Cutting competitor response time from two weeks to twelve minutes

The challenge

This bank's life insurance division was consistently late to competitor pricing and product moves. By the time competitive intelligence reached the product team, it had been assembled by hand from fragmented sources and was already stale, so whatever response followed was aimed at a market that had already moved on.

Tracking competitors was a manual, inconsistent process, stitched together from scattered public sources rather than a live, structured view the team could rely on. Market blind spots were common, with pricing and product changes sometimes only spotted once they'd already affected the division's own position.

The division wanted live visibility of what competitors were doing, and the ability to act on it in minutes rather than weeks, without building a team of people to watch competitor websites around the clock.

Our approach

1. 01

   Mapping competitors and building monitoring agents

   We began by mapping the competitors and product categories that mattered most, then built AI monitoring agents that scan competitor sites and public content continuously, orchestrated on a schedule through Power Automate rather than relying on anyone checking manually.

2. 02

   Building the differential detection engine

   We built a differential detection engine that tracks what has actually changed between scans and scores each change for relevance, so the team sees material pricing and product moves rather than being buried in noise from routine page updates.

3. 03

   Creating a structured intelligence repository

   Every detected change, its source and its assessment is stored as structured content in a SharePoint intelligence repository, giving the division a running, searchable history of competitor activity in place of a scattered series of one-off emails.

4. 04

   Routing alerts and hooking into internal workflows

   We built Power Automate flows that push prioritised alerts directly into Microsoft Teams and raise items in the product, legal and compliance review workflows, with escalation rules applied automatically based on severity.

5. 05

   Building the Power Apps review interface

   We built a Power Apps interface where product and legal reviewers triage, action and sign off on alerts, with every decision written back to SharePoint Lists rather than tracked separately outside the system.

6. 06

   Personalising tone and building in continuous learning

   We matched agent messaging tone and urgency to the division's internal language and procedure, then calibrated detection and expanded coverage over time, with the system learning from every review decision so alerts keep getting more relevant the longer it runs.

The outcome

Response time has dropped from around two weeks to roughly 12 minutes from detection to action, with three major pricing updates made within hours of a competitor move being detected.

Signal-to-noise on competitive intelligence has improved by 87%, and more than 200 escalations a month are now handled without manual triage, freeing the team from sifting through fragmented sources by hand.

Legal and product teams now treat the AI system as frontline monitoring rather than a reporting afterthought, trusting alerts written in the firm's own voice enough to act on directly, and relevance keeps improving as the system learns from every internal review decision made.