Enspyre 安石國際 AEO Assessment Report
by Novastacks AI
enspyre.com | Taiwan Market
July 17, 2026 | Prepared by Novastacks AI
Site Readiness: 7.3 · LLM Visibility: 5.2
Enspyre Is #1 on Google for Every Question We Tested — ChatGPT Names Them on Three of Five
Enspyre 安石國際 has spent 24 years building Taiwan's most focused B2B lead-generation and phone-answering practice. It shows in search: you hold the #1 Google result on all five buyer questions we tested.
Ask ChatGPT those same five questions and your name appears in three. On the two worth the most — the buyer asking who does telemarketing outsourcing, and the one asking who answers their missed calls — ChatGPT recommends a telecom giant, a competitor whose site hasn't published anything since 2013, and a firm you outrank on Google. You win the search. Someone else wins the answer.
The gap isn't reach. Your articles are already quoted in 11 Google AI answers — the machines read you fine. They just never reach for the pages that sell. In a category where an AI answer now sits above every result you own, that is the whole contest.
| Domain | Ranked Keywords | Est. Traffic (ETV) | #1 Positions | #2-3 | #4-10 |
|---|---|---|---|---|---|
| enspyre.com | 35 | 275 | 3 | 1 | 5 |
| tpemarketing.com | No data available | No data available | No data available | No data available | No data available |
| chainsea.com.tw | 1,117 | 16,764 | 7 | 41 | 223 |
You're #1 on Google for Phone Answering. ChatGPT Answers That Question Without You.
Six prompts tested on ChatGPT and Google in Traditional Chinese from Taiwan — plus the 11 AI answers already quoting your blog
ChatGPT Query Results
| Prompt Type | Query | Mentioned? | Who Was Cited |
|---|---|---|---|
| Branded | 安石國際 Enspyre 是一家什麼樣的公司？他們提供哪些服務？在台灣的評價如何？ | Yes | Named and described accurately — 2002 founding, correct legal name, correct service lines. Cited enspyre.com's homepage, /about-us/ and /digital-marketing/, plus the government business registry and 台灣就業通, a government job portal. |
| Competitor-Branded | 程曦資訊 Chainsea 是一家什麼樣的公司？他們提供哪些客服委外服務？ | No | Chainsea and its affiliate chainsky.com.tw only. Enspyre absent — which is the correct behaviour for a competitor-branded question. This row is diagnostic and is excluded from every score in this report. |
| Category | 台灣有哪些推薦的 B2B 商機開發外包公司？我想把業務開發的工作外包出去，該找哪幾家？ | Yes | Enspyre listed 4th, alongside Salesworks. But the answer was sourced from callbox.com.sg, ensun.io and ycp.com — foreign aggregators. Not one Enspyre page was cited. |
| Category | 台灣的電話行銷外包公司有哪些推薦？中小企業想找專業團隊做電話開發，有什麼選擇？ | No | 頂準創新行銷顧問, 台灣達唯諮 and 程曦資訊 recommended. Enspyre never named. The source cited for the top recommendation was tpemarketing.tw66.com.tw — a third-party directory listing. |
| Comparison | Enspyre 安石國際、頂準創新行銷 (tpemarketing)、程曦資訊 (Chainsea) 這三家公司有什麼差別？做 B2B 業務開發外包的話，哪一家比較適合？ | Yes | Enspyre rated ★★★★★ for B2B business development against 頂準 ★★★★☆ and 程曦 ★★☆☆☆, citing enspyre.com/b2b-lead-generation-service/. The only independent source in the answer was a competitor's directory profile. |
| Long-Tail / Transactional | 我的公司人手不足常漏接客戶來電，想找專業的電話代接、電話祕書代接服務，台灣有哪些推薦的廠商？費用大概多少？ | No | 中華電信 (cht.com.tw) and MOT (mot-net.tw) recommended, with MOT's price list quoted down to NT$525-2,100/month. Enspyre never named — on the exact question where it holds #1 on Google with a purpose-built service page. |
All six prompts triggered live web search, so these are real retrieval results, not model memory. Of the five scored prompts (the competitor-branded one is diagnostic and excluded), ChatGPT named Enspyre on three and cited an Enspyre page as a source on two — the branded question, and the comparison that already contains the company's name. On the three questions a buyer asks when they do not yet know who Enspyre is, the model cited zero Enspyre URLs. The pattern is precise: ChatGPT knows this brand well and describes it accurately, but it does not reach for Enspyre when the question is about the category rather than the company. That is a citation problem, not an awareness problem — and the sample is six prompts, so read it as a signal rather than a rate.
