2026 AI Search Data Study · GEO

If the official site is cited more often, does that mean AI recommends the brand?

30 days, 7 AI platforms, 66,866 answers: when the official site is cited more often, does that mean the brand is already being recommended?

AuthorDageno teamPublishedNovember 2026DataDageno AI answer database, Oct 19 to Nov 17, 2026Scale66,866 answers · 7 AI platforms

Key findings

Short answer: being cited is not the same as being recommended. If a GEO team treats official-site citations as proof that AI already recommends the brand, it will overstate eSign.AI's position. In this dataset the official site is cited 175 times, while the brand appears in only 99 answers, a 0.15% mention rate. It ranks 121st on e-signature questions, and it does not appear once in 16,986 answers about contract lifecycle management.
Figure 1. Official-site citations outnumber brand mentions; the CLM mention rate is 0%
Figure 1. Official-site citations outnumber brand mentions; the contract-lifecycle mention rate is 0%. Oct 19 to Nov 17, 2026, 7 AI platforms. Card 01 is counts; cards 02–04 are mention rates under their own denominators.
  1. 175 vs 99. The official site is cited 175 times; the brand is named 99 times. esign.ai is the 46th most cited website, appearing in 0.26% of all answers.
  2. 0% vs 0.3%. In 16,986 answers on contract lifecycle management, eSign.AI's mention rate is 0%. In 30,697 e-signature answers, it is 0.3%, rank 121 of 1,115 brands. DocuSign's mention rate in e-signature answers is 52.9%.
  3. 0.25% vs 0%. We opened esign.ai for this article and found template-management and API pages, but the mention rate is still 0.25% on "API and system integration" (4,338 answers, rank 129) and 0% on "contract templates and document generation" (1,469 answers). Having the pages does not mean these questions name the brand.
  4. 1.4% vs 0.06%. Google AI Mode names eSign.AI in 1.4% of answers (37/2,637). Perplexity names it in 0.06% (19/30,678). Each rate uses that platform's own answers as the denominator; neither can stand in for the other.

Brands selling e-signature or contract-management products often start with the more flattering number: how often the official site is cited. Agency weekly reports also tend to put citation counts and mention counts on the same line, as if they measured the same kind of visibility.

For eSign.AI the two numbers move in opposite directions: citations outnumber mentions. We used 30 days of contract-management and e-signature answers in the Dageno AI answer database, 66,866 in total, to see what a citations-only view misses. The window is Oct 19 to Nov 17, 2026, including both end dates, across 7 AI platforms. Each answer records which brands were named, which brand came first, and which pages were cited.

This piece uses three metrics. Mention rate: the share of answers that name the brand. Top pick: the first brand an answer recommends. Citation: a web page given as a source in an answer. These metrics describe how AI answers. They do not judge product quality or show causation. The report has no separate figures for Hong Kong or Southeast Asia, and this article does not draw visibility conclusions for those markets.

01|If the official site is cited more often, does that mean AI recommends the brand?

No. In this dataset the official site is cited 175 times, and the brand is named only 99 times.

eSign.AI appears in 99 of 66,866 answers, a mention rate of 0.15%. 77.6% of all answers named a brand (51,877 answers). DocuSign's mention rate is 37.8%, and it is the first recommendation in 18.4% of those answers; eSign.AI's top-pick share is 0.01%.

At the same time, 23.1% of answers cited a web page. Counted as answer × website, cited sources break down as 74.88% brand and company sites, 24.09% media, directories and other sites, and 1.03% community, video and review sites. These three shares are chart values and add up to 100%. esign.ai is the 46th most cited website, at 0.26% of all answers. 0.26% × 66,866 is about 174, close to the 175 citations in the report body. This article uses 175.

Figure 2. Official-site citations outnumber brand mentions
Figure 2. Official-site citations outnumber brand mentions. Source types are chart values, counted as answer × website.

When citations outnumber mentions, some linked answers opened the official site without recommending eSign.AI as a brand, or named it less often than they cited it. That should not be read as "already recommended 175 times".

The direction matches an earlier, larger Dageno study: Do AI answers that cite external pages also name brands? The platforms, industries and periods differ, so only the direction can be compared, not the figures. The sites each platform cites most also differ: Which websites do AI answers cite most? ChatGPT and Google's AI don't cite the same ones.

