Your weekly ops report, written by agents

What is Weekly Ops Report?

Weekly Ops Report is a scheduled four-agent pipeline that turns product metrics into a structured Excel report. It computes period-over-period and year-over-year change, flags anomalies with root-cause candidates, and ranks recommended actions by priority.

Weekly Ops Report runs a four-agent pipeline on your product metrics — collect, clean, analyze, report — and returns a four-sheet Excel workbook (Data Overview, Trend Analysis, Anomaly Annotation, Operations Recommendations). No dashboards to build, no prompts to write.

How it works

Four specialised agents, run in sequence, with explicit handoffs between them.

1. 1

   Data Collection Agent

   Pull the week

   Collects the key metrics from your configured sources — user growth, revenue, activity and retention — and records provenance (source, frequency, last update) for every value.

   Handoff → Raw metric set + source ledger

2. 2

   Data Cleaning Agent

   Trust the inputs

   Checks for anomalies and missing values, annotates each with a candidate root cause and severity, and repairs what can be repaired deterministically (documenting every repair).

   Handoff → Cleaned dataset + anomaly register

3. 3

   Analysis Agent

   Find the signal

   Computes week-over-week and year-over-year change, classifies multi-week trend direction, and distils the movements into ranked insights with confidence.

   Handoff → Trend series + insight table

4. 4

   Report Agent

   Ship the artefact

   Assembles the four sheets — Data Overview, Trend Analysis, Anomaly Annotation, Operations Recommendations — into one Excel workbook ready to send.

   Handoff → weekly-ops-report.xlsx

What you receive

One Excel workbook, four sheets, structured the same way every week.

Features

• ✓Four-agent pipeline (collect → clean → analyze → report)
• ✓Four-sheet Excel: Overview / Trend / Anomaly / Recommendations
• ✓WoW and YoY change with trend classification
• ✓Anomaly detection with root-cause candidates
• ✓Priority-ranked operations recommendations

Use cases

Compared with doing it by hand

When not to use it. Weekly Ops Report does not replace a warehouse or a BI tool for exploratory analysis. It produces one structured weekly artefact; use your BI stack for open-ended slicing.

Data sources

The pipeline reads from four logical source groups. In the demo these are simulated with a documented generator; in production you point them at your own systems.

• User growth

  Signups, activation, funnel conversion

• Revenue

  Total, subscription, ads, value-added, ARPU

• Activity

  DAU, WAU, MAU, session length, opens

• Retention

  D1, D7, D30, NPS

Security & privacy

• • Inputs are processed to produce your report and are not retained as a dataset.
• • No values are invented: every reported metric traces to a collected value.
• • Anomaly explanations are candidates to verify, never confirmed root causes.
• • Payment is handled by Waffo Pancake; card data never touches this application. See the Privacy page.

Pricing

See full pricing and checkout →

Frequently asked questions

What does Weekly Ops Report actually produce?

A four-sheet Excel workbook: Data Overview (metric dashboard), Trend Analysis (multi-week trends and insights), Anomaly Annotation (flagged values with root-cause candidates), and Operations Recommendations (ranked P0/P1/P2 actions).

Do I need to connect my real data?

For the demo, no — the pipeline runs on documented simulated sources. For production you connect your own sources; inputs are processed to produce the report and are not retained as a dataset.

Does it invent numbers?

No. Every metric in the report traces to a collected value. Anomaly explanations are labelled as candidates, not facts.

What is the agent workflow?

Four specialised agents in sequence: Data Collection (pull metrics from sources), Data Cleaning (detect anomalies and missing values), Analysis (compute WoW/YoY, identify trends), Report (assemble the Excel workbook).

Is there a free tier?

Yes — rule-based report generation with a fair-use daily cap. Model-assisted narrative insights are on paid plans.

What language is supported?

Interface, prompts and report output are English (v1.x).

Is my data stored?

Inputs are used only to produce your report. See the Privacy page for retention and your rights.