Claude Track
Module 23
Claude Enterprise -- Module 23
Measuring What Matters: Six months in, ShopMate is writing 4,000 product descriptions per month, handling 2,000 customer chats per week, and generating 8 email campaigns per month. Maya reports to investors: staff time on repetitive tasks down 65%, customer response time down from 6 hours to 3 minutes, conversion rate on AI descriptions up 12%.

Enterprise Rollout Plan

A phased rollout de-risks enterprise Claude adoption by validating each layer before expanding to the next. Each phase has clear entry criteria, deliverables, and exit gates -- preventing the most common failure: expanding too fast before foundations are solid.

Rollout Timeline

Four-Phase Enterprise Claude Rollout
Foundation Wk 1โ€“6 Controlled Pilot Wk 7๏ฟฝ๏ฟฝ14 ยท 20โ€“50 users Dept. Expansion Mo. 4โ€“6 ยท 100โ€“500 users Enterprise Scale Mo. 7+ ยท Org-wide Policy Gov RAG + PII layer MCP + Platform Agents + ROI board Each phase requires successful exit criteria before advancing

KPI Framework

Productivity Metrics

Time saved per task type -- measure before and after. Tasks completed per hour. Time from request to first draft. These are the most convincing metrics for sceptical stakeholders at board level.

Financial Metrics

Total API cost vs value delivered. Cost per completed task. ROI as (value generated minus Claude cost) divided by Claude cost. Target a 3-10x ROI to justify expansion to the next phase.

Adoption Metrics

Monthly active users vs allocated seats. Session frequency per user. Feature adoption rate -- are users using advanced features or only basic chat? NPS from quarterly user surveys.

Quality and Safety

Human override rate -- what percentage of Claude outputs are edited before use, a proxy for quality. Compliance incidents. Hallucination rate on validated test sets. User-reported error rate.

RACI -- Roles and Responsibilities

ActivityExec SponsorCoE LeadEngineeringLegal/SecurityDept Head
Strategy and visionR/ACIII
Use case approvalARCRC
Policy authoringARCRI
Technical deploymentIARCI
Employee trainingIR/ACIR
Cost managementACRIC
ROI reportingARCIR

R = Responsible / A = Accountable / C = Consulted / I = Informed

Claude Track Complete.

You have completed all Claude modules including the full enterprise onboarding programme.

Start Windsurf Track --> Back to Foundations

ShopMate -- KPI Dashboard

Python -- scripts/shopmate_kpi_report.py
# scripts/shopmate_kpi_report.py -- Maya's monthly report for investors
import json
from pathlib import Path
from collections import defaultdict

logs = [json.loads(l) for l in Path("logs/shopmate_audit.jsonl").read_text().splitlines() if l]

by_feature = defaultdict(lambda:{"calls":0,"cost":0})
for log in logs:
    f = log["feature"]
    by_feature[f]["calls"] += 1
    by_feature[f]["cost"]  += log["cost_usd"]

total_cost = sum(v["cost"] for v in by_feature.values())

print("=== ShopMate Monthly KPI Report ===")
print(f"
{'Feature':<30} {'Calls':>7} {'Cost':>10} {'$/call':>8}")
print("-" * 58)
for feat, s in sorted(by_feature.items(), key=lambda x: -x[1]["cost"]):
    print(f"{feat:<30} {s['calls']:>7,} ${s['cost']:>9.2f} ${s['cost']/s['calls']:>7.4f}")
print(f"
Total AI cost this month: ${total_cost:.2f}")

# Business impact estimates
desc_calls  = by_feature.get("product_description",{}).get("calls",0)
chat_calls  = by_feature.get("customer_chat",{}).get("calls",0)
hours_saved = (desc_calls * 8 / 60) + (chat_calls * 5 / 60)
print(f"
Estimated staff hours saved:  {hours_saved:.0f} hrs")
print(f"Cost per hour saved:          ${total_cost/max(hours_saved,1):.2f}")