BidScope AI Command Center
An AI-powered Google Ads intelligence system that turns raw ad data, lead quality, CRM revenue, competitor signals, landing page behavior, and human approvals into safer, smarter campaign decisions.
🔒 Demo Mode
Integration Lock Active
No Live Google Ads Writes
No API Keys Stored
Human Approval Required
436/436 Self-Tests Passing
9 Pages · 15 JS Modules
24 Mock Data Sources
⚙️ What BidScope Does
Nine core capabilities — all browser-safe, no live API connections required to demo.
Finds profitable keyword opportunities
Scores 150+ keyword ideas across 9 dimensions including intent, location, emergency value, commercial value, and profit prediction. Surfaces only campaign-ready candidates.
→ Keyword Intelligence
Detects wasted spend vs tracking gaps
Compares Google Ads conversion claims against Leads Gear calls, CRM jobs, and GA4 sessions to identify phantom conversions and attribution blind spots.
→ Google Ads Intel (Phase 2)
Builds campaign-ready keyword clusters
Groups keywords into 29 clusters across 9 services and 10 locations with campaign/ad group recommendations, exact/phrase splits, and negative risks mapped.
→ Cluster Explorer
Reviews negative keyword risks safely
Negative Keyword Lab scores 42 candidates across 14 categories with intent conflict checking, match type guidance, and human approval before any upload.
→ Negative Keyword Lab
Diagnoses landing page and UX issues
Checks LCP, CLS, FID, Clarity heatmap signals, and bounce behavior. Maps landing page gaps to specific keywords and ad groups with severity scoring.
→ GA4 + Clarity (Phase 3)
Adds competitor SERP intelligence
Surfaces competitor ad angles, impression share pressure, local pack presence, and SERP position data across 30 tracked keyword-location combinations.
→ Competitor SERP Module
Tracks decision accuracy over time
Records every recommendation, approval, and 7/14/30-day result. Builds accuracy scores and score modifiers that make future decisions more reliable.
→ Learning Dashboard
Prepares Google Ads Editor CSV previews
Generates CSV-structured output for campaigns, keywords, ad groups, and negatives — ready for human review before any Editor import. No auto-upload.
→ Export Center (Phase 1)
Keeps integrations locked until approved
All 13 evaluated connectors remain locked. No API keys stored. No live calls made. Every write action requires explicit human confirmation before enabling.
→ Settings / Connector Gate
✅ Completed Modules Snapshot
All modules are in demo mode with mock data. No real integrations are active.
| Module | Status | Data Mode | Key Value | Page |
|---|---|---|---|---|
| Daily Briefing / Decision Center / Approval Queue | Complete | Mock | Rule engine fires 80 rules; approval queue shows urgency, entity, export status | Phase 1 |
| Google Ads Intelligence | Complete | Mock | Search terms, dual conversions, auction IS, competitor PPC, ad assets | Phase 2 |
| GA4 + Clarity + Landing Page Diagnosis | Complete | Mock | Session quality, heatmap signals, LCP/CLS, speed fix roadmap per page | Phase 3 |
| Keyword Intelligence | Complete | Mock | 150 keywords, 9 scores, intent taxonomy, negative lab, LP gap finder | Keyword Intel |
| Keyword Cluster Explorer | Complete | Mock | 29 clusters, 9 services, 10 locations, 4 launch-ready, 17 CSV-ready | Cluster Explorer |
| Continuous Learning Dashboard | Complete | Mock | 346 learning records, 14 marts, decision history, accuracy scoring | Learning |
| Settings / Connected Accounts | Complete | Locked | 13 connector cards, 10-phase roadmap, safety matrix, integration gate | Settings |
| AI Prompts Library | Complete | Mock | 20 copy/paste prompts, 5 workflows, live self-test, category filter | AI Prompts |
| MCP / Connectors / Skills Evaluation | Complete | Locked | 13 connectors researched, phasing decided, no accounts connected | Settings |
| Google Ads Advanced Data Coverage | Complete | Mock | 18 coverage rows, P1/P2/P3 prioritisation, 41/41 self-tests | Settings |
| Export Center / CSV Preview | Complete | Mock | 7 export types, JSON/CSV download, export summary panel, no live upload | Phase 1 |
| Safety Governance | Complete | Active | 17 safety rules, integration lock, no 100% accuracy claims, write lock | Settings |
🗄️ Data Foundation
All data sources below are currently represented as mock/planning data. The architecture is integration-ready — no live connections are active.
