# Weekly Ops Digest Agent

**Agent Name:** Weekly Ops Digest Agent
**Role:** Operational trend analyst
**Frequency:** Every Monday at 8:00 AM
**Output Channel:** Slack (#product-ops)
**Run Time:** ~15 minutes

---

## PURPOSE

Your daily agents catch fires. They tell you which Tier 1 features have adoption issues, which customers are at risk, which bugs are critical. But they don't connect the dots.

You're hearing about the same customer escalation for the third time this week. Support resolution times are degrading, you had two fast resolutions Monday and Tuesday, then the bottom fell out Wednesday and Thursday. One feature's adoption curve is flat, but you can't tell if it's because it's genuinely not valuable or because there's a hidden UX blocker.

The Weekly Ops Digest Agent looks at all your daily reports from the past week and asks: What's the pattern here? What's getting worse? What's repeating? Where is the system breaking down?

This is how you spot systemic problems before they become crises.

---

## DATA SOURCES

### 1. **All Daily Agent Outputs** (Last 7 Days)
**Source:** Slack, Email, Notion database
**Query:**
- Daily Focus agent reports (M-F)
- Daily Health Check reports (M-F)
- Customer escalation logs
- Support ticket summaries
- Extract: Issues flagged, trends noted, severity levels

### 2. **Support Ticket Data & Trends**
**Query:**
- Total tickets this week vs. last week (count, trend)
- Average resolution time (daily granularity)
- Top issues by frequency
- Escalation rate
- First-response time trend
- Extract from: Zendesk, Intercom, or support database

### 3. **Engineering Completion & Bug Data**
**Query:**
- Story points completed this week vs. estimated (velocity trend)
- Critical bug count (trend week-over-week)
- Bug fix time (opened to closed, average)
- PRs merged daily (velocity trending)
- Test coverage and failure rates
- Build health (pass rate)

### 4. **Customer Escalation Pattern Data**
**Query:**
- All escalations opened in last week
- Time to escalation (first contact to escalation)
- Escalation resolution time
- Escalation reason (feature request, bug, churn risk, etc.)
- Whether escalations are related (same customer, same feature, etc.)

### 5. **Operational Metrics**
**Query:**
- Customer support SLA compliance
- Engineering deployment frequency
- Feature adoption curves (from daily dashboards)
- Known tech debt items affecting operations
- System reliability (uptime, error rate)

---

## REPORT STRUCTURE

### 1. WEEK-OVER-WEEK TREND SUMMARY

One-paragraph assessment of operational health.

**Format:**
```
[WEEK OF MMM DD]: Operations [improving/stable/degrading].
Positive: [1-2 improving metrics].
Concern: [1-2 degrading metrics].
Action: [1 priority focus for next week].
```

### 2. KEY METRICS - WEEK OVER WEEK

All metrics trended, with visual indicators.

**Format:**
| Metric | This Week | Last Week | Trend | Status |
|---|---|---|---|---|
| Support Tickets | 47 | 42 | ↑ 12% | 🟡 |
| Avg Resolution Time | 18 hrs | 14 hrs | ↑ (slower) | 🔴 |
| Customer Escalations | 3 | 2 | ↑ | 🟡 |
| Engineering Velocity | 87 pts | 92 pts | ↓ 5% | 🟡 |
| Critical Bugs | 2 | 4 | ↓ | 🟢 |
| Avg Bug Fix Time | 2.3 days | 2.8 days | ↓ (faster) | 🟢 |
| Build Health (pass rate) | 97% | 96% | ↑ | 🟢 |
| Deployment Frequency | 5x/week | 4x/week | ↑ | 🟢 |

**Status Indicators:**
- 🟢 On track or improving
- 🟡 Stable but watch
- 🔴 Degrading or concerning

### 3. RECURRING PATTERNS (The Systemic Issues)

This is the core of the report. What's repeating? What does that signal?

**Format:**
```
**Pattern 1: [Pattern Name]**
- Issue: [Description of pattern]
- Frequency: [Appearing in N daily reports / affecting N customers / N support tickets]
- First noticed: [Date]
- Likely root cause: [What's actually broken]
- Impact: [Customer impact, business impact]
- Action: [What to fix / investigate]
- Owner: [Name]
- ETA: [When this will be resolved]

**Pattern 2: [Pattern Name]**
[Same format]
```

**Examples of patterns you might identify:**
- "Payment processing is failing for [feature] in Chrome on Windows, 3 separate customer reports of the same issue"
- "Support resolution time jumped from 14 hrs to 22 hrs starting Thursday; correlates with [team member] out sick"
- "New customers in [region] are asking the same 5 questions about [feature]; indicates documentation gap"
- "Adoption of [feature] is flat but usage per adopter is 10x the target; indicates discovery problem, not value problem"
- "Escalation reason in 6 of 8 escalations was 'not aware feature exists'; indicates GTM/communication failure"

### 4. TOP UNRESOLVED ITEMS (Carried Over)

What from last week is still open and why.

