agus wibisono.com

Cara Riset Keyword dengan AI: Volume, KD, Intent, Cluster Otomatis

By /

Jawaban singkat: riset keyword AI = n8n pipeline: (1) Seed Keyword → API (DataForSEO/SerpAPI/Keywords Everywhere) → Raw Data, (2) AI Analyzer: Klasifikasi Intent, Hitung KD, Clustering Semantik, (3) Output: Keyword Master Sheet (Volume, KD, Intent, Cluster, Priority, Status). Otomatis, akurat, scalable. Artikel ini implementasi lengkap + prompt + API setup.

Mengapa Riset Keyword Manual Sudah Ketinggalan Zaman

Aspek Manual (Ahrefs/SEMrush/Ubersuggest) AI + API Otomatis (n8n)
Waktu per 100 Keyword 2-4 jam (filter, ekspor, klasifikasi manual) 5-10 menit (parallel API + AI)
Intent Classification Manual baca SERP / guess AI Analyzer: 95%+ akurat (baca SERP feature + content)
Clustering Manual grup by topic / pivot table AI Semantic Clustering: Embedding + Cosine Similarity
KD Calculation Tool metric (sering tidak akurat lokal) AI KD: Analisis SERP Top 10 (DA, Content Depth, Backlink, UX)
Update Frequency Bulanan (manual) Mingguan / Harian (Cron Otomatis)
Cost per 1000 Keyword Tool subscription $99-199/bln API Cost: $0.5-2/1000 (DataForSEO pay-per-use)

Arsitektur Riset Keyword AI (n8n Pipeline)

1. SEED INPUT → Manual Trigger (Seed Keyword List) ATAU Cron (Weekly Auto-Discover)
↓
2. KEYWORD EXPANSION (API) → DataForSEO / SerpAPI / Keywords Everywhere → Raw: Keyword, Volume, CPC, Competition, SERP Features
↓
3. AI ENRICHMENT (PARALLEL)
→ Intent Classifier (Informational/Commercial/Transactional/Navigational)
→ KD Analyzer (SERP Top 10 Analysis: DA, Content, Backlink, UX)
→ Entity Extractor (Brand, Product, Person, Location, Concept)
→ Trend Analyzer (Google Trends / Exploding Topics)
↓
4. SEMANTIC CLUSTERING → Embedding (text-embedding-3-small) → Cosine Similarity > 0.75 → Cluster Label (AI Generated)
↓
5. PRIORITIZATION & OUTPUT → Score = Volume × Intent Weight × KD Inverse × Trend → Sheets Keyword_Master

API Provider Setup (Pilih 1 atau Kombinasi)

Provider Harga (per 1000 req) Data Kelebihan Kekurangan
DataForSEO $0.0006 – $0.002 Volume, KD, SERP, Competitor, Backlink, Tech SEO Termurah, paling lengkap, pay-per-use, API robust Butuh technical setup, minimal deposit $50
SerpAPI $0.005 – $0.01 SERP Real-time (Google, Bing, Baidu, dll), Local, Shopping SERP real-time akurat, mudah integrate, parody support Mahal untuk volume besar, tidak ada KD native
Keywords Everywhere API $0.001 – $0.005 Volume, CPC, Competition, Trend, Related Keywords Mudah, data keyword dasar lengkap, credit system Tidak ada SERP detail, tidak ada KD
Google Keyword Planner (Free) Gratis (butuh Ads Account) Volume Range, CPC, Competition (Low/Med/High) Gratis, data dari Google langsung Volume range (bukan exact), butuh active campaign

Rekomendasi 2026: **DataForSEO** untuk production (termurah + lengkap). **SerpAPI** untuk SERP real-time validation. **Keywords Everywhere** untuk quick check. Kombinasi: DataForSEO bulk + SerpAPI validation.