Google AI Overview Results
| Query Type | Query | AIO Triggered? | Prospect Rank | Top Results |
|---|---|---|---|---|
| Branded | 安石國際 Enspyre 是一家什麼樣的公司？ | Yes | #1 | Enspyre homepage #1 and its lead-generation category page #5. Every other page-1 slot is a third-party profile OF Enspyre: 104人力銀行 (5.0★, 2 votes), Facebook (2,150+ followers), an AI TAIWAN exhibitor page, Yourator, Qollie, LinkedIn. Neither competitor appears. Related searches: 安石國際評價, 安石國際ptt, 安石國際dcard. |
| Competitor-Branded | 程曦資訊 Chainsea 是一家什麼樣的公司？ | No | Absent (expected — diagnostic) | chainsea.com.tw #1. The only query of six where no AI Overview fired — Google served a featured answer from ict.careernet.org.tw, a third-party directory, instead. Even Chainsea's own 'what do they do' answer box is served from a trade-association magazine, not from Chainsea. |
| Category | 台灣有哪些推薦的 B2B 商機開發外包公司？ | Yes | #1 | Enspyre #1 — with a blog post dated 2020-10-10, not a service page. Neither competitor ranks. The rest of page 1 is diluted by generic outsourcing marketplaces, job boards, and freelance threads on Reddit and PTT. |
| Category | 台灣的電話行銷外包公司有哪些推薦？ | Yes | #1 | Enspyre #1 (blog post, 2025-07-09); chainsea.com.tw #3 with an on-topic column page. Chainsea also owns Google's answer box for 「如何做好電話行銷？」 with a second column piece. The '2026年十大台灣行銷公司推薦排名清單' listicle at nabi.104.com.tw does not include Enspyre. |
| Comparison | Enspyre、頂準創新行銷、程曦資訊 這三家公司有什麼差別？ | Yes | #1 | Enspyre holds FOUR page-1 organic slots (#1, #3, #5, #9). Neither named competitor ranks at all, despite both being named in the query. meet.bnext.com.tw, which describes Enspyre's service by name, sits just below page 1. |
| Long-Tail / Transactional | 想找專業的電話代接、電話祕書代接服務，台灣有哪些推薦的廠商？費用大概多少？ | Yes | #1 | Enspyre #1 with its actual service page, /phone-answering-service/ — the best asset-to-question match in the whole audit. Google quotes the value proposition verbatim: 「系統化來電進線處理，完美解決漏接與人力問題」. Neither competitor ranks. ChatGPT answers this same question without ever naming Enspyre. |
An AI Overview fires above five of the six queries we ran, and above all three core commercial keywords in this category — Google is now synthesizing an answer above every result Enspyre owns. Underneath that answer, Enspyre is #1 organic on all five scored queries. That is an exceptional position and it deserves to be said plainly. One honest limitation: Taiwan AI Overviews load after the initial page response, so we could observe that the AI answer fired but not which sources it named. We therefore cannot tell you whether Enspyre sits inside those specific AI Overviews, and we will not guess. What we can tell you comes from a separate record of live AI answers: enspyre.com has been quoted in 11 Google AI Overviews, several of them using exact text fragments lifted straight out of /blog/leads/ and /blog/b2b-marketing-6-steps-guide/. Every one of those 11 points at a blog URL. Not one points at a service page.
Citation Surface Analysis
| Platform | Presence | Strength | Notable |
|---|---|---|---|
| Yes | 2,150+ followers · 5.0★ (3 votes) | Your strongest third-party asset by a wide margin. Posted 2 weeks before this audit; genuine buyer-facing case-study video (翱翔智慧, 梭易科), AmCham interviews, InnoVEX and Meet Taipei coverage, a TaiwanPlus founder interview. ChatGPT cited facebook.com ZERO times across all six prompts. | |
| Yes | 1,010+ followers | The platform where B2B buyers actually live. Ranks on your branded page 1. Cited by ChatGPT zero times in six prompts. | |
| 104人力銀行 | Yes | 5.0★ (2 votes) | A recruitment profile that outranks everything except your own site on brand searches, and is quoted twice inside Google's AI Overviews for this category. It also holds a page-1 slot on your core commercial keyword — with a job ad. |
| YouTube | Yes | 40+ subscribers | Real but negligible weight. The most-watched asset in the channel's history (3,830+ views) is a 2013 recruitment video. Recent buyer-facing case-study shorts (2025-08, 2025-09) earn 50-240 views. youtube.com is co-cited twice in this category's AI answers. |
| PTT | Yes | 10 of 10 results are about employment | Reputation risk. All 10 brand results concern working at Enspyre; zero discuss hiring Enspyre. The #1 result on Taiwan's most influential forum is a 2010 post tagged 「黑名」 alleging 90% part-time staff without labour insurance. It is sixteen years old and it still ranks first. |
| Dcard | No | 0 genuine brand matches of 10 | A named gap. Buyers search '安石國際dcard', Google offers it as a related search, and dcard.tw is co-cited twice in this category's AI answers — and there is nothing about Enspyre there. All 10 results are collisions on the generic token 安石. |
| Mobile01 | No | 0 genuine brand matches of 10 | Checked for completeness of the Taiwan forum set. A consumer-electronics board; not a realistic surface for B2B outsourcing. All matches were token collisions (王安石, 安石牧場, '5000毫安石'). |
Two things stand out. First, the best asset is the least cited: Facebook and LinkedIn together carry 3,160+ followers, a 5.0★ rating and a steady stream of genuine buyer-facing case-study video — and ChatGPT cited neither, once, across six prompts. Second, the footprint is recruitment-shaped, not buyer-shaped: 104人力銀行, Yourator, Qollie and 1111人力銀行 dominate the brand's third-party surface, and 10 of 10 PTT threads discuss Enspyre as an employer. When a model researches this company, it disproportionately reads job ads.
