Novastacks
Talk to Our Team
// AEO ASSESSMENT

Profet AI AEO Assessment Report
by Novastacks AI

profetai.com | Taiwan Market

February 28, 2026 | Prepared by Novastacks AI

4.2 /10
Warning

Site Readiness: 5.1 · LLM Visibility: 3.6

Compared against: chimes.ai aiwin.com.tw Profet AI
Share
Scroll
// 01 EXECUTIVE SUMMARY

Profet AI Knows Its Brand. AI Doesn't Know Its Category.

Profet AI (傑倫智能) holds a defensible niche in Taiwan's manufacturing AI space — ranked #3 for automl (SV: 590), #1 for branded queries, and correctly cited by ChatGPT when asked directly. The problem is what happens without the brand name in the query.

In unprompted category searches — "best manufacturing AI platforms in Taiwan", "no-code AutoML for semiconductor" — Profet AI loses ground. Competitor Chimes AI appears in Google's AI Overview via government startup registry citations. Profet AI does not. In 4 of 6 ChatGPT queries without the brand name, Profet AI was either absent or buried.

Three technical issues compound this: (1) the site is JavaScript-rendered, so AI crawlers receive empty shells instead of content; (2) LCP of 10.8s is the slowest among all three audited domains; and (3) the English subdomain en.profetai.com splits domain authority rather than consolidating it. These are fixable problems with measurable impact on both LLM citation frequency and organic traffic.

The foundation is stronger than the AEO score suggests. BenQ, Maxxis, and industry trade media (iThome, CIO Taiwan, asiamanufacturingreview.com) already cite Profet AI in English. That authority needs to be connected to the right technical infrastructure for LLMs to surface it reliably.

DomainRanked KeywordsEst. Traffic (ETV)#1 Positions#2-3#4-10
profetai.com368913111
chimes.ai9638310
aiwin.com.tw30114000

Keyword rankings and ETV measured via DataForSEO. Location: Taiwan (location_code: 2158). ETV = Estimated Traffic Volume based on position and search volume.

LLM visibility tested via ChatGPT (GPT-4o with web_search enabled) using 6 Traditional Chinese query types. Google AIO tested via DataForSEO SERP with AI Overview extraction. Assessment date: February 28, 2026.

Site Speed Gap

10.8s

Largest Contentful Paint — 3.6× above the 3-second threshold

Chimes AI loads in 3.0s. Profet AI loads in 10.8s. Slow LCP directly suppresses Google rankings and reduces the content AI crawlers can successfully index.

→ Section 04: Site Infrastructure

AI Citation Gap

0 / 4

Non-branded queries where Profet AI appeared in Google AIO or unprompted ChatGPT

ChatGPT cites Profet AI when asked directly. But in open-ended category queries — where most buyer journeys start — Chimes AI appears and Profet AI does not.

→ Section 02: AI Visibility

Crawlability Barrier

JS

Client-side rendering blocks AI bot content extraction

GPTBot, ClaudeBot, and PerplexityBot typically cannot execute JavaScript. The site's content renders client-side, meaning crawlers see an empty shell rather than product descriptions and case study text.

→ Section 04: Site Infrastructure

† Organic keyword and ETV data sourced from DataForSEO Labs (Taiwan, location_code: 2158). ‡ LLM visibility tested via ChatGPT GPT-4o with web_search enabled and Google SERP AI Overview queries using 6 Traditional Chinese query types. Performance measured via DataForSEO Lighthouse. Assessment date: February 28, 2026.

// 02 AI VISIBILITY

Invisible in the Queries That Matter

ChatGPT knows Profet AI by name. Category queries tell a different story.