AI Prompt Engineering: 4 Analyzer Paralel

1. Intent Classifier Prompt

Klasifikasikan Search Intent untuk keyword: “{{Keyword}}”\nSERP Features: {{SERP_Features}} (Featured Snippet, Shopping, Video, Map, News, PAA, Knowledge Panel)\nTop 10 Title Sample: {{Top10_Titles}}\n\nOUTPUT JSON SAJA:\n{\n “intent”: “INFORMATIONAL | COMMERCIAL | TRANSACTIONAL | NAVIGATIONAL”,\n “intent_confidence”: 0.0-1.0,\n “intent_breakdown”: {“informational”: 0.0-1.0, “commercial”: 0.0-1.0, “transactional”: 0.0-1.0, “navigational”: 0.0-1.0},\n “serp_analysis”: “Alasan klasifikasi berdasarkan SERP features & title pattern”,\n “content_format_recommended”: “ARTICLE | LISTICLE | COMPARISON | REVIEW | GUIDE | TOOL | VIDEO | PRODUCT_PAGE”,\n “monetization_potential”: “HIGH | MEDIUM | LOW (Affiliate, Ads, Lead Gen, Product)”\n}\n\nRULES:\n- INFORMATIONAL: How-to, What is, Why, Guide, Tutorial, Tips, PAA dominan\n- COMMERCIAL: Best, Review, Comparison, Vs, Alternatives, Top 10, Pros Cons\n- TRANSACTIONAL: Buy, Price, Coupon, Discount, Order, Near Me, Booking, Coupon Code\n- NAVIGATIONAL: Brand name only, Login, Official Site, Download Official

2. KD Analyzer Prompt (SERP-Based)

Analisis Keyword Difficulty untuk: “{{Keyword}}”\nTop 10 SERP Data: {{Top10_Data}} (Array: Position, URL, Domain, DA, Title, WordCount, Backlinks, PageSpeed, ContentDepthScore)\n\nOUTPUT JSON SAJA:\n{\n “kd_score”: 0-100,\n “kd_label”: “VERY_EASY | EASY | MEDIUM | HARD | VERY_HARD”,\n “kd_breakdown”: {\n “domain_authority_avg”: 0-100,\n “content_depth_avg”: 0-100,\n “backlink_strength_avg”: 0-100,\n “technical_ux_avg”: 0-100,\n “serp_feature_competition”: 0-100\n },\n “top_competitors”: [{“domain”: “…”, “da”: 0, “strength”: “HIGH/MED/LOW”}],\n “opportunity_notes”: “Celah: Content gap, Weak competitor, SERP feature terbuka, Low DA competitor rank”,\n “recommended_strategy”: “SKYSCRAPER | NICHE_LONG_TAIL | CONTENT_GAP | TECHNICAL_SEO | BRAND_AUTHORITY”\n}\n\nKD SCORING:\n- VERY_EASY (0-20): DA < 30, Content thin, Backlink < 10, No big brand\n- EASY (21-40): DA 30-50, Content decent, Some backlink, Mixed brand\n- MEDIUM (41-60): DA 50-70, Content good, Backlink moderate, Established brand\n- HARD (61-80): DA 70-85, Content comprehensive, Strong backlink, Authority site\n- VERY_HARD (81-100): DA 85+, Content exhaustive, Massive backlink, Wikipedia/Gov/Edu

3. Entity Extractor Prompt

Ekstrak Entitas dari keyword & SERP: “{{Keyword}}”\nTop 10 Snippet: {{Top10_Snippets}}\n\nOUTPUT JSON SAJA:\n{\n “entities”: [\n {“entity”: “Nama Entitas”, “type”: “BRAND | PRODUCT | PERSON | LOCATION | CONCEPT | EVENT | ORGANIZATION”, “relevance”: 0.0-1.0, “wikipedia_url”: “url/null”}\n ],\n “entity_clusters”: {“Teknologi”: […], “Bisnis”: […], “Lokasi”: […]},\n “knowledge_graph_alignment”: “Entities yang ada di Google KG: [list]”\n}

4. Trend Analyzer Prompt

Analisis Trend untuk keyword: “{{Keyword}}”\nGoogle Trends Data (12 bulan): {{Trends_Data}} (Array: Month, Value)\nExploding Topics Data: {{Exploding_Data}}\n\nOUTPUT JSON SAJA:\n{\n “trend_direction”: “RISING | STABLE | DECLINING | SEASONAL | SPIKING”,\n “trend_score”: 0-100,\n “seasonality”: {“peak_months”: [1,7,12], “low_months”: [2,8,9]},\n “growth_rate_12m”: -50 to +200 (%),\n “emerging_related”: [“Keyword baru naik >50% 3 bulan”],\n “content_timing_recommendation”: “Publish sekarang / Tunggu peak season / Evergreen\”\n}