Citation Source Landscape: Who AI Cites Instead
Across six ChatGPT prompts, not one non-competitor third-party domain was cited more than once. There is no incumbent set of publications this model trusts for this category in Taiwan. Nobody owns the citation surface — which means it is available.
| Domain | Type | Appears For | You Present? | Opportunity |
|---|---|---|---|---|
| cht.com.tw | Telecom incumbent (中華電信) | Phone answering / 電話代接 | No | A different promise entirely — an automated service against your human team. That comparison is unwritten, and you are the only party who can write it. |
| mot-net.tw | Cloud PBX vendor | Phone answering + pricing | No | ChatGPT quotes their published tiers (NT$525-2,100/month) to answer a question you rank #1 for. Published prices get quoted. |
| tpemarketing.tw66.com.tw | Business directory listing | Telemarketing outsourcing | No | A directory profile for a competitor whose own site has no tracked keywords and hasn't published since 2013 — and it beats you into the answer. The clearest proof in this audit that AI visibility is won off-domain. |
| darwinz.tw | Category competitor (台灣達唯諮) | Telemarketing outsourcing | No | Named ahead of you on your own category question. |
| callbox.com.sg · ensun.io · ycp.com | Foreign aggregators / vendor lists | B2B lead-generation outsourcing | Named, not cited | ChatGPT lists you 4th but sources the answer from foreign directories rather than your own site. Directory inclusion is a controllable input. |
| chainsea.com.tw/column/ | Competitor editorial | 如何做好電話行銷 | No | Owns Google's answer box for a how-to question inside your category, using a page type you already publish. |
| cheerway.tw | Independent publisher (智匯家) | Lead lists, B2B marketing | Co-cited 4× beside you | The warmest citation route found. Already inside four of the eleven AI answers you are already inside. Not a competitor. |
| web.cheers.com.tw | Independent publisher (Cheers) | Telemarketing careers, cold outreach | Co-cited 4× beside you | Second-warmest. Occupies the same AI answers you do, four times over. Not a competitor. |
You Cover Every Page Type Chainsea Does — and 91.9% of Its Traffic Comes From Excel Tutorials
Page-type coverage across all three domains, 14 pages analyzed
| Page Type | Enspyre 安石國際 | tpemarketing.com | chainsea.com.tw |
|---|---|---|---|
| Homepage | ✓ 3,971 chars · 32 H2s | ✗ 247 chars · no meta | ⚠ 880 chars · 2 H2s |
| Service / Money Pages | ✓ 3 lines · 3,827-5,352 chars | ✗ 252 chars | ✓ 3,079 chars + Service markup |
| Case Studies | ✓ 3 hubs · named clients | ⚠ 1 page · 62 words | ✓ /showcase/ |
| Blog / Informational | ✓ ~260 posts · on-category | ✗ ~26 posts · last one 2013 | ✓ ~165 posts · weekly |
| FAQ / Help | ⚠ Content only, unmarked | ⚠ Q&A page · 74 words | ✗ None |
| Video Hub (on-site) | ✗ None | ✗ None | ✓ Drives 91.9% of its traffic |
| Tools / Interactive | ✗ None | ✗ None | ✓ fastbot chatbot · /webchat/ |
| Multilingual (zh-TW / EN) | ✓ Full /en/ tree | ✗ None | ✗ None |
Your Site Is the Fastest and Best-Built of the Three. It Also Carries Zero FAQ Markup.
Implementation quality across 14 pages and three domains
Rendering & Bot Access: Nothing Is Blocking the Machines
Bright SpotEvery check that could explain an AI visibility gap comes back clean. enspyre.com is server-rendered WordPress — the full text of every page is present in the raw HTML before a single line of JavaScript runs, so any crawler that requests a page receives the content. Google's own page-quality audit scores it 96 performance, 100 SEO, 84 accessibility. FCP 0.9 s, LCP 1.2 s, CLS 0.005, Speed Index 1.1 s, zero blocking time, and the root document responds in 30 ms. Language targeting (zh-Hant / en / x-default) and canonical URLs are correct on every page we analyzed. On this evidence no rendering cap applies to the score.