ChatGPT Query Results

Prompt TypeQueryMentioned?Who Was Cited
Branded告訴我關於Profet AI傑倫智能YesProfet AI cited in detail; sourced from profetai.com, en.profetai.com, en.icaa.org.tw, asiamanufacturingreview.com
Competitor告訴我關於詠鋐智能Chimes AINoChimes AI (詠鋐智能) featured exclusively; no mention of Profet AI
Category台灣製造業AI轉型平台哪家最好Yes8 vendors listed; Profet AI AILM ranked #2 — not the lead recommendation
Technical/AutoML半導體製造業如何用AutoML做品質預測NoGeneric industry answer; no Taiwan vendors cited (training data only — web search disabled)
ComparisonProfet AI vs 詠鋐智能 vs 慧穩科技 哪個好YesProfet AI ranked #1 in comparison table; strongest in ease-of-use and case study depth
Long-tail製造業導入無代碼機器學習需要多久多少成本NoGeneric implementation guidance; no vendor citations (training data only — web search disabled)

Profet AI wins on direct queries but loses on discovery. 2 of 6 queries returned training-data-only answers (web_search disabled) — meaning ChatGPT's knowledge of Profet AI's category is frozen at training time. For technical queries about semiconductor AutoML and implementation cost, Profet AI doesn't exist in ChatGPT's world unless the user already knows the brand name.

Google AI Overview Results

Query TypeQueryAIO Triggered?Prospect RankTop Results
CategoryAutoML無代碼製造業台灣最佳平台YesN/AChimes AI cited via startup.sme.gov.tw; Profet AI absent from AIO sources
Technical製造業AI轉型平台台灣YesN/AIndustry overview articles; no specific vendor citations in AIO
BrandedProfet AI 傑倫智能NoN/AStandard organic results; profetai.com ranks #1 organically
Informationalautoml是什麼 台灣企業NoN/ANo AIO triggered; profetai.com /automl/ ranks #8-9 organically

Google's AI Overview surfaces vendor recommendations based on government registries (startup.sme.gov.tw), G2 listings, and third-party comparison articles — not company homepages. Chimes AI is listed on startup.sme.gov.tw; Profet AI is not. This single citation source gap is why Chimes AI appears in AIO for category queries while Profet AI is absent.

ChatGPT Mentions
3 / 6
Queries where Profet AI was cited
Google AIO Citations
0 / 4
AIO queries where Profet AI appeared
Training Data Blind Spots
2 queries
ChatGPT returned cached answers (no web search)
Category Query Position
#2 of 8
Ranking in ChatGPT category recommendation list

Citation Surface Analysis

PlatformPresenceStrengthNotable
iThome / CIO TaiwanYesStrongMultiple product announcements and AILM coverage — primary domestic B2B citation surface
asiamanufacturingreview.comYesModerateEnglish-language trade media coverage — key for LLM English training data
en.icaa.org.twYesModerateIndustry association listing actively cited by ChatGPT in branded queries
YouTubeYesWeak70 subscribers; product demos exist but minimal engagement or reach
startup.sme.gov.twNoAbsentChimes AI listed here — this is the direct source of their Google AIO citation advantage
G2NoAbsentNo G2 listing; critical gap for B2B SaaS LLM citation infrastructure
DcardYesWeak10+ threads — all intern recruitment, zero product authority or buyer context
PTTNoAbsentTaiwan's primary tech forum; no brand presence detected

Profet AI's citation surface is B2B-appropriate (trade media, industry associations, enterprise case studies) but missing the two platforms that drive AIO and LLM category citations: government startup registries and G2. Chimes AI's AIO advantage comes entirely from a single startup.sme.gov.tw listing. Adding Profet AI there is a 1-week fix with outsized impact.

Full AI Visibility Breakdown

We monitor what ChatGPT, Perplexity, and Google AI say about brands in your category. Our team runs proprietary LLM tracking across every major AI platform — the same systems we deploy for growth-stage companies.

Talk to Our Team to Unlock

30 min | Free | No commitment

// 03 SITE READINESS

Fast Enough for Google; Too Slow for AI

A 92 SEO score hides a 10.8s LCP and JS rendering that blocks AI crawlers.

Site Readiness5.1/10
SignalProfet AIchimes.aiaiwin.com.tw
Title Tag
Meta Description
Structured Data / Schema✗ None detectedUnknownUnknown
LCP (Largest Contentful Paint)✗ 10.8s✓ 3.0s⚠ 5.8s
Content Extractability✗ JS-rendered✓ Static/SSR⚠ Partial
robots.txt (AI crawlers)⚠ Blocks /solutions✓ Open✓ Open
Blog on Main Domain✓ /blog/UnknownUnknown
HTTPS
Canonical TagsUnknownUnknown
Hreflang (zh/en)✓ ImplementedUnknownUnknown

profetai.com scores well on standard SEO signals (title tags, meta descriptions, hreflang, canonicals) — which explains the 92 SEO score in Lighthouse. What the SEO score misses is JavaScript rendering: AI bots that don't execute JS see empty product pages. Adding server-side rendering or prerender middleware for AI user agents is the single highest-ROI technical fix available.