**Format:**
```
**[Item Name]** (Opened [date])
- Status: Still in progress / blocked on / waiting for
- Days open: [N]
- Why it's still open: [Root cause]
- Impact if not resolved soon: [Customer/business impact]
- Action needed: [Next step]
- Target resolution: [Date]
```

### 5. TEAM PERFORMANCE HIGHLIGHTS

Wins worth calling out. Show your teams they're winning.

**Format:**
```
**[Team name]**: [Achievement]
- Detail: [What they did and why it mattered]
- Impact: [Metric improvement, customer outcome, or time saved]

**[Team name]**: [Achievement]
[Same format]
```

### 6. ESCALATION SUMMARY & ANALYSIS

Every escalation from the week, with outcome.

**Format:**
```
Total Escalations: [N] (vs [N] last week)

| Customer | Account Value | Reason | Severity | Status | Owner | Resolution Time |
|---|---|---|---|---|---|---|
| [Name] | [$ARR] | [Reason] | Critical/High/Medium | Open/Resolved | [Name] | [Time] |

Escalation Trends:
- Most common reason: [Reason] ([N] instances)
- Fastest resolution: [Escalation name] ([X hours])
- Slowest resolution: [Escalation name] ([X days])
- Pattern: [Is there a pattern in reasons / customers / timing?]
```

### 7. OPERATIONAL RECOMMENDATIONS (For Next Week)

What should the team focus on based on this week's data.

**Format:**
```
**Priority 1: [Focus Area]**
- Why: [Metric is degrading / pattern is recurring / risk is growing]
- Action: [Specific thing to do]
- Owner: [Name]
- Target: [Success metric]

**Priority 2: [Focus Area]**
[Same format]

**Priority 3: [Focus Area]**
[Same format]
```

---

## HOW TO SET IT UP

### Step 1: Aggregate Daily Outputs
- Create a Slack channel (#ops-dashboard) where all daily agents post summaries
- Or: Create a Notion database that auto-ingests daily agent outputs
- Give the Weekly Ops Digest agent read access to all channels/docs

### Step 2: Connect Data Sources
- **Support tickets**: Connect to Zendesk/Intercom API for ticket trends
- **Engineering**: Connect to GitHub/Linear for velocity and bug data
- **Escalations**: Create a shared escalation log (Google Sheet or Notion) that agents feed into

### Step 3: Define "Patterns"
- Set thresholds for what counts as recurring (e.g., same issue in 3+ daily reports = pattern)
- Define "degrading" (e.g., resolution time up 25% = concern)
- Define severity (critical = affects revenue, high = affects adoption, medium = nice to fix)

### Step 4: Customize Metrics
- Decide which metrics matter most to your org
- Set "healthy" vs "at risk" ranges for each metric
- Decide what kind of trends you want to flag (up, down, volatile)

### Step 5: Schedule
- Set to run every Monday at 8:00 AM (after executive report if you run that too)
- Set timezone to your primary office location

---

## SAMPLE PROMPT (Customizable)

```
You are an operations analyst. Your job is to look at a week of daily product reports and identify patterns, trends, and systemic issues.

DATA INPUTS:
- Daily focus reports (M-F)
- Daily health check reports (M-F)
- Support ticket summaries
- Customer escalation logs
- Engineering velocity data

INSTRUCTIONS:
1. Calculate week-over-week trend for: support tickets, resolution time, escalations, velocity, critical bugs, bug fix time, build health, deployment frequency
2. Identify recurring patterns: What issue appeared in multiple daily reports? What customer problem appeared more than once? What metric is consistently degrading?
3. For each pattern, state: what it is, how often it's appearing, what's causing it, and what should be done about it
4. List unresolved items from last week that are still open. Why are they still open?
5. Call out 2-3 team wins from the week
6. Summarize all escalations: what were the reasons, how long to resolve, are there patterns?
7. Based on trends, recommend 3 priorities for next week

TONE: Analytical, data-driven. Focus on patterns and systemic issues, not one-off problems.
OUTPUT: Markdown formatted for Slack and email.
```

---

## FREQUENCY & TIMING

- **Frequency:** Every Monday
- **Time:** 8:00 AM (1 hour after executive report, if you run that)
- **Runtime:** ~15 minutes
- **Timezone:** Your company's primary timezone

---

## WHAT SUCCESS LOOKS LIKE

✓ Systemic issues get surfaced early (before they cascade)
✓ You can tell the difference between a one-off problem and a pattern that needs fixing
✓ Team performance visibility increases (wins are celebrated, concerns are addressed)
✓ Escalation resolution times improve because you're identifying patterns and fixing root causes
✓ Operational conversations in team meetings are more data-driven and less reactive