📊 Google Ads (9 sources)
Campaigns · Ad groups · Keywords · Search terms · Ads & assets · Landing page mapping · Conversion actions · Geo/device/hour breakdowns · Advanced coverage map (18 rows, P1/P2/P3)
📈 Analytics & UX (3 sources)
GA4 sessions (quality, bounce, engagement) · Clarity heatmaps (rage clicks, dead clicks, scroll depth) · PageSpeed vitals (LCP, CLS, FID, speed fix roadmap)
💰 Revenue & Leads (2 sources)
Leads Gear call logs (duration, answered, qualified, keyword attribution) · CRM opportunities (revenue, profit, margin, job status, GCLID match)
🔎 Search Intelligence (3 sources)
GSC queries (impressions, clicks, CTR, position) · Keyword Planner CSV planning · Competitor SERP intelligence (30 keyword-location checks, ad angles, local pack)
🧠 Learning & Feedback (6 sources)
Decision history · Approval/rejection feedback · Result tracking (7/14/30 day) · Score modifiers · Mistake log · Winning/losing patterns
💬 AI Prompt Library (1 source)
20 expert copy-paste prompts, 13 modules, 12 categories. All safety-locked: no live writes, no auto-apply, human approval required on every output.
⚠️ Real integrations are not connected yet.
The architecture is integration-ready. All data sources above are mocked with realistic structures. Connecting real data begins only when the user explicitly says: start real integrations
🔑 Keyword Intelligence Summary
BidScope does not just list keywords — it scores, clusters, and approves them before any campaign action.
150
Keyword ideas
16
Intent types
9
Services
42
Negative candidates
25
LP gap findings
30
Competitor SERP checks
9 Dimensions Scored Per Keyword
Intent type · Emergency value · Commercial value · Location score · Negative risk · Landing page fit · Competitor SERP pressure · Profit prediction · Approval readiness — each with a formula and threshold.
Negative Keyword Lab
42 candidates across 14 intent categories. Each entry includes match type recommendation, conflict score, and why auto-upload is never used. Human approval required on every batch.
Landing Page Gap Finder
25 gaps detected across 8 check types (headline relevance, CTA clarity, trust signals, speed, mobile, schema, offer match, booking flow). Severity: critical / high / medium / low.
Competitor SERP Intelligence
30 keyword-location SERP snapshots. Surfaces top competitor ad angles, impression share gaps, local pack presence, and headline patterns BidScope hasn't used yet.
🗂️ Keyword Cluster Explorer Summary
Clusters are campaign-ready packages, not just keyword lists.
29
Clusters
9/9
Services represented
10
Locations
4
Launch-ready
7
Controlled test
17
CSV-ready
38/38
Self-tests PASS
Each Cluster Includes
Campaign & ad group recommendation · Exact/phrase keyword split · Negative risk flags · Landing page gap mapping · Competitor SERP pressure score · Learning signal history · Advanced Google Ads data needs · Budget plan · Pre-launch checklist · What NOT to do warnings · CSV export preview · Approval status
🧠 Continuous Learning Engine Summary
BidScope learns from every decision — without claiming 100% accuracy.
346
Learning records
14
Learning marts
78/78
Mart tests PASS
43/43
Data tests PASS
Learning Loop
Recommendation issued → Human approves or rejects → CSV/export action taken → 7 / 14 / 30 day result measured → Accuracy score updated → Score modifiers adjusted → Better future decisions. The system records what worked and what didn't — including mistakes — without claiming certainty.
Decision History
Every recommendation, its reasoning, approval status, and outcome tracked in a structured mart.
Approval Feedback
Human override reasons are stored and used to update rule weights and flag patterns.
Result Tracking
7, 14, and 30-day performance snapshots tied to specific decisions — no guessing.
Mistake Log
False positive recommendations are explicitly recorded to prevent repeat errors in the same category.
🔌 Connector / MCP Roadmap Summary
13 connectors evaluated. None are active. All are integration-locked.
