08.11 AI-Assisted GTM Tool List and Practical Prompt Library
Governance metadata | Owner: Head of Research | Last reviewed: 2026-10-22 | Next review: 2027-01-31
Tool positioning: the tool matrix and implementation workflows that empower GTM operators with AI (see 08.10 Frontier Topic: Human-Machine Collaboration Workflows in the AI Agent Era).
1. The Application Matrix of AI Across GTM Steps
| GTM step | What AI can do | Recommended tools | Frequency / threshold | Owner | Execution notes | Related |
|---|---|---|---|---|---|---|
| Product definition insight | Run market sizing, scrape competitor data, mine user demand keywords | ChatGPT / Claude / Perplexity | 1 time per product definition, data refreshed 1 time per week; threshold: TAM estimate deviation >10% requires human review | Growth lead | Use chained prompting: give the background first → then request the data → then request the conclusion; conclusions must be cross-validated against the original data sources | See 02.06 Competitor Continuous Monitoring Radar §1 price/volume monitoring |
| Competitor monitoring | Scrape competitor price/review/BSR changes daily and alert automatically | ChatGPT + Web Search / AgentReach + Feishu Bitable | Scrape daily, weekly report every Monday; thresholds: price move >5% alerts the same day, BSR move >20% alerts the same day (aligned with 16.11§2) | Growth lead | Configure price/BSR monitoring tables in Feishu Bitable, with anomalies auto-pushed to the WeCom group | See 02.06 Competitor Continuous Monitoring Radar §2 weekly monitoring |
| Content generation | Batch-generate Listing copy, ad creatives, and social media post drafts | Claude / ChatGPT / Midjourney | 2 batch generations per week; threshold: human review pass rate ≥90% before publishing, 0 tolerance for prohibited wording | Growth lead, content operations reviews | Human review is the floor; AI output ≠ publish-ready; run the compliance checklist before publishing | — |
| Data analysis | Upload sales tables to run attribution analysis, ad ROI analysis, and inventory forecasting | ChatGPT + Code Interpreter / Python + Feishu Bitable | Attribution 1 time per week, daily during big promos; thresholds: red alert at ACOS>30% or DOS<15 days (note: a tighter emergency threshold for promo periods; the daily mature steady state aligns with ACOS<18% and DOS 45~75 days) | Growth lead | Upload the CSV to Code Interpreter → have AI analyze → human-verify the conclusions, then write them to the Feishu Bitable dashboard | See 02.06 Competitor Continuous Monitoring Radar §1 |
| Customer insight | Batch-run review sentiment analysis, negative review classification, and NPS trend extraction | Claude / ChatGPT + Feishu Bitable | Batch-process new Reviews 1 time per week; thresholds: negative review rate >5% or NPS<30 triggers CS intervention (aligned with the 15.5§4 dashboard) | CS lead, growth supporting | Batch-process reviews to extract high-frequency keywords, output the Top 3 improvement priorities, and open work orders | See 06.02 Post-Sales Reputation: Complaint Tiering, NPS Tracking, and the VOC Closed Loop §4 |
| Document processing | Run contract review, compliance document translation, and multilingual localization drafts | Claude / DeepL / ChatGPT | On demand, contract items completed within T+1 day; threshold: 100% human review rate on key clauses | Legal lead, growth supporting | Key clauses must be human-reviewed, and tariff/compliance clauses require legal sign-off | — |
2. AI Workflow Examples for the GTM Operator
1. Workflow 1: New-Product Definition Insight (30 minutes)
Step 1: Market scan (5 minutes)
Prompt: "I am a smart hardware GTM operator planning to launch a [category] product in the US market.
Please help me analyze: 1) TAM/SAM/SOM estimation framework 2) main competitors and their price bands 3) channel selection recommendations"
Tools: Perplexity + Claude
Step 2: Competitor deep dive (15 minutes)
Prompt: "Please analyze the following for [competitor name] on Amazon: 1) pricing strategy 2) high-frequency Review keywords
3) advertising playbook 4) supply chain weaknesses"
Tools: AgentReach + Helium 10 + Keepa
Step 3: Output the first BP draft (10 minutes)
Prompt: "Based on the analysis above, generate a one-page business plan (BP) for me,
covering: market opportunity, competitive landscape, profit model, and risk points"
Tools: Claude + Feishu Bitable (BP template)Implementation: Owner: growth lead | Frequency: 1 time per product definition | Deadline: BP first draft output and entered into Feishu Bitable within T+2 days | Thresholds: competitor prices cross-validated against Keepa/CamelCamelCamel, with deviations >5% requiring a second verification; TAM deviations >10% requiring human review | Related: monitoring data is synced to 02.06 Competitor Continuous Monitoring Radar §1 price/volume monitoring
2. Workflow 2: Big-Promo Post-Mortem Review (1 hour)
Step 1: Upload the data (10 minutes)
Action: export the advertising report + sales report as CSV and upload to ChatGPT Code Interpreter
Tools: Shopify/Amazon backend + Feishu Bitable
Step 2: Analysis instruction (20 minutes)
Prompt: "Please analyze this big-promo data: 1) advertising ACOS trend 2) ROI comparison across channels
3) inventory turnover 4) the gap against target 5) improvement suggestions for the next big promo"
Tools: ChatGPT Code Interpreter + Python
Step 3: Generate the post-mortem report (30 minutes)
Prompt: "Based on the analysis above, generate a structured big-promo post-mortem report,
covering: core data, what worked, what did not, and Action Items"