LCP (Page Load)
10.8s
Critical — 3.6× above 3s threshold
Performance Score
57 / 100
Chimes AI: 80. Slowest of 3 audited domains.
SEO Score
92 / 100
Best of 3 — strong foundational SEO
Accessibility
87 / 100
Above average; minor WCAG gaps remain

Site Readiness Analysis

Our senior operators audit every technical signal AI crawlers evaluate — from schema markup to JavaScript rendering to Core Web Vitals. We identify what to fix and in what order.

Talk to Our Team to Unlock

30 min | Free | No commitment

// 04 SITE INFRASTRUCTURE

Three Technical Barriers Blocking AI Indexation

JavaScript rendering, page speed, and robots.txt fragmentation compound each other.

JS-Rendered Pages = Empty Content for AI Crawlers

Critical

profetai.com uses client-side JavaScript rendering (WordPress + JS framework). When AI crawlers — GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot — visit product pages like /ailm/, /solutions/, or case studies like /success-case/maxxis/, they receive an empty HTML shell. The actual content renders after JavaScript executes, which most AI crawlers never trigger.

Evidence: Despite ranking #3 for automl and holding 36 keywords in Taiwan, ChatGPT required the brand name to provide accurate product details. Without the brand name (technical/category queries), ChatGPT either gave generic answers or returned training-data-only responses — consistent with AI crawlers failing to extract page content.

Fix: Implement server-side rendering (SSR) or prerender middleware that serves static HTML to bots based on User-Agent detection. WordPress plugins like "Prerender.io" or a Next.js/Nuxt migration of key product pages are viable paths. Priority pages: /ailm/, /automl/, /domaintwin/, and all /success-case/ pages.

LCP of 10.8s: Slowest Among All Three Audited Domains

High

Largest Contentful Paint of 10.8s places profetai.com in the bottom performance tier. Google's Core Web Vitals threshold is 2.5s for "Good" and 4.0s for "Needs Improvement." At 10.8s, profetai.com is well into "Poor" territory.

DomainLCPPerf ScoreTTI
profetai.com10.8s57/10013.0s
chimes.ai3.0s80/1003.0s
aiwin.com.tw5.8s54/1005.9s

Fix: Identify and defer non-critical JavaScript, enable image lazy loading, add font preloading, leverage browser caching, and consider a CDN for Taiwan-based users. Target: LCP under 4.0s in 30 days.

robots.txt Blocks High-Value Product Pages From Crawlers

Medium

profetai.com's robots.txt disallows /solutions/ and /industry-solutions/ paths — the exact pages that explain Profet AI's product to new visitors and AI crawlers. Additionally, Lighthouse was unable to download profetai.com's robots.txt file entirely (CDN or WAF configuration issue), meaning some AI crawlers may receive error responses when attempting to check crawl rules.

No explicit AI bot blocking rules detected (no User-agent: GPTBot or User-agent: ClaudeBot disallow blocks) — but inaccessible critical paths reduce the content volume available for LLM indexation.

Fix: Remove /solutions and /industry-solutions from Disallow. Ensure robots.txt is served with a 200 status code without CDN interference. Add explicit Allow: / for GPTBot, PerplexityBot, and ClaudeBot.

Infrastructure Deep-Dive

We don't just identify issues — we build and implement fixes. Our team deploys schema markup, SSR/prerender solutions, and technical SEO improvements so AI platforms can actually index and cite your content.

Talk to Our Team to Unlock

30 min | Free | No commitment

// 05 CONTENT COMPETITIVENESS

Category Authority is the Missing Layer

Blog and case study content ranks. Decision-stage category content does not.