For brands: put official-site citations and brand mentions on two separate rows in the weekly report. Keep both 175 and 99 in view; do not keep only the citation count.
For agencies: do not write a client's official-site citation rank as if it were a recommendation rank. esign.ai is the 46th most cited site, and the brand ranks 121st on e-signature questions. Those are two different metrics.

02|Are contract management and e-signature the same list?

No. The brand is rare in e-signature answers and absent from contract-lifecycle questions. A blended 0.15% hides both facts.

The report splits questions into 2 product categories. Their answer counts add up exactly to 66,866, with no overlap.

SliceAnswersNamed a brandMention rateRankTop three
E-signature (category)30,69783.3%0.3%121 of 1,115 brandsDocuSign 52.9%, Adobe 31.5%, Dropbox 18.3%
Contract management (category)36,16972.7%0.02%Outside the top 200Ironclad 27.8%, DocuSign 24.9%, Icertis 18.6%
Contract lifecycle management (topic)16,986—0%—DocuSign 34.8%, Ironclad 32.3%, Agiloft 23.6%
E-signature platforms and signing experience (topic)17,238—0.48%81DocuSign 61.9%, Adobe 39.3%, Dropbox 26.1%

Category and topic denominators differ and cannot be subtracted. An answer can be counted in more than one topic.

Figure 3. Mention rate is 0% in 16,986 CLM answers, and 0.3% in the e-signature category
Figure 3. Mention rate is 0% in 16,986 CLM answers, and 0.3% in the e-signature category. Category and topic denominators differ and cannot be subtracted.

We opened the Chinese and English homepages on esign.ai on Nov 20, 2026. The homepages describe a platform that prepares, signs and archives contracts, and they list template management, bulk send, identity verification and compliance. The report's industry classification, contract management, is correct. Those homepage claims do not place the brand in the 16,986 contract-lifecycle answers.

For brands: do not use the blended 0.15% to decide whether the brand is on the list. Report at least two rows: 0.3% and rank 121 for e-signature, and 0% for contract lifecycle management.
For agencies: do not build one mixed "contract management" prompt set and then give a single rank. The client is absent from CLM questions and outside the top 100 for e-signature. Those are different diagnoses, and they point to different next steps.

03|If the official site already has API and template pages, will those questions name the brand?

Not automatically. The pages are on the official site, and mention rates on the matching topics are still very low, including a zero.

One claim in the source report needs a correction. In the "eSign.AI coverage" column, the report said API, embedded signing and third-party integrations were "not found on the official site", and that a template library was "not found" either. We opened the official site for this article and that judgment does not hold. This article does not use it. Pages we verified include template management and bulk send on the homepage; API documentation in the developer centre for listing templates and generating files from templates; and a pricing page that states the Professional plan includes API capabilities (this article does not discuss prices). The homepage also lists SOC 2 Type II, ISO 27001, GDPR and eIDAS.

TopicAnswersMention rateRankTop three / notes
API and system integration4,3380.25%129DocuSign 57%, Adobe 27.2%, Dropbox 22.8%
Contract templates and document generation1,4690%—DocuSign 57.1%, PandaDoc 34.6%, Anvil 31.9%. Fewer answers; the zero is a clear absence, but rank commentary should stay cautious
Identity verification, security and signing compliance4,2300.47%93DocuSign 50.7%, OneNotary 25.4%, Adobe 19.7%
Software pricing and licensing1,2600.63%56Highest rate among the listed topics, but only 1,260 answers; directional only, not a core finding

Each rate uses that topic's answers as the denominator.

Figure 4. API and template pages exist on the official site; mention rates on those topics stay very low
Figure 4. API and template pages exist on the official site; mention rates on those topics stay very low. Official site checked on Nov 20, 2026.

For brands: making "publish an API page and a template page" the first GEO action is not what this dataset supports — those pages are already there. A more useful first check is which pages were among the 175 citations, whether those answers also named eSign.AI, and whether third-party reviews and directories mention the brand. Finding the layer where the gap sits is not a promise that filling it will produce recommendations.
For agencies: do not write an execution list from a report line that says a page type is "missing". Open the official site first. In this report, at least two "not found on the official site" claims do not match the current pages.