Tools: Claude + Feishu Bitable (post-mortem dashboard)Implementation: Owner: growth lead | Frequency: run within T+3 days after every big promo | Deadline: post-mortem report output by T+3 days | Thresholds: red alert at ACOS>30% or DOS<15 days (a tighter emergency threshold for promos), requiring an improvement work order the same day; data conclusions require human verification | Related: inventory/ACOS thresholds align with 02.06 Competitor Continuous Monitoring Radar §2 anomaly flagging
3. AI Usage Red Lines (Non-Negotiable)
| Red line | Why | The correct way |
|---|---|---|
| ❌ Treating AI-generated data as a decision basis | AI will "hallucinate" false data | The original data sources must be cross-validated |
| ❌ Publishing AI-generated Listings directly | They may contain inaccurate or non-compliant content | Human review → compliance check → A/B test |
| ❌ Skipping field validation after AI competitor analysis | Online data can be stale or distorted | Combine with actual purchase experience and user interviews |
| ❌ Using AI in place of user research | AI cannot replace real user feedback | AI assists the analysis; users validate the conclusions |
| ❌ Feeding sensitive business data into public AI | Data security risk | Use the enterprise version or a private deployment |
4. Prompt Template Library (GTM Scenarios)
1. Template 1: Competitor Analysis
You are a senior smart hardware GTM operator. Please analyze the following competitor:
- Competitor name: [fill in]
- Platform: Amazon / independent site / other
- Analysis dimensions: pricing strategy, Review keywords, advertising playbook, supply chain weaknesses
Please provide: 1) a summary of the competitor's strengths and weaknesses 2) our differentiation opportunities 3) specific action recommendationsImplementation: Tools: Claude/ChatGPT + AgentReach + Keepa/Helium 10 | Owner: growth lead | Frequency: 1 time per week (Monday) | Deadline: competitor snapshot output and entered into Feishu Bitable within T+1 day | Thresholds: price move >5% or BSR move >20% alerts the same day (aligned with 16.11§2); data must be cross-validated against Keepa | Related: see 02.06 Competitor Continuous Monitoring Radar §1 and §2
2. Template 2: Listing Optimization
Please optimize the following Amazon Listing:
- Original copy: [paste]
- Target keywords: [fill in]
- Top 3 competitor titles: [paste]
Please output: an optimized title (within 200 characters), 5 bullet points, and a product description
Requirements: highlight differentiated selling points, plant the core keywords, and comply with Amazon rulesImplementation: Tools: Claude/ChatGPT + Helium 10 (keywords) | Owner: growth lead, content operations reviews | Frequency: 1 time per Listing iteration | Deadline: optimization completed and submitted for compliance check within T+3 days | Thresholds: keyword coverage ≥90%, A/B testing only after human review passes, 0 tolerance for prohibited wording | Related: pre-publish compliance check refers to 02.06 Competitor Continuous Monitoring Radar §3 Listing change monitoring
3. Template 3: Review Analysis
Please analyze the following user reviews and extract: 1) high-frequency positive keywords 2) high-frequency negative keywords
3) the TOP 3 features users care about most 4) improvement priority recommendations
Review data: [paste or upload the CSV]Implementation: Tools: Claude/ChatGPT + Feishu Bitable (review dashboard) | Owner: CS lead, growth supporting | Frequency: batch-processed 1 time per week | Deadline: analysis output and work orders opened within T+2 days | Thresholds: negative review rate >5% or NPS<30 triggers CS intervention (aligned with 15.5§4); the Top 3 high-frequency negative terms need a same-day response | Related: see 02.06 Competitor Continuous Monitoring Radar §1 Review increment monitoring and 06.02 Post-Sales Reputation: Complaint Tiering, NPS Tracking, and the VOC Closed Loop §4
5. Tool Selection Recommendations
| Scenario | Recommended tools | Why |
|---|---|---|
| Daily conversation and analysis | ChatGPT Plus / Claude Pro | Strong all-round capability and a mature ecosystem |
| Long-document processing | Claude (200K context) | Excellent long-text comprehension |
| Real-time data and search | Perplexity / AgentReach | Web-connected search with fresh data |
| Image and design generation | Midjourney / DALL-E | Ad creatives and social media imagery |
| Data analysis | ChatGPT Code Interpreter | Upload the CSV and analyze it directly |
| Code and automation | Claude + Python / n8n | Complex workflow automation |
💡 Recommendation: start with Workflow 1 (product definition insight) to get the feel, then expand to other scenarios once you are comfortable. AI is a tool; the operator's judgment is the real core competency.
Execution Checklist
The following are self-check items for landing the key actions in this chapter. Check the box after execution and note the owner and completion date.
- [ ] Application matrix landed: at least one scenario per function (market, product, marketing, sales, service) actually in use
- [ ] Workflow reproducible: steps and prompts can be followed directly, not a conceptual description
- [ ] Red lines posted: the five forbidden uses (data, compliance, decisions) written into team policy
- [ ] Prompt templates copyable: each scenario template has complete fields, usable after swapping variables
- [ ] Tool selection justified: selection criteria (cost, data compliance, integration effort) map to the conclusion rather than a list
- [ ] Human review retained: AI output is reviewed by a person before external publication; no fully automated channel