The proof is not theoretical. Google's AI Overviews have quoted enspyre.com in 11 recorded answers, several of them using exact text fragments lifted straight out of /blog/leads/ and /blog/b2b-marketing-6-steps-guide/. Machines can reach this site, read it, and quote it verbatim — which independently confirms what Enspyre already suspected about getting leads from AI. Whatever is limiting AI visibility here, it is not access. That single fact reframes the entire engagement: this is not a repair job, it is a targeting job.
One structural gap is worth recording. None of the five pages we analyzed declares a <main> landmark, and four of five declare no <section>. Extraction engines use those landmarks to separate the article from the navigation and the footer. On a Divi build that wraps every element in layout containers, drawing that boundary explicitly is cheap — and it is the only technical deduction in this finding.
| Signal | enspyre.com | tpemarketing.com | chainsea.com.tw |
|---|---|---|---|
| Rendering method | SSR (WordPress + Divi) | Static (Classic ASP) | SSR (WordPress + Elementor) |
| Bot-accessible HTML | ✓ | ✓ | ✓ |
| Performance score (Google page-quality audit) | 96 | 88 | 58 |
| SEO score (Google page-quality audit) | 100 | 83 | 85 |
| Accessibility score (Google page-quality audit) | 84 | 57 | 81 |
| FCP | 0.9 s | 0.8 s | 3.2 s |
| LCP | 1.2 s | 1.1 s | 5.1 s |
| CLS | 0.005 | 0.093 | 0.038 |
| Server response | 30 ms | 510 ms | 1,270 ms |
| Canonical URL | ✓ all 5 pages | ✗ none | ✓ all 4 pages |
| Language targeting | ✓ zh-Hant / en / x-default | ✗ none | ✗ none |
| <main> landmark | ✗ 0 of 5 | ✗ 0 of 5 | ⚠ 1 of 4 |
Zero FAQ Markup on All 14 Pages, on All Three Domains
HighAcross the 14 pages analyzed on the three domains, not one carries FAQ markup. Not Enspyre, not Chainsea, not TPE Marketing. The whole category is unclaimed.
That matters most on /phone-answering-service/, which already contains genuine question-and-answer content — the page itself advertises a 「客製問答FAQ」 feature — but publishes it as ordinary prose. This is the page holding #1 on Google for exactly the question a buyer asks when they keep missing calls. It is also the page ChatGPT answers by recommending 中華電信 and MOT, without ever saying the word Enspyre. The answer is already written on the page. Nothing on it tells a machine that it is an answer.
Two further gaps are competitive rather than categorical. Chainsea's service page declares Service markup and its homepage declares LocalBusiness markup; enspyre.com declares neither on any of its five pages, despite selling three named service lines from a Taipei address. On every other markup type — Organization, WebSite, SearchAction, BreadcrumbList, and Article on the blog — Enspyre matches or beats both competitors, and TPE Marketing carries no structured data at all on any page.
| Markup type | enspyre.com | tpemarketing.com | chainsea.com.tw | What it unlocks |
|---|---|---|---|---|
| Organization | ✓ 5 of 5 pages | ✗ | ⚠ service page only | Entity identity — who the company is |
| WebSite + SearchAction | ✓ 5 of 5 | ✗ | ✓ 4 of 4 | Sitelinks search box |
| BreadcrumbList | ✓ 5 of 5 | ✗ | ✓ 4 of 4 | Page hierarchy for crawlers |
| Article / BlogPosting | ✓ on the blog | ✗ | ✗ none | Authorship and freshness on editorial |
| Service | ✗ 0 of 5 | ✗ | ✓ service page | Machine-readable statement of what you sell |
| LocalBusiness | ✗ 0 of 5 | ✗ | ✓ homepage | Taipei address, hours, contact |
| FAQPage | ✗ 0 of 5 | ✗ 0 of 5 | ✗ 0 of 4 | Unclaimed by the entire category |
The Page Type Earning All 11 of Your AI Citations Carries No H1
MediumFour of the five enspyre.com pages we analyzed declare a clean H1 — the homepage ('Accelerate Success.'), the lead-generation service page, the phone-answering page and the case-study hub. The exception is the blog post /blog/outsourcing-b2b-cold-outreach/, currently #1 on Google for the telemarketing-outsourcing question, which declares no H1 at all. Its two H2s carry the entire hierarchy.
That would be a footnote on most sites. It is not one here, because every one of the 11 Google AI Overview citations enspyre.com has earned points at a blog URL. The page type doing all of this company's AI visibility work is the one with the weakest heading signal. We analyzed one representative post, so treat this as a template question to confirm across the ~260-post library rather than a proven site-wide count.