No Ranking Presence for Core Category Purchase-Intent Queries

High

profetai.com ranks #3 for automl (SV: 590) and holds strong branded positions. But critical purchase-intent queries have zero visibility:

  • 製造業AI轉型 (manufacturing AI transformation) — not ranking
  • 製造業AI軟體 (manufacturing AI software) — not ranking
  • 台灣AI平台比較 (Taiwan AI platform comparison) — not ranking
  • 工業AI無代碼 (industrial AI no-code) — not ranking

These are the queries that generate vendor consideration. LLMs train on pages that rank for these terms. Without organic visibility here, Profet AI is invisible in the upper funnel of AI-assisted buying journeys.

Fix: Create a dedicated comparison/category landing page targeting these queries. Include head-to-head feature comparison vs Chimes AI and aiwin.com.tw. Target 1,500+ words with structured H2/H3 headings that LLMs can extract.

English Content Stranded on Subdomain — Authority Split

Medium

en.profetai.com exists as a separate subdomain for English content. ChatGPT actively cited en.profetai.com as a separate source from profetai.com — confirming the authority split is real. Backlinks earned by the English site (including en.icaa.org.tw and asiamanufacturingreview.com citations) do not flow to the main domain's SEO authority.

This creates two problems: (1) the main domain loses link equity from international press coverage; (2) LLMs treat en.profetai.com and profetai.com as separate entities, diluting brand authority signals.

Fix: Migrate English content to profetai.com/en/ with hreflang annotations. Redirect en.profetai.com/* to profetai.com/en/* with 301s. Consolidates all authority under one domain for both SEO and LLM citation purposes.

Blog Content Engine Works — Needs Category Targeting Pivot

Competitive Intel

profetai.com ranks #5 for digital twin (SV: 2,400) via a single blog post — Digital Twin vs Domain Twin. This is proof the content engine can rank for competitive generic terms. The challenge: the blog targets technical curiosity content (what is AutoML, what is digital twin) rather than decision-stage queries.

Content TypeCurrentNeeded for LLM Citation
Educational explainersStrong ✓Maintain
Case study depth pagesModerate (BenQ, Maxxis, etc.)Expand with ROI metrics
Category comparison pagesNone ✗Priority 1 — drives LLM recommendation
Implementation guidesMinimal ✗Priority 2 — targets long-tail queries

Competitive Content Analysis

We benchmark your content against competitors using the exact signals LLMs prioritize — structure, depth, category authority, and third-party corroboration. Then we design content systems that systematically close the gap.

Talk to Our Team to Unlock

30 min | Free | No commitment

// 06 BRAND & POSITIONING

Strong Brand Authority, Weak Category Signals

Industry trade media knows Profet AI. Government startup registries don't.

Invisible in Training Data for Technical/Semiconductor Queries

High

Two of six ChatGPT queries — semiconductor AutoML quality prediction and manufacturing ML implementation cost — returned training-data-only answers with web_search explicitly disabled. These are the highest-intent queries in Profet AI's ICP.

LLM training data is predominantly English. Profet AI's primary web presence is Traditional Chinese. The result: ChatGPT has no learned knowledge of Profet AI in the context of semiconductor AI or manufacturing ML implementation — it only knows the brand when forced to search the web.

Fix: Create English-language technical content targeting these exact query types. A 2,000-word guide titled "How Taiwan Manufacturers Use AutoML for Quality Prediction: Implementation Guide" would serve both LLM training crawls and web_search retrieval.

AIO Category Disadvantage: Missing One Government Registry Listing

Medium

The direct cause of Chimes AI's Google AIO appearance for category queries is traceable: startup.sme.gov.tw — Taiwan's SME Bureau startup registry. Chimes AI is listed there. Profet AI is not. Google's AI Overview actively pulls from this source for manufacturing AI vendor queries.

This is a free government program. Submitting a listing takes less than a week. The AIO impact could be visible within 30-60 days as Google re-indexes the registry.

Additional platforms missing: G2, Capterra, and Gartner Peer Insights for the "AutoML platforms" and "Manufacturing AI" categories. These are the primary sources LLMs use for B2B SaaS vendor recommendations.

