Case Study — Insurance PMO

From 50 hours of manual analysis to 5 minutes with AI

How ResonAi built an AI agent that automates cross-project failure analysis for a Canadian insurance company's Program Management Office.

93%
Time reduction
7
Projects analyzed
5
Repeat offenders found
100%
Extraction accuracy
The challenge

A critical report that took weeks to produce

The client's PMO produces a "Repeat Offenders" report — a lessons learned analysis that identifies recurring failures across all completed projects. This report drives corrective actions and accountability for systemic issues.

Before ResonAi

  • A senior PM manually reads every project closure report
  • Cross-references issues across all projects to find patterns
  • Categorizes findings by theme, audience, and severity
  • Writes the final report in a specific corporate format
  • Process takes 30–50+ hours per cycle
  • Report produced only once per year — too late to act

After ResonAi

  • AI agent reads all closure reports automatically
  • Extracts and categorizes every improvement issue
  • Cross-references patterns and links to historical data
  • Generates the report in the exact corporate format
  • Complete analysis in under 5 minutes
  • Enables quarterly or monthly reporting cycles
How it works

Three-stage AI pipeline

The agent uses a purpose-built prompt chain — each stage feeds the next, producing progressively refined output from raw documents to presentation-ready report.

1
Extract
Read each closure report, extract issues with categories
2
Analyze
Cross-reference patterns, flag repeat offenders
3
Generate
Produce the full report in corporate format
Results

What the AI found across 7 projects

The agent analyzed 7 project closure reports and identified 5 repeat offenders — including patterns that had persisted for over 5 years without resolution.

#Repeat OffenderSeverityStatusTracked Since
1High Code Defects Across ProjectsHighNew2024
2UAT Resource & Environment ReadinessMediumContinues2024
3Development Environment SharingMediumContinues2021
4Circular Business DecisionsMediumContinues2021
5External Stakeholder Communication GapsMediumContinues2020

See the agent in action

This is a simulation of the actual AI agent built for the client. Click the buttons to see how it works.

R
ResonAi Agent
Online — PMO Repeat Offenders Analysis
R
Welcome! I'm the Repeat Offenders Agent. I analyze project closure reports to identify recurring issues across your portfolio.

What would you like to do?
Return on investment

The business case

Direct labor savings are significant, but the strategic value — catching repeat failures earlier — is where the real ROI lives.

Time savings per cycle

93%
From 30–50 hours to under 5 minutes of AI processing + 2–4 hours of PM review

Reporting frequency

Quarterly instead of annual — patterns caught while there's still time to act

Cost avoidance potential

$50K–$200K+
Per prevented project repeat failure (rework, overruns, team productivity)

Payback period

1 cycle
Investment recovered within the first reporting period
Technology

Built on existing infrastructure

No new platforms. No external tools. The entire solution runs within the client's existing Microsoft 365 environment — Teams, SharePoint, and Copilot Studio.

What was built

  • 4 purpose-built AI prompts (extraction, pattern analysis, report generation, quick analysis)
  • 7 conversation topics with button-based navigation
  • SharePoint integration for dynamic document access
  • PowerPoint generation script for presentation-ready output

ResonAi methodology applied

  • Assess — Mapped the manual workflow and closure report structure
  • Prioritize — Validated RO report as the highest-value first build
  • Design — Engineered prompts, taxonomy, and agent architecture
  • Enable — Built, tested, and deployed with team training

Ready to turn AI into
operational advantage?

Every organization has processes like this — manual, time-intensive, and waiting for AI to make them better.

Book an AI Strategy Call