How to Build a Center of Excellence for Automation

Most automation programs fail the same way: they start with promising pilots, generate early excitement, then quietly stall out. The root cause isn't technology but rather the absence of a structure to govern, scale, and measure automation as a strategic capability. That structure is a center of excellence (CoE).

Without an automation center of excellence, organizations tend to hit predictable walls:

• Different teams automate the same process in different ways
• Nobody knows which automations are running, breaking, or delivering value
• Development standards vary wildly, creating technical debt and security exposure
• Business stakeholders lose confidence because results don't match promises

With a CoE, organizations can create the conditions for automation to move from scattered experiments to cohesive enterprise-wide impact.

This article explains how to build a CoE that works and is ready to advance to agentic automation, guided by best practices from organizations that have built successful automation programs. You'll learn how to structure roles and governance, how to implement a framework that scales, what metrics matter and what they mean for CoE maturity, and how purpose-built platforms support the full lifecycle.

What is a center of excellence?

A CoE is more than a team, though it includes one. It's more than a department, though it may be structured as one.

At its core, a CoE is an operating model: the set of structures, processes, and decisions that determine how automation gets done at scale.

The CoE team decides what gets automated and in what order. They set the standards developers follow, track whether automations actually work and deliver value, determine who can build what, and how the organization learns from each implementation.

Without this operating model, it is difficult for organizations to keep track of what's been built, what works, or what's broken — let alone scale automation. When someone asks, "What's our automation ROI?" finding the answer takes weeks (and a spreadsheet archeology project).

An automation center of excellence (CoE) is an operating model, including dedicated teams, governance structures, and standardized processes, that determines how automation gets built, deployed, and scaled across an organization.

The evolving CoE

Until recently, most CoEs started with pure robotic process automation (RPA), creating automations to handle structured, repetitive work like data entry, report generation, and moving information between disconnected systems. With traditional automation, the CoE's job is primarily technical: build automations that work reliably, establish coding standards, and track results like time saved.

As automation technology and programs mature, the CoE's scope expands beyond individual automations to orchestrating enterprise automation capabilities. Adding AI to this foundation has pushed CoEs to evolve further, moving beyond "Can we automate this task?" to transforming processes end-to-end. And now, agentic AI is driving an even greater shift.

What agentic automation means for the CoE

Agentic AI redraws the boundaries of automation, making the limitations of structured workflows obsolete. It processes unstructured information, makes decisions, and takes action, handling exceptions that would trip up traditional automation by learning from patterns in data to adjust behavior. For example, an agentic system can understand a customer email, determine the right response, take action across multiple systems, and escalate only when absolutely necessary.

For CoEs, this means processes that were too complex, too variable, or too dependent on judgment become candidates for automation. It expands automation to the complex work that traditional approaches can't touch — and expands the scope of