Strong Enterprise Case Study Credentials That LLMs Can Cite

Bright Spot

profetai.com has published case studies for BenQ, Maxxis, and BenQ Materials — all recognizable enterprise names that carry brand authority signals. ChatGPT cited Profet AI's comparative strength as "strongest in ease-of-use and case study depth" in the direct comparison query. This is a genuine competitive advantage.

The gap: these case studies lack the quantitative ROI metrics that LLMs use to justify recommendation rankings. Statements like "reduced defect rate by 23%" or "cut model training time from 6 weeks to 3 days" are citation-worthy data points. Currently missing from public-facing case study pages.

Fix: Work with BenQ and Maxxis to publish before/after metrics. Add a results summary section at the top of each case study with 2-3 quantitative outcomes. LLMs consistently cite pages with specific metrics over generic success stories.

Brand Perception Analysis

AI platforms form opinions about brands based on third-party signals — not your own website. We track sentiment, citation sources, and positioning across LLMs, then build the authority signals that shift how AI describes you.

Talk to Our Team to Unlock

30 min | Free | No commitment

// 07 ROADMAP & IMPACT

90-Day Path to AI Category Visibility

From JS-rendered invisibility to LLM category citations.

Site Readiness Score

Current5.1/10
Projected7.5/10

LLM Visibility Score

Current3.6/10
Projected6.5/10

Horizon 1: Technical Foundation (0-30 days)

0-30 Days

Unlocks AI crawler access to all content. Addresses the root cause of invisible category performance.

01

Implement prerender middleware for AI bot User-Agents (GPTBot, ClaudeBot, PerplexityBot) to serve static HTML on product pages

02

Remove /solutions and /industry-solutions from robots.txt Disallow; fix robots.txt accessibility for CDN/WAF bypass

03

Add JSON-LD structured data: Organization on homepage, FAQPage on /automl/, Product schema on /ailm/ and /domaintwin/, BreadcrumbList site-wide

04

Submit to startup.sme.gov.tw Taiwan startup registry (free, 1-week impact on AIO)

Horizon 2: Content Authority (30-60 days)

30-60 Days

Builds category keyword rankings and LLM recommendation presence for high-intent queries.

01

Migrate en.profetai.com to profetai.com/en/ with 301 redirects and hreflang — consolidate domain authority

02

Create a 2,000-word English-language category page: 'AutoML for Manufacturing: Taiwan Platform Comparison' — targeting semiconductor, digital twin, and AutoML implementation queries

03

Optimize page speed: defer non-critical JS, enable image lazy loading, add font preloading — target LCP < 4.0s

04

Create a G2 listing under 'AutoML Platforms' and 'Manufacturing Intelligence' categories

Horizon 3: Citation Infrastructure (60-90 days)

60-90 Days

Converts case study authority into LLM citation signals. Establishes long-term AEO moat.

01

Add quantitative ROI metrics to BenQ, Maxxis, and BenQ Materials case studies (before/after defect rate, time-to-insight, cost savings)

02

Publish implementation guides targeting training-data-blind queries: 'How to Implement AutoML in Semiconductor Quality Control — 90-Day Guide'

03

Build English-language technical documentation for AILM and DomainTwin products (increases LLM training data presence)

04

Submit Profet AI to Capterra and Gartner Peer Insights in 'Industrial AI' and 'AutoML Platforms' categories

Your Prioritized Roadmap

Every engagement starts with a clear plan: what to fix first, expected impact, and measurable milestones. We Discover, Design, Build, and Operate — you get results, not deliverables.

Talk to Our Team to Unlock

30 min | Free | No commitment

// WHO WE ARE

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

01
Head-Level Strategy

Solutions designed by operators who've led growth marketing and SEO at the Director/VP level. Not juniors following playbooks.

02
Agentic Execution

AI-powered workflows that move at machine speed. Audits, content, optimization, and reporting that would take a team weeks, delivered in days.

03
Flexible Engagement

No bloated retainers. Scope of work tailored to your stage, budget, and goals. Start small, scale when you see results.

Ready to Be
Cited by ChatGPT and Google AI?

Book a 30-minute call with us. We'll walk through this assessment, answer your questions, and map out exactly what it takes to get Profet AI recommended in the AI-driven buying journeys that matter.

Book Discovery Call

30 min | Free | No commitment

WhatsApp Email