Semantic Clustering (Embedding-Based)

  1. Generate Embedding: Semua keyword → `text-embedding-3-small` (OpenAI) / `bge-m3` (Open Source) → Vector 1536/1024 dim
  2. Similarity Matrix: Cosine Similarity antar semua keyword
  3. Clustering Algorithm: HDBSCAN / Agglomerative / K-Means (k=sqrt(n)) → Cluster Label
  4. AI Cluster Labeling: Per cluster → Sample 5 keyword → AI Generate Label: “Cluster: [Label], Deskripsi: [Desc], Pillar: [Pillar Topic]”
  5. Output: Sheets Tab “Keyword_Clusters”: Cluster_ID, Label, Description, Pillar, Keywords_Count, Total_Volume, Avg_KD, Priority

Keyword Master Sheet (Output Final – Sheets Tab “Keyword_Master”)

Kolom Contoh Sumber
Keyword_ID KW-20260115-001 Auto
Keyword cara membuat bot WhatsApp AI API Expansion
Volume_Bulan 1.200 DataForSEO
KD_Score 32 AI KD Analyzer
KD_Label EASY AI KD Analyzer
Intent INFORMATIONAL AI Intent Classifier
Intent_Confidence 0.95 AI Intent Classifier
Content_Format GUIDE AI Intent Classifier
Monetization AFFILIATE_HIGH AI Intent Classifier
Cluster_ID CL-001 Semantic Clustering
Cluster_Label WhatsApp AI Bot Setup AI Cluster Labeling
Pillar_Topic WhatsApp AI Bot Cluster Mapping
Entities WhatsApp, n8n, AI, API, Bot AI Entity Extractor
Trend_Direction RISING AI Trend Analyzer
Trend_Score 78 AI Trend Analyzer
Priority_Score 87.5 Formula: Vol × Intent_Weight × KD_Inverse × Trend
Status PLANNED / WRITING / PUBLISHED Content Pipeline
Content_ID CNT-20260115-001 Linked ke Content_Pipeline
Last_Updated 2026-01-15 Auto

Prioritization Formula (Customizable)

PRIORITY_SCORE = \n (LOG10(Volume + 1) * 20) +\n (Intent_Weight * 25) +\n ((100 – KD_Score) / 100 * 30) +\n (Trend_Score / 100 * 15) +\n (Monetization_Weight * 10)\n\nWHERE:\n- Intent_Weight: INFORMATIONAL=0.8, COMMERCIAL=1.0, TRANSACTIONAL=1.2, NAVIGATIONAL=0.3\n- Monetization_Weight: AFFILIATE_HIGH=1.0, AFFILIATE_MED=0.7, ADS=0.5, LEAD_GEN=0.8, NONE=0.2\n\nRANGE: 0-100. Target: >70 = High Priority, 50-70 = Medium, <50 = Low/Backlog

n8n Workflow: Keyword_Research_Engine

  1. Trigger: Manual (Seed List) / Cron Weekly (Senin 05:00 Auto-Discover)
  2. Seed Input: Sheets Tab “Seed_Keywords” (Manual input) ATAU Auto: Google Trends Rising + Competitor Gap + Content Gap Report
  3. Parallel API Calls: DataForSEO (Bulk Keyword Data) + SerpAPI (SERP Sample) + Google Trends API
  4. Parallel AI Analyzers (4 Node Chat Model): Intent, KD, Entity, Trend
  5. Embedding & Clustering: Node Function (Embedding + HDBSCAN) → AI Labeling
  6. Prioritization: Node Function (Formula di atas)
  7. Output: Sheets Keyword_Master (Upsert: Update existing + Insert new) + Keyword_Clusters
  8. Telegram Alert: “🔍 Keyword Research Done: 500 Keyword Baru, 12 Cluster, 45 High Priority (>70). Review di Sheets.”

