Blue Prism
Insight
A multi-tab operational intelligence dashboard built for Blue Prism CoEs. Surfaces fleet health, queue status, per-process RAG scoring, exception trend analysis, infrastructure monitoring, and ROI reporting — none of which is available natively in Control Room.
Blue Prism's native tooling — Control Room and Process Analytics — gives you individual bot statuses and session logs. It doesn't give a CoE the unified operational picture it needs to run the programme confidently. Exception logs are flat lists with no trend analysis. There's no ROI reporting, no infrastructure visibility, and no way to answer "are we healthy right now?" without clicking through multiple screens.
For a mature CoE running 200+ processes, that gap is a real operational risk. Problems surface late, effort is spent manually compiling MI for leadership, and infrastructure degradation often isn't noticed until bots start failing.
Built a Streamlit dashboard that plugs directly into the BP v7 REST API and Zabbix, designed for three distinct audiences: automation teams who need to operate the estate day-to-day, leadership who need ROI and MI, and admins who manage access.
The dashboard ships with a full mock data provider — JSON fixtures covering every view — so it runs out of the box for demos and UAT without any BP access. Switching to live data is a single environment variable change. No code changes, no demo setup, no data scrubbing.
"Are we healthy right now?"
Control Room shows individual bot statuses, but there's no single view correlating fleet health, queue backlogs, infrastructure metrics, and schedule adherence. Insight's Overview tab puts all of this on one screen with live status pills, fleet-wide KPIs, and clearance ETAs for every monitored queue.
"Which exceptions actually need attention?"
BP's exception logs are flat lists. Insight adds trend analysis (spiking vs normal), velocity tracking, severity ranking, and Business vs System classification — so you can prioritise by impact, not just recency. A spiking System exception from a single SAP timeout is different from a rising Business exception rate across five processes.
"What's the ROI of our automation programme?"
There's no native ROI reporting in Blue Prism. Insight models cost avoidance against your annual programme cost, with per-process breakdowns, contribution analysis, and plain-English interpretation of ROI ratios for non-technical stakeholders. Exportable as CSV.
"Is our infrastructure about to cause problems?"
BP doesn't monitor its own infrastructure. Insight pulls server metrics from Zabbix and derives health status from real thresholds — CPU, RAM, disk, and Windows service states. It highlights degradation before bots start failing, not after.
Fleet-wide KPI cards, per-pool digital worker grid with status indicators, work queue cards with progress bars and clearance ETAs, and upcoming schedule table — all on one screen.
Per-process RAG health scoring against configurable STP and exception rate thresholds, AHT and volume trends, and an Exception Intelligence section with trend-aware KPIs, velocity tracking, and Business vs System classification.
Projected annual cost avoidance, ROI ratio with plain-English interpretation, per-process contribution analysis, cumulative savings over time, and CSV export for reporting. Designed for a leadership audience with no automation background.
Server health cards with derived status from real CPU, RAM, disk, and service metrics. API endpoint response times and database connection health — all pulled from Zabbix.
Bot utilisation bar chart, hourly activity heatmap showing when each bot is working, and daily trend lines. Fleet-wide KPIs with configurable thresholds. Date range filtering.
Mock/live provider swap
A factory pattern switches between static JSON fixtures and the live BP API + Zabbix via a single environment variable. The dashboard runs fully out of the box for demos and UAT — no BP environment, no data scrubbing, no manual setup. One config change and it's pointing at production.
Derived status, not declared status
Server health cards compute status from real CPU, RAM, disk, and Windows service metrics against configurable thresholds. A server isn't 'healthy' because no one reported a problem — it's healthy because the numbers are within bounds. Any stopped service triggers Critical regardless of resource usage.
Security by design
bcrypt password hashing, JWT session cookies, brute-force protection (5 attempts), server-side session invalidation on password reset and user removal, role-based tab access, and structured audit logging of every auth event. Designed to pass a basic security review before going on a corporate network.
- Architecture brief and component design
- Multi-tab Streamlit dashboard — Overview, Process Intelligence, Infrastructure, Utilisation, ROI & MI, Admin
- Mock data provider with JSON fixtures — fully runnable without a Blue Prism environment
- Live data provider — Blue Prism v7 REST API + Zabbix JSON-RPC integration
- Role-based authentication — admin, automation, leadership — with bcrypt hashing and JWT session cookies
- Server-side session registry with invalidation on password change and user removal
- Per-process RAG scoring with configurable STP and exception rate thresholds
- Exception Intelligence — trend analysis, velocity tracking, Business vs System classification, switchable time windows
- ROI & MI reporting — projected cost avoidance, ROI ratio, per-process contribution table, cumulative savings charts, CSV export
- Infrastructure health monitoring — derived server status from Zabbix metrics, API response times, database connection health
- Bot utilisation heatmap, bar charts, and trend lines with date range filtering
- Admin panel — user management (create, reset password, remove) with role assignment
- IIS + NSSM deployment guide for Windows Server environments
- 128-test suite with 92% coverage (pytest + pytest-cov), GitHub Actions CI pipeline (ruff + pytest)
- SECURITY.md — OWASP Top 10 mapping and pre-deployment hardening checklist