04|Can the mention rate on one platform stand in for the others?

No. The mention rate on Google AI Mode is about 20x the rate on Perplexity. Each figure can only be read against its own denominator.

Mention rates below use each platform's own answers as the denominator.

PlatformAnswersNamed a brandCited a pageMentionsMention rateNotes
Google AI Mode2,63786.4%54.7%371.4%Top three: DocuSign 44.7%, CloudSign 37%, GMO GlobalSign 28.1%
Google AI Overviews2,25886%46.5%50.22%
Gemini11,79789.5%16.4%170.14%
ChatGPT13,81190.4%24.8%150.11%
Copilot5,07168.4%18.1%40.08%
Perplexity30,67867.4%21.4%190.06%
Grok58977.4%23.1%10.17%Too few answers; directional only

Each rate uses that platform's own answers as the denominator. Answer counts describe sample size only.

Figure 5. The mention rate on Google AI Mode is much higher than on Perplexity
Figure 5. The mention rate on Google AI Mode is much higher than on Perplexity, which has the most answers. Answer counts describe sample size, not platform importance.

1.4% divided by 0.06% is about 23x (calculated). That is not because Google AI Mode is the largest slice of the sample — it has 2,637 answers, far fewer than Perplexity's 30,678. The gap is in the rate, under each platform's own denominator. The report's line that the brand "depends almost entirely on one platform" is not accurate and should not be treated as a finding: Perplexity, Gemini and ChatGPT also name the brand, just at lower rates.

Platform-level mention counts add up to 98, which is 1 short of the 99 in the report body. This article uses 99. The gap is flagged for backend confirmation.

One explanation we cannot confirm: Google AI Mode cites a web page in 54.7% of answers, versus 21.4% on Perplexity, and may therefore open esign.ai more often. That is an inference. The report does not break down esign.ai citations by platform, so it cannot be confirmed.

For brands: report Google AI Mode separately from Perplexity, ChatGPT and Gemini. Do not use either 1.4% or 0.06% as the figure for "all AI platforms".
For agencies: platform comparisons must state the denominator. Google AI Mode's 2,637 answers are enough to read directionally, but they are well below Perplexity and ChatGPT. Grok's 589 answers should not go into a KPI.

05|Does AI recommend the same brands in every market?

Not necessarily. The US still shows a very low mention rate. Japan has 5,922 answers and zero mentions. There is no separate data for Hong Kong or Southeast Asia, and this article does not draw conclusions for those markets.

MarketAnswerseSign.AI mention rateTop three / notes
United States35,0320.21% (75 mentions)78.8% named a brand. Top three: DocuSign 38.3%, Ironclad 19.9%, PandaDoc 17.8%
Germany8,3270.04%Top three: DocuSign 37%, Skribble 22.9%, Adobe 22.3%. A local brand is in the top three
Japan5,9220%Top three: CloudSign 73.6%, GMO GlobalSign 68.7%, GMO Internet Group 61%. A zero with a large enough denominator to read
India6011%Too few answers to support a conclusion; directional only
Spain + Mexico1,922Named 3 timesToo few to support a separate conclusion

Answer counts describe sample size, not market size. There is no separate data for Hong Kong or Southeast Asia.

The official site says the company is headquartered in Hong Kong and covers APAC. That is brand positioning, not visibility data from this report. The report has no separate figures for Hong Kong or Southeast Asia, and those markets cannot be inferred from Japan or the US.

For brands: if Japan is a priority market, the US figure of 0.21% is not Japan's position. Japan is at 0% in 5,922 answers, and the top three brands are local.
For agencies: check each market's answer count before it goes into a formal comparison. Japan's zero has a sufficient denominator and can be reported. India's 601 answers, and 3 mentions across Spain and Mexico, are clues only.

06|How can brands and agencies use these findings?

The points below refer back to the figures above. They do not promise that mention rates will rise.