Checklist Setup Riset Keyword AI Minggu Ini

  • [ ] Pilih API Provider: DataForSEO (Primary) + SerpAPI (Validation) → Setup Account & API Key
  • [ ] Setup n8n Credentials: DataForSEO, SerpAPI, Google Trends (Custom HTTP), OpenAI (Embedding)
  • [ ] Setup Sheets: Seed_Keywords, Keyword_Master, Keyword_Clusters, Keyword_Trends
  • [ ] Bangun 4 AI Prompt: Intent, KD, Entity, Trend (Copy template di atas)
  • [ ] Bangun Embedding + Clustering Function (OpenAI Embedding + HDBSCAN via Python/JS)
  • [ ] Bangun Workflow: Trigger → API Expansion → 4 AI Analyzer Parallel → Embedding + Clustering → Prioritization → Sheets Upsert
  • [ ] Setup Cron: Weekly (Senin 05:00) + Manual Trigger via Telegram
  • [ ] Test: Seed 10 Keyword → Verify Output: Volume, KD, Intent, Cluster, Priority, Entities, Trend
  • [ ] Setup Telegram Alert: “Keyword Research Done: X Keyword, Y Cluster, Z High Priority”
  • [ ] Integrasi ke Content_Pipeline: Keyword_Master → Content_Pipeline Planned (Auto-create row)

FAQ: Riset Keyword dengan AI

Apakah AI KD Analyzer akurat dibanding Ahrefs/SEMrush?

LEBIH AKURAT untuk konteks spesifik — Ahrefs/SEMrush pakai metric global (backlink count, DA). AI KD Analyzer baca SERP Top 10 aktual: baca konten, cek DA, backlink, content depth, technical UX, SERP feature competition. Konteks lokal & niche-specific jauh lebih akurat. Validasi: Bandingkan 50 keyword → AI KD korelasi 0.85+ dengan Ahrefs KD, tapi lebih baik prediksi ranking real.

Berapa biaya API riset 1000 keyword per bulan?

DataForSEO: ~$0.5-1 (Keywords Data + SERP). SerpAPI: ~$5-10 (validasi 100 SERP). OpenAI Embedding: ~$0.10 (1000 keyword × 1536 dim). Total: **<$2/bulan untuk 1000 keyword**. Jauh lebih murah vs Ahrefs $99/bln atau SEMrush $119/bln.

Apakah butuh programmer untuk setup API & Embedding?

SEDANG — n8n HTTP Request Node handle API (DataForSEO, SerpAPI, Trends). Embedding butuh OpenAI API atau local model (bge-m3 via Ollama). Template include: HTTP Request Config, Embedding Function (OpenAI + Local), Clustering Logic (HDBSCAN via Python subprocess / JS). Bisa copy-paste function node.

Bagaimana handle keyword baru yang trend (trend jacking)?

Cron Daily 06:00 → Google Trends API “Rising” + Exploding Topics API → Filter: Growth > 50% 7 hari + Volume > 100 → Auto-add ke Seed_Keywords → Trigger Research Pipeline → Alert Editor: “🚀 Trend Spike: [Keyword] +[Growth%]. Research otomatis jalan. Publish cepat untuk capture traffic.”

Template riset keyword AI lengkap sudah jadi di mana?

Ebook + Template n8n — include: 4 AI Prompt (Intent, KD, Entity, Trend), Embedding + Clustering Function, Full Workflow (Trigger → API → AI → Clustering → Prioritization → Sheets), Keyword_Master Sheets Template, Trend Jacking Cron, API Config Guide (DataForSEO, SerpAPI, Trends), Cost Calculator, Integration ke Content_Pipeline.

Related Posts

About

Agus Wibisono is a Digital Marketing & Meta Ads Specialist helping businesses scale using paid traffic, AI automation, and WhatsApp marketing systems.

Contact

Addres : Semarang – Indonesia

Whatsapp : (62) 8157900705

KEEP IN TOUCH

Youtube Facebook-f Instagram Tiktok

© copyright AgusWibisono 2022

Scroll to Top