Context, firmly in Enspyre's favour: neither competitor declares an H1 on any page we analyzed, and tpemarketing.com serves the same title tag on all five pages with no meta description, no canonical tag and no social preview image anywhere. On title tags, descriptions, language targeting, canonicals and heading density, enspyre.com is the best-implemented domain of the three by a wide margin. This is a narrow correction on a strong build, not a rebuild.
| Signal | enspyre.com | tpemarketing.com | chainsea.com.tw |
|---|---|---|---|
| Pages with an H1 | 4 of 5 | 0 of 5 | 0 of 4 |
| H2s — homepage | 32 | 0 | 2 |
| H2s — main service page | 37 | 0 | 2 |
| H2s — blog post | 2 | 0 | 7 |
| Unique title tag per page | ✓ 5 of 5 | ✗ same title on all 5 | ✓ 4 of 4 |
| Meta description | ✓ 5 of 5 | ✗ 0 of 5 | ✓ 4 of 4 |
| Social preview image | ✓ 5 of 5 | ✗ 0 of 5 | ✓ 4 of 4 |
| Visible breadcrumbs | ✗ | ✗ | ✗ |
The Answers Buyers Want Are on Your Blog. The Pages That Sell Don't Have Them.
Content depth, composition and question coverage across page types
Four Times More Text Than Chainsea's Homepage — Wrapped Around 87 Images
MediumThe volume story runs in Enspyre's favour, and it is worth stating plainly because it changes what the fix is. enspyre.com's homepage carries 3,971 characters of extractable text against Chainsea's 880 and TPE Marketing's 247 — 4.5x and 16x respectively. The phone-answering page carries 5,352 characters and 45 H2 headings; Chainsea's equivalent service page carries 3,079 characters and 2 H2s. On text depth and heading density, Enspyre leads both competitors on every page type we measured. This is not a thin site.
The cost is composition. The homepage carries 87 images against those 3,971 characters — roughly one image every 46 characters — and the phone-answering page carries 123. Images are an asset: they drive Google Image search, Discover and visual results, and they are a large part of why this site converts human visitors. But an image only tells a machine what it depicts through the text around it: its alt text, its caption, the heading it sits under. At one image per 46 characters there is not a great deal of surrounding text doing that work. The blog inverts the ratio entirely — 11 images to 2,080 characters. And the blog is the only page type Google's AI answers ever quote.
One measurement correction, recorded because a naive crawl gets it backwards: enspyre.com emits absolute URLs for its own internal links, which most automated tools file as external. The homepage's real internal link count is 71, not 0. Internal linking is not a problem on this site, and any report that tells you otherwise has misread its own tooling.
| Page type | Text (chars) | Images | Chars per image | H2s | Internal links |
|---|---|---|---|---|---|
| enspyre.com — homepage | 3,971 | 87 | 46 | 32 | 71 |
| enspyre.com — phone answering | 5,352 | 123 | 44 | 45 | 51 |
| enspyre.com — lead generation | 3,827 | 90 | 43 | 37 | 46 |
| enspyre.com — case studies | 4,415 | 82 | 54 | 28 | 88 |
| enspyre.com — blog post | 2,080 | 11 | 189 | 2 | 42 |
| chainsea.com.tw — homepage | 880 | 13 | 68 | 2 | 62 |
| chainsea.com.tw — service (BPO) | 3,079 | 14 | 220 | 2 | 57 |
| chainsea.com.tw — column | 2,283 | 12 | 190 | 7 | 106 |
| tpemarketing.com — homepage | 247 | 16 | 15 | 0 | 17 |
| tpemarketing.com — service | 252 | 20 | 13 | 0 | 17 |
First-Mover Opportunity: The Q&A Is Already Written, on the Wrong Page Type
Competitive IntelOf the 14 pages analyzed across all three domains, exactly one contains real question-and-answer content: enspyre.com/phone-answering-service/. It is the page holding #1 on Google for the phone-answering question. It is the page whose value proposition Google quotes verbatim in its own snippet — 「系統化來電進線處理，完美解決漏接與人力問題」. And it is the page ChatGPT answers by recommending 中華電信 and MOT, without ever naming Enspyre.
TPE Marketing has a Q&A page carrying 74 words of extractable text and no structure. Chainsea has no Q&A surface at all. Nobody in this category has marked up a single question. That is unusual — and it is the textbook definition of an uncontested opening. The first domain here to publish machine-readable answers to the questions buyers actually type — 「電話代接多少錢？」, 「外包電話行銷值不值得？」, 「B2B 商機開發外包該找誰？」 — does not have to displace an incumbent, because there isn't one. Every AI answer in this category is currently being assembled from sources that are not experts in it.
