PathFinder AI
Intelligent incident prioritisation platform that uses explainable AI to score, correlate, and recommend actions on IT incidents. Built for UK SMEs who can't afford a 20-person NOC.
Status
Early access
AI Scoring
Explainable
Pricing
Coming soon
Built & Hosted
UK
The Problem
Every SME with more than a handful of services eventually hits the same wall: alert fatigue. Monitoring tools fire hundreds of alerts per week. Most are noise. But buried in that noise are the critical incidents that could take down your business. Without a round-the-clock team, you either monitor everything (and burn out) or ignore most alerts (and miss the one that matters).
What We Saw Managing 4 Production Platforms
The Reality
- • Hundreds of alerts a week across our own platforms
- • Most were informational or duplicates
- • Many required investigation but no action
- • Only a small fraction were genuinely actionable incidents
- • Triage cost minutes per alert, every alert
The Cost
- • Hours every week spent on triage alone
- • P1 incidents missed because they looked like P3
- • On-call engineers burning out
- • No consistent scoring = inconsistent response
- • Enterprise tools priced well beyond SME budgets
How We Built PathFinder AI
The Core Idea: Explainable AI Scoring
Instead of another dashboard that shows you everything, PathFinder AI tells you what matters and why. Every incident gets an AI-generated score (0-100) with a plain-English explanation: "Scored 87/100 because: database connection errors correlate with 3 previous P1 outages, affects payment processing service, occurred during peak hours."
Step 1: Ingest
Connect your monitoring tools (email alerts, Slack, webhooks, API). PathFinder AI ingests everything into a unified incident stream.
Step 2: Score
AI analyses each incident using 15+ signals: historical patterns, service criticality, time of day, blast radius, correlation with recent changes.
Step 3: Act
Get prioritised recommendations: auto-resolve noise, escalate P1s immediately, group related alerts into a single incident thread.
The AI Scoring Engine
Not a black box. Every score comes with a breakdown so your team understands why an incident was flagged as critical or dismissed as noise.
Signal Analysis
15+ weighted signals scored in real-time:
- ✓ Service criticality tier (revenue-impacting services score higher)
- ✓ Historical incident patterns (has this failed before?)
- ✓ Blast radius estimation (how many users/services affected?)
- ✓ Time correlation (peak hours vs maintenance windows)
- ✓ Change proximity (was anything deployed in the last 2 hours?)
Correlation Engine
Groups related alerts automatically:
- ✓ 5 "disk space low" alerts from same cluster → 1 incident
- ✓ API timeout + database slow query + CPU spike → correlated root cause
- ✓ Duplicate suppression (same alert fired 10 times in 2 minutes)
Recommendations
Actionable next steps, not just alerts:
- ✓ "Auto-resolved: CPU spike normalised after 3 minutes"
- ✓ "Escalate: Payment gateway timeout affecting checkout — similar to P1-2024-031"
- ✓ "Investigate: Unusual login pattern detected — possible credential compromise"
Core Features
Multi-Source Ingestion
Email, Slack, webhooks, PagerDuty, Opsgenie, ServiceNow, custom APIs
AI Severity Scoring
0-100 score with plain-English explanation for every incident
Alert Correlation
Automatically groups related alerts into single incident threads
Smart Escalation
Route P1s to on-call, suppress noise, auto-resolve transients
Team Dashboard
Real-time MTTR, incident volume, response times, team workload
Runbook Integration
Link runbooks to incident types for instant response guidance
Post-Incident Reports
Auto-generated timeline, impact analysis, and improvement suggestions
Custom Scoring Rules
Override AI defaults for your specific infrastructure and services
Shift Management
On-call schedules, rotation management, coverage gaps detection
SLA Tracking
Monitor response/resolution times against your SLA commitments
Mobile Alerts
Push notifications for P1/P2 with one-tap acknowledge
API & Webhooks
Full REST API for automation and custom integrations
How PathFinder AI Compares
| Feature | PathFinder AI | PagerDuty | Opsgenie |
|---|---|---|---|
| AI Incident Scoring | Explainable | Basic | Not included |
| Alert Correlation | Automatic | Manual rules | Basic |
| Setup Time | Hours, not weeks | Weeks | Weeks |
| Pricing (SME) | SME-focused, coming soon | Enterprise pricing | Enterprise pricing |
| UK Data Residency | Yes | US-based | US-based |
| Explainable Recommendations | Yes | Not included | Not included |
| No Enterprise Lock-in | Monthly | Annual contracts | Annual contracts |
Pricing: Coming Soon
PathFinder AI is currently in early-access beta. It will launch with tiered plans designed for teams of every size — simple, transparent, and SME-focused.
Starter
For small teams getting started
- ✓ Multi-source ingestion
- ✓ AI scoring & correlation
- ✓ Email + Slack notifications
- ✓ Email support
Professional
For growing operations teams
- ✓ Advanced AI + custom rules
- ✓ Unlimited integrations
- ✓ On-call scheduling
- ✓ Post-incident reports
- ✓ API access
Enterprise
For larger teams and MSPs
- ✓ Multi-tenant support
- ✓ SSO / SAML
- ✓ Dedicated account manager
- ✓ Custom SLA
Join the waitlist to be among the first to know pricing and availability.
Why We Built PathFinder AI
We needed it ourselves
Managing our own production platforms generating a constant stream of alerts, we needed a way to separate signal from noise without hiring a round-the-clock operations team.
Enterprise tools don't fit SMEs
Enterprise incident tools are designed for large engineering teams. They carry enterprise price tags, require weeks of setup, and assume you have dedicated SRE staff. Most UK SMEs have 1-3 people managing IT alongside everything else.
AI should explain itself
"AI-powered" tools that just show a number aren't useful. If an incident scores 87/100, you need to know why. Our scoring model provides plain-English reasoning so your team can validate the AI's judgment and learn from it.
Technical Architecture
Stack
Frontend: Next.js 14, TypeScript, Tailwind CSS
Backend: Node.js / Express API
AI Engine: Custom scoring model (rule-based + ML hybrid)
Database: PostgreSQL with time-series extensions
Queue: Redis for real-time alert processing
Infrastructure: Docker, Traefik, automated CI/CD
Hosting: UK-based data centres
Design Principles
✓ Explainability over accuracy — every score has reasoning
✓ Fast setup — no consulting engagement needed
✓ SME-first pricing — no enterprise tax
✓ UK data residency — GDPR compliant by default
✓ Open integrations — works with what you already use
✓ Progressive complexity — start simple, add rules over time
PathFinder AI Is in Early Access
Join the waitlist for early access to the beta programme. No credit card required.
Visit pathfinderai.co.uk →Start with a straight answer.
A fixed-price healthcheck tells you exactly where your IT and security stand — devices, backups, email security and where your data sits — in a plain-English report within five working days. Credited against your first month if you go on to a retainer.