The question is not where a page ranks but whether ChatGPT, Perplexity, Gemini and Google’s AI Overviews name the insurer when a buyer asks about term cover. That needs a fixed list of buyer questions and a record of the answers over time.

Third, the compliance line sits across both. Every answer block that can be quoted alone carries the registered name, and the review is done block by block. Written this way, GEO and AEO become one programme with one owner, which is what the finance head needed to see.

A step-by-step checklist for starting GEO and AEO work

This order follows the history: build the answer layer first, then measure and grow presence in generated answers.

1. List the buyer questions you must own. Start with the questions a customer asks before choosing a policy, loan or fund.
2. Match each question to one page. One question, one page, with the answer in the first sentence under a question heading.
3. Fix the structure. Clear headings, short paragraphs, tables where they help, and schema that describes the page honestly.
4. Put the compliance lines inside each answer block. Registered name, risk warning or APR pointer, wherever the block could be quoted alone.
5. Choose the engines you will track. For example Google’s AI Overviews, ChatGPT, Perplexity and Gemini.
6. Record a baseline. Ask each engine the same questions and note whether your brand is named, how it is described and which source is cited.
7. Review and repeat. Update pages where answers are wrong or missing, then ask the same questions again on a fixed schedule.

Regulator rules for advertisements apply however an answer is shown. Confirm the current text with your compliance team before each batch of answer pages goes live.

How to brief an agency or team on GEO and AEO

A short, specific brief keeps the work on one programme rather than two. Include these points:

• The buyer questions you want to own, ranked by business value.
• The pages that should answer them today, and any that are missing.
• The regulators whose advertising rules apply to you, and who signs off in compliance.
• The engines to track and how often to check them.
• How you will judge progress: being named, being described accurately and being cited.

Ask the team how it will separate the AEO base from the GEO layer in its reporting. A good answer names the pages it will fix first and the questions it will track. Be wary of anyone who promises a place in a generated answer. Engines decide what they cite, and no one outside them controls that.

People also ask

Is generative engine optimisation the same as answer engine optimisation?

No, though they overlap. AEO aims to be the single answer an engine quotes, as in a featured snippet or a voice reply. GEO aims to be one of the sources a generative engine reads, trusts and names when it writes its own answer.

Who coined the term generative engine optimisation?

The term was set out by a team of academic researchers in the paper “GEO: Generative Engine Optimization”, posted to arXiv on 16 November 2023. It is the earliest widely cited use; informal uses before it may exist.

Do financial brands still need SEO if they invest in GEO and AEO?

Yes. Generative engines and answer boxes still draw on pages that search engines can crawl and trust. SEO keeps those pages findable; AEO and GEO make them quotable and citable.

Which came first, GEO or AEO?

AEO came first. It grew out of Google’s featured snippets, introduced in January 2014, and the spoken answers voice assistants read aloud later in the decade. GEO was named in a research paper posted to arXiv on 16 November 2023, almost a year after ChatGPT’s release. The two now work as layers of one programme.

When did AI Overviews reach India?

Google announced AI Overviews for India and five other countries on 15 August 2024. Before that, from August 2023, people in India could opt into Google’s generative Search experiment in English and Hindi. For Indian financial brands, the 2024 launch is the point at which generated answers began to appear by default above organic results.

Is AEO the same as optimising for featured snippets?

Featured snippets were the first big target for AEO, but the idea is wider. AEO covers any format where an engine gives one direct answer, including voice replies and answer boxes. The method is the same across them: a clear question, a direct first sentence, clean structure and a page the engine can trust.

Why does the start date of GEO and AEO matter to a financial brand?

Because it separates old work from new. The AEO base, clear structure and direct answers, is work a site should already have done. The GEO layer, tracking whether AI engines name and cite the brand, is newer and needs its own measure. Knowing which is which helps a marketing head explain the budget to finance and compliance.

Did voice search start answer engine optimisation?

Voice search was part of it. In January 2018 Google described how a spoken answer on Google Home cited its source page, which made being the single answer a clear goal. But featured snippets came earlier, in 2014, and the Knowledge Graph before that, in 2012. AEO grew from all three rather than from one launch.