The raw material already exists, which is what makes this the highest-leverage item in the audit rather than a content programme. Enspyre's blog answers these questions at length and in depth — that is precisely why Google's AI Overviews quote it 11 times, reaching for a blog URL every single time and for a service page never. The content is written. The expertise is real and 24 years deep. It is simply sitting on the wrong page type, in the wrong shape, one step removed from where the buying decision happens.
| Question-content signal | enspyre.com | tpemarketing.com | chainsea.com.tw |
|---|---|---|---|
| Pages with real Q&A content | 1 of 5 (phone answering) | 1 of 5 (74 words) | 0 of 4 |
| Pages with FAQ markup | 0 of 5 | 0 of 5 | 0 of 4 |
| Owns a Google answer box in-category | Not observed | Not observed | ✓ 「如何做好電話行銷？」 |
| Quoted by Google's AI answers | 11 times — blog URLs only | Not observed | Not observed |
| On-category library depth | ~260 posts | ~26 posts, last 2013-04-03 | ~165 posts, weekly |
AI Researching Enspyre Reads Job Ads: Recruitment Profiles Own Your Branded Page One
Positioning signals across category search, brand search and the competitive footprint
On the Three Short Keywords That Convert, You're #1, #4 and #3 — Only One Is Held by a Service Page
MediumWe tested three core commercial keywords derived from Enspyre's own homepage self-description. Enspyre appears on page one of all three: #1 organic for 「B2B 商機開發 外包」, #4 for 「電話行銷 外包 服務」, #3 for 「電話祕書 代接電話 服務」. This is emphatically not a domain with a ranking problem, and any proposal that starts there is solving the wrong thing.
What repays examination is which asset does the ranking. The #1 position on the B2B outsourcing keyword is held by a blog post published 2020-10-10 — six years old. The 「電話行銷 外包 服務」 position is held by a 2025 blog post, and on that keyword Chainsea takes #1 with a category column page, then also takes Google's answer box for 「如何做好電話行銷？」 with a second column piece. Only the phone-answering keyword is won by an actual service page. Two of three core-keyword positions are carried by editorial that was never built to sell.
Two page-one slots on the B2B keyword are worth naming. A 1111人力銀行 job posting for Enspyre sits at #3 organic, and a directory profile for TPE Marketing at #4 — outranking TPE Marketing's own website, which does not appear on the page at all. Third parties are ranking on Enspyre's core category with pages Enspyre does not control and cannot edit.
An AI Overview fires at position zero on all three. Every one of these hard-won rankings now sits underneath a synthesized answer.
| Core keyword | Enspyre rank | Asset doing the ranking | Date | Competitors on page 1 |
|---|---|---|---|---|
| B2B 商機開發 外包 | #1 (and #7) | Blog post — 「業務開發外包」到底行不行？ | 2020-10-10 | tpemarketing via a directory profile at #4; its own domain absent |
| 電話行銷 外包 服務 | #4 | Blog post — 新型態電話陌生開發服務 | 2025-07-09 | chainsea #1 (column page); tpemarketing #9 (own domain) |
| 電話祕書 代接電話 服務 | #3 | Service page — /phone-answering-service/ | — | Neither competitor in the top 20; contested by phone-system vendors and virtual-office providers |
Your Off-Domain Footprint Sells Jobs at Enspyre, Not Enspyre
HighEnspyre sells B2B lead generation. Its off-domain footprint sells jobs at Enspyre.
On the branded search we ran, page one below enspyre.com's own two results reads: 104人力銀行 (a recruitment profile, 5.0★ from 2 votes), Facebook, an AI TAIWAN exhibitor page, Yourator (a job board), Qollie (an employee-review site) and LinkedIn. When we asked ChatGPT to describe the company, it reached for the government business registry and 台灣就業通 — a government job portal. On PTT, all 10 brand results concern working at Enspyre; not one discusses hiring Enspyre. The most-watched video the company has ever published — 3,830+ views — is a 2013 recruitment reel, and its 2025 buyer-facing case-study shorts earn 50 to 240 views apiece.
This has already propagated into the models. 104人力銀行 is co-cited twice inside Google's AI Overviews for this category, and its Enspyre profile is the source of the line an AI is most likely to repeat about the company: 「自2002年以來，Enspyre已為甲骨文(Oracle)、樂天、eBay等數千家全球頂尖公司提供商機開發行銷服務」. That sentence is excellent — Oracle, Rakuten and eBay as named clients is a serious credential. It is also on a job board, written for candidates. A recruitment profile is currently one of the strongest brand-defining assets Enspyre owns inside an AI's field of view, and Enspyre does not control the page.
