RPA CoE Setup
An RPA Centre of Excellence (CoE) is a dedicated team and governance structure that owns, manages, and scales the organisation's RPA programme. The CoE sets standards, maintains best practices, governs which processes get automated, builds bots, manages the platform, and measures results. Without a CoE, RPA projects grow chaotically — different teams use different tools, different standards, and duplicated effort.
Think of the CoE as the headquarters of your automation army — strategy, standards, and operations all under one roof.
Why a CoE Matters
WITHOUT CoE: WITH CoE:
─────────────────────────────────────────────────────────
Different departments buy One platform, one licence
different RPA tools agreement, volume discount
Every developer codes differently Shared standards, reusable
libraries, faster builds
No central visibility into Dashboard: all bots, all
what bots are running metrics, one view
Bots break and nobody owns them Clear ownership and SLAs
for every bot
Nobody measures ROI Centralised benefit tracking
and reporting to leadership
CoE Operating Models
Organisations structure their CoE in one of three ways depending on their size and maturity:
Centralised CoE
A single team handles all RPA activities for the entire organisation — discovery, development, deployment, and operations. All requests go through the CoE.
- Pros: Consistent standards, economies of scale, clear ownership
- Cons: Can become a bottleneck as demand grows; business teams lose autonomy
- Best for: Organisations starting their RPA journey (under 50 bots)
Federated CoE
A central CoE team sets standards and manages the platform. Individual business units have their own RPA developers who build automations for their departments, following the central standards.
- Pros: Scales well; business teams move faster; CoE focuses on governance
- Cons: Requires strong governance to prevent inconsistency
- Best for: Mature programmes with 50+ bots
Hybrid CoE
Complex bots are built by the central CoE. Simpler automations are built by citizen developers in business units using low-code tools, under CoE guidance.
CoE Team Structure
┌─────────────────────────────────────────────────────────┐ │ RPA CoE TEAM │ │ │ │ ┌──────────────┐ ┌──────────────┐ │ │ │ CoE LEAD / │ │ RPA BUSINESS │ │ │ │ PROGRAMME │ │ ANALYST │ │ │ │ MANAGER │ │ (Discovery, │ │ │ │ (Strategy, │ │ PDD, UAT) │ │ │ │ Governance, │ └──────────────┘ │ │ │ Stakeholders│ │ │ └──────────────┘ ┌──────────────┐ │ │ │ RPA DEVELOPER│ │ │ ┌──────────────┐ │ (Build, test,│ │ │ │ PLATFORM / │ │ deploy) │ │ │ │ INFRA ADMIN │ └──────────────┘ │ │ │ (Orchestrator│ │ │ │ VMs, access)│ ┌──────────────┐ │ │ └──────────────┘ │ RPA SOLUTION │ │ │ │ ARCHITECT │ │ │ ┌──────────────┐ │ (Design, │ │ │ │ OPERATIONS │ │ standards) │ │ │ │ ANALYST │ └──────────────┘ │ │ │ (Monitor, │ │ │ │ support) │ │ │ └──────────────┘ │ └─────────────────────────────────────────────────────────┘
CoE Responsibilities
Strategy and Governance
- Define the RPA strategy aligned with business goals
- Maintain the automation pipeline — prioritise and approve new automations
- Set development standards, naming conventions, and quality gates
- Report automation ROI and pipeline progress to leadership
Platform Management
- Own and manage the Orchestrator environment (dev, UAT, production)
- Manage bot machine infrastructure and licences
- Control access and permissions across the platform
- Manage RPA tool upgrades and patches
Development Services
- Conduct process discovery and feasibility assessments
- Create PDDs and SDDs
- Build and test bots following the implementation lifecycle
- Maintain shared component libraries and reusable workflows
Operations and Support
- Monitor all live bots and respond to alerts
- Manage exception queues and escalations
- Perform corrective and preventive maintenance
- Maintain SLAs for bot uptime and performance
CoE Governance: The Intake Process
The Intake Process is how new automation requests enter the CoE pipeline. Without a formal intake process, every department submits requests informally and the team cannot prioritise effectively.
AUTOMATION INTAKE PROCESS:
─────────────────────────────────────────────────────────
Business unit submits request via intake form
│
▼
CoE BA reviews for basic suitability (1–2 days)
│
▼
Feasibility assessment and ROI calculation (1–2 weeks)
│
▼
CoE Steering Committee reviews and prioritises
│
▼
Approved: Added to development backlog with priority rank
Rejected: Feedback provided with reasons
│
▼
Development scheduled based on capacity and priority
Measuring CoE Success
| Metric | Description | Target |
|---|---|---|
| Bots in Production | Total live automations | Grows quarter on quarter |
| FTE Equivalent Saved | Total hours saved ÷ working hours per FTE | Tracked per bot and in total |
| ROI | Net benefit ÷ total investment | > 200% in year 1 |
| Automation Pipeline | Number of approved automations awaiting build | Healthy backlog, not excessive |
| Bot Success Rate | Successful transactions across all bots | > 95% |
| Time to Automate | Average days from intake to go-live | Decreasing over time |
Summary
An RPA CoE is the organisational structure that turns isolated bot projects into a scalable, governed automation programme. It brings together strategy, development, platform management, and operations under one team with clear responsibilities. The operating model — centralised, federated, or hybrid — evolves as the programme matures. A formal intake process prioritises automation requests objectively. Measuring CoE performance through FTE savings, ROI, and bot success rates demonstrates value to leadership and justifies continued investment. A strong CoE is what separates organisations with 5 bots from those with 500.