✅ Phase 1 — Read-Only Safe
Google Ads Scripts → Sheets · Official Google Ads MCP (read-only) · Google Ads API Direct (read-only)
✅ Phase 2 — Analytics Read
Supermetrics (read-only) · GA4 · GSC · Google Business Profile
⚠️ Phase 3 — Paid / Optional
DataForSEO ($30/mo cap) · Leads Gear · CRM read-only · Offline Conversion preview
🔒 Deferred / High Risk
Supermetrics write access · Offline conversion upload (legacy endpoint blocked Jun 15 2026) · Zapier · Community MCPs
Recommended Safe Integration Order
- Google Ads Scripts → Sheets — read-only reporting, zero write risk, free
- Google Ads MCP (read-only) — campaign/keyword data via official API, mutations gated by env flag
- GA4 / GSC / PageSpeed — read-only analytics and search data, free tier available
- Leads Gear + CRM read-only — call and revenue attribution, no write access
- Offline conversion preview only — view and validate without uploading
- Offline conversion upload (with human approval) — requires Data Manager API (legacy endpoint blocked June 15 2026)
- Paid / optional connectors — DataForSEO, Supermetrics — only after cost controls and read-only verification
🛡️ Safety & Governance Summary
Safety is not a layer — it is built into every module, prompt, and export.
No real integrations connected
No API keys stored anywhere
No live API calls
No Google Ads writes
No auto keyword launch
No auto negative uploads
No auto budget changes
No offline conversion uploads
Broad match disabled by default
Competitor terms require human review
Human approval required for every action
No 100% accuracy claims ever made
CRM revenue truth overrides AI estimate
All 20 prompts: write_automation_allowed: false
All 20 prompts: human_approval_required: true
Integration lock visible on every page
📊 What Is Mock vs What Is Real
Current state: all data is mock. The system is architecture-ready for real connections.
| Area | Current Status | Real Integration Needed? | Risk If Missing | Next Step |
|---|---|---|---|---|
| Google Ads data | Mock | Yes — Phase 1 | No real campaign decisions possible | Google Ads Scripts → Sheets or MCP read-only |
| GA4 | Mock | Yes — Phase 2 | Session quality and bounce data unavailable | GA4 API read-only after Phase 1 |
| Clarity UX | Mock | Yes — Phase 2 | Heatmap and rage-click data unavailable | Clarity export or Supermetrics connector |
| Leads Gear | Mock | Yes — Phase 3 | Call attribution and lead quality missing | Leads Gear API read-only after GA4 |
| CRM | Mock | Yes — Phase 3 | Revenue and profit truth missing | CRM read-only after Leads Gear |
| GSC | Mock | Yes — Phase 2 | Organic vs paid overlap unavailable | GSC API read-only via Supermetrics or direct |
| PageSpeed | Mock | Optional — Phase 2 | LCP/CLS verification only available in mock | PageSpeed Insights API (free, read-only) |
| Keyword Planner CSV | Mock | Manual export only | Volume estimates remain approximate | Download from Google Ads UI, import CSV |
| Competitor SERP / DataForSEO | Mock | Optional — Phase 3 | Competitor intelligence remains illustrative | DataForSEO at $30/month cap after cost controls |
| Offline Conversion Upload | Locked | Yes — Phase 3+ with approval | True ROAS measurement blocked without it | Data Manager API required (legacy blocked Jun 15 2026) |
| Prompt Library | Live (Mock) | No — copy/paste only | None — fully usable in current state | Optional: expand to 30+ prompts |
| Learning Engine | Mock | Improves with real data | Accuracy scores are illustrative without real results | Connect after Phase 1–2 integrations complete |
🎯 Readiness Score / Gate
Where BidScope stands today across six dimensions.
Demo Readiness
✅ Complete
UI Readiness
✅ Complete
Mock Data Readiness
✅ Complete
Safety Readiness
✅ Complete
Self-Test Coverage
436/436 PASS
Navigation
✅ 10 Pages Linked
Real Integration
🔒 Locked
Production Deploy
— Not Started
Google Ads Writes
🔒 Locked / Not Allowed
🔒 Integration Lock
Real integrations only begin when the user explicitly says: start real integrations
Until then, all connectors remain locked, all data remains mock, and no live calls are made.
➡️ Recommended Next Steps
All options below are safe to proceed with under the integration lock.