One consequence is a live liability. The #1 PTT result for the brand — on Taiwan's most influential forum, on that same employment surface — is a 2010 post tagged 「黑名」 alleging that 90% of staff were part-timers with no labour insurance. It is sixteen years old and it still ranks first. Google's related searches show buyers looking for exactly this: 安石國際評價, 安石國際ptt, 安石國際dcard, Enspyre 的評論. Aggregate sentiment across 94 third-party mentions is fine — 10 positive against 3 negative, no sentiment problem at all. The problem is not the average. It is which one ranks.
| Branded page-1 slot | Domain | What it is | Audience it speaks to |
|---|---|---|---|
| 1 | enspyre.com | Homepage | Buyer |
| 2 | 104.com.tw | Company recruitment profile · 5.0★ (2 votes) | Candidate |
| 3 | facebook.com | Brand page · 2,150+ followers · 5.0★ | Buyer |
| 4 | ai-taiwan.com.tw | 2026 AI TAIWAN exhibitor page | Buyer |
| 5 | enspyre.com | Lead-generation category page | Buyer |
| 6 | yourator.co | Job board listing | Candidate |
| 7 | qollie.com | Employee review site | Candidate |
| 8 | linkedin.com | Company page · 1,010+ followers | Buyer |
Chainsea's 16,764 Monthly Visits Look Like a Moat. 91.9% of Them Are Gamma and Excel Tutorials.
Competitive IntelRead the headline numbers and Chainsea looks unassailable: 1,117 tracked keywords against Enspyre's 35, and roughly 16,764 estimated monthly organic visits against 275 — a 61x gap. Those numbers are real and we report them as measured.
They are also almost entirely off-category. We bucketed Chainsea's top 150 keywords by the section of the site that earns them. 91.9% of that traffic — 15,019 of 16,341 estimated visits — lands on /video/, a library of software tutorials ranking for 「gamma」 (135,000 monthly searches; that one keyword alone is 54% of Chainsea's entire top-150 traffic), 「dreamina」, 「excel 公式」 and 「word行距調整」. Its /column/ section — the only part of Chainsea that competes with Enspyre on subject matter — accounts for 0.2%: 31 estimated visits, from a single keyword.
So the honest read is this. Nobody is winning this category on search volume. Chainsea is winning an unrelated office-software tutorial niche that happens to sit on the same domain as a customer-service outsourcing business. Meanwhile TPE Marketing has no whole-domain organic footprint on record at all and has published nothing since 2013 — and ChatGPT still names it ahead of Enspyre for telemarketing outsourcing, via a third-party directory profile.
Neither the biggest keyword footprint nor the deadest website predicts who gets named in the AI answer. That is the finding, and it is the single most important sentence in this report for anyone deciding where to spend next.
| chainsea.com.tw section | Est. monthly visits | Share of top-150 | Keywords | What it actually ranks for |
|---|---|---|---|---|
| /video/ | 15,019 | 91.9% | 136 | gamma, gamma 簡報, dreamina, excel 公式, word行距調整 — office-software tutorials |
| / (homepage, mostly branded) | 1,052 | 6.4% | 7 | Brand terms |
| /about/ | 141 | 0.9% | 4 | Company information |
| /webchat/ | 75 | 0.5% | 1 | Chatbot product |
| /column/ — the on-category editorial | 31 | 0.2% | 1 | The only section competing with Enspyre |
| /news_center/ | 23 | 0.1% | 1 | Press releases |
Bright Spots: Two Publishers Already Sit Beside You Inside Google's AI Answers
Bright SpotThree assets are already doing real work and can be leveraged rather than built from nothing. This is why the roadmap that follows is six months and not two years.
Two independent publishers already share your AI answers. Across the 11 Google AI Overviews that cite enspyre.com, two domains appear alongside you four times each: cheerway.tw (智匯家) and web.cheers.com.tw (Cheers 快樂工作人). Neither is a competitor. They are already inside the answers you are already inside — the warmest citation routes this audit found anywhere.
Business media already describes your service by name. meet.bnext.com.tw (Meet 創業小聚) ranks #6 organic on 「B2B 商機開發 外包」 and #2 on 「電話行銷 外包 服務」 with articles that name you directly: 「安石的B2B 商機開發連結服務就可以解決目前中小企業及SOHO 族電話行銷上的問題」. That is earned media ranking above you, on your own category, saying precisely what you would want it to say. It is a relationship, not a cold pitch.
ChatGPT already prefers you when the buyer knows your name. Asked to compare Enspyre, TPE Marketing and Chainsea, ChatGPT rated Enspyre ★★★★★ for B2B business-development capability against ★★★★☆ and ★★☆☆☆, and cited your own service page to justify it. On Google, that same comparison question returns four Enspyre results on page one and zero from either named competitor. The judgment is already in your favour. It only needs to be reachable at the moment the buyer does not yet know your name — which is the entire distance between 3 of 5 and 5 of 5.
| Asset | Measured strength | Why it is leverage |
|---|---|---|
| cheerway.tw (智匯家) | Co-cited 4× beside you | Most co-cited independent publisher in your AI-answer neighbourhood. Not a competitor. |
| web.cheers.com.tw (Cheers) | Co-cited 4× beside you | Joint-most co-cited. Already occupies the answers you occupy. |
| meet.bnext.com.tw | #6 and #2 organic, names you | Earned media that already quotes your service by name and outranks you on your own category. |
| 2,150+ followers · 5.0★ · active | Real buyer-facing case-study video already produced. Uncited by ChatGPT — an unconverted asset, not a missing one. | |
| ChatGPT head-to-head verdict | ★★★★★ vs ★★★★☆ / ★★☆☆☆ | The model already rates you first in your category when asked directly. |
| Third-party sentiment | 10 positive · 3 negative (of 94) | No sentiment problem to repair. Only a ranking problem, on one 2010 post. |
From 6.0 to 7.6: The Rankings Are Already Won. The Answers Are Not.