  1. Report official-site citations and brand mentions on two separate rows. In this dataset that is 175 versus 99. Reporting citations alone will overstate how often the brand is already recommended (see 01).
  2. Split e-signature and contract lifecycle management; do not use only the blended 0.15%. The former is 0.3% and rank 121; the latter is 0% in 16,986 answers (see 02).
  3. Open the official site before writing "publish API and template pages" as the first action. Those pages are already there, and mention rates on the matching topics are still 0.25% and 0% (see 03).
  4. Report platforms and markets separately, with sample sizes. Google AI Mode at 1.4% cannot stand in for Perplexity at 0.06%, and Japan's 0% cannot be replaced by the US figure of 0.21% (see 04, 05).
  5. Do not accept a single screenshot as proof. Another Dageno study found that when the same question was asked twice on the same day (June to September 2026, 5 AI platforms, 7,030 comparisons), after removing pairs where neither answer named a brand, ChatGPT returned the exact same brand list only 7.4% of the time: Ask AI the same question again the same day: are the recommended brands the same?

Want citation, mention, topic and platform splits for your own category? Public market data is free forever after sign-up at Dageno, covering 10 AI platforms, 86 countries and regions, and 12,000+ market segments. Agencies that want to use this kind of breakdown in client proposals and reviews can look at the Dageno partner program; white-label reports for agencies are coming soon.

Check your AI search market intelligence for free at Dageno →

Methodology

Period30 daysOct 19 to Nov 17, 2026 (inclusive)
Answers66,866Contract management 36,169 · E-signature 30,697
AI platforms7Perplexity, ChatGPT, Gemini, Copilot, AI Mode, AI Overviews, Grok
Prompt set2 categoriesPlus 10 topics; an answer can be in more than one topic

Definitions: mention rate is answers naming the brand divided by answers in that slice. Top pick is the first brand recommended in an answer; top-pick share uses answers that named a brand as the denominator (51,877). Source types are counted as answer × website. Official pages are taken from the Chinese and English esign.ai homepages, the developer centre and the pricing page, opened on Nov 20, 2026. The report's claim that API and template pages were "not found on the official site" does not match the current site and is not used as a finding.

Sample composition and limits: answers cover 55 countries and regions, with the US at 35,032 answers, or 52.4%, so overall figures reflect the US to a large extent. This is how the database is collected, not market share. Germany has 8,327 answers and Japan 5,922; other markets have fewer. Grok has only 589 answers; India and similar markets are directional only. Prompt mixes differ by platform and market. The report has no separate data for Hong Kong or Southeast Asia, and this article does not draw visibility conclusions for those markets. All findings are correlations, not causation.

Platform-level eSign.AI mention counts add up to 98, versus 99 in the report body, a gap of 1. Flagged for backend confirmation; this article uses 99.

Sources

Dageno AI answer database, Oct 19 to Nov 17, 2026. Brand report: eSign.AI AI visibility assessment.
eSign.AI official site https://www.esign.ai/zh-CN, https://www.esign.ai/en-US/, pricing https://www.esign.ai/pricing, and developer-centre API docs (template list and generate-file-from-template), checked on Nov 20, 2026.
Dageno team, Do AI answers that cite external pages also name brands? https://www.dageno.ai/research/ai-citations-vs-brand-mentions
Dageno team, Which websites do AI answers cite most? ChatGPT and Google's AI don't cite the same ones. https://www.dageno.ai/research/ai-most-cited-websites-chatgpt-vs-google
Dageno team, Ask AI the same question again the same day: are the recommended brands the same? https://www.dageno.ai/research/ai-same-day-repeat-queries

Items pending confirmation (preview only; can be removed before publishing)

Pink highlights in the body mark chart values, calculated figures, items flagged for backend confirmation, and stated inferences. The author byline uses the default "Dageno team". CTA wording should be reconfirmed before publishing.

  • Chart values: source types 74.88% / 24.09% / 1.03%; esign.ai at 0.26% of all answers.
  • Calculated: 1.4 / 0.06 ≈ 23x; US answers 52.4%; 0.26% × 66,866 ≈ 174 (article uses 175).
  • Backend: platform-level mentions add up to 98, 1 short of 99. The collection end date follows the report body (Nov 17), not the filename (Nov 18).
  • The report's claim that API and template pages were "not found on the official site" does not match the site as opened on Nov 20, 2026. This article corrects it and does not use it as a finding. There is no separate data for Hong Kong or Southeast Asia, and no conclusions are drawn for those markets.
  • Inference: Google AI Mode cites pages more often and may therefore open esign.ai more often. Citations of esign.ai are not broken down by platform, so this cannot be confirmed.