Three horizons, ordered by what moves AI visibility — not by what is easiest to file
Site Readiness Score
LLM Visibility Score
Horizon 1: Make the Answers You Already Wrote Machine-Readable
0-30 days+1.0 to +1.4 Site Readiness (7.3 → 8.3-8.7). Structured data 6.5 → 9.0. Low effort, all of it on an already-strong build.
Mark up the question-and-answer content already live on /phone-answering-service/ — the opening named in Section 05, 'First-Mover Opportunity: The Q&A Is Already Written, on the Wrong Page Type'. No domain in this category has done it.
Add Service and LocalBusiness markup across the three service lines — the two competitive gaps named in Section 04, 'Zero FAQ Markup on All 14 Pages', where Chainsea currently leads.
Restore the H1 on the blog template and audit it across the ~260-post library — Section 04, 'The Page Type Earning All 11 of Your AI Citations Carries No H1'.
Declare main landmarks so extraction engines can separate the article from the navigation — the single technical deduction in 'Rendering & Bot Access: Nothing Is Blocking the Machines'.
Horizon 2: Move the Answers Onto the Pages That Sell
30-90 days+0.2 Site Readiness (→ 8.5) and +1.0 to +1.5 LLM Visibility. Core keyword visibility 3.5 → 5.5. This is where the two absent ChatGPT answers get contested.
Rebuild the two money pages around the exact questions ChatGPT currently answers without you — 'who does telemarketing outsourcing' and 'who answers my missed calls, and what does it cost'. Both transcripts are in Section 02.
Publish the pricing and selection content the models reach elsewhere for: ChatGPT quotes MOT's NT$525-2,100/month tiers to answer a question you hold #1 for on Google.
Refresh the 2020 blog post still carrying your #1 position on 'B2B 商機開發 外包' — named in 'On the Three Short Keywords That Convert'. Chainsea publishes weekly; your best core asset is six years old.
Raise the text-to-image ratio on the money pages — 'Four Times More Text Than Chainsea's Homepage, Wrapped Around 87 Images' in Section 05. Keep the images; give them text to stand on.
Horizon 3: Reshape the Off-Domain Surface From Recruitment to Buyer
3-6 months+1.3 to +1.8 LLM Visibility (→ 7.0). Third-party review presence 3.0 → 5.5; platform citation surface 6.0 → 7.5. Combined 6.0 → 7.6.
Pursue the two publishers already co-cited beside you four times each — cheerway.tw and web.cheers.com.tw, named in 'Bright Spots'. Warm routes, not cold outreach.
Extend the meet.bnext.com.tw relationship, which already ranks above you on your own category saying what you would want it to say.
Build buyer-side presence on Dcard and PTT: both are quoted in this category's AI answers, both carry live branded search demand, and neither currently carries anything about Enspyre as a vendor — see 'Your Off-Domain Footprint Sells Jobs at Enspyre, Not Enspyre'.
Address the 2010 PTT post that still ranks #1 for the brand, against live buyer demand for 安石國際評價 — and get Facebook's existing case-study library onto surfaces the models actually read.
Agentic Marketing Systems, Built by Senior Operators
Novastacks is not an agency selling AI as a buzzword. We are senior marketing operators with decades of experience at Expedia, Tencent, Klook, and Traveloka who built enterprise-grade AI marketing systems from the ground up.
What We Do
- ▸AEO (Answer Engine Optimization) — Get your brand cited when prospects ask ChatGPT, Perplexity, and Google AI about your category
- ▸SEO Integration — Traditional search visibility that compounds with AI visibility
- ▸Custom AI Growth Systems — Agentic workflows, content engines, and data pipelines built for your business
- ▸Fractional Growth Partner — Senior strategic leadership without the full-time overhead
What You Get From Us
Solutions designed by operators who've led growth marketing and SEO at the Director/VP level. Not juniors following playbooks.
AI-powered workflows that move at machine speed. Audits, content, optimization, and reporting that would take a team weeks, delivered in days.
No bloated retainers. Scope of work tailored to your stage, budget, and goals. Start small, scale when you see results.
Ready to Own
the Answer Too?
Book a 30-minute call. We'll walk through the two buyer questions where you hold #1 on Google and go unnamed by ChatGPT, show you the eleven AI answers already quoting your blog, and map what it takes to claim a citation surface that nobody in your category has claimed yet.
Book Discovery Call30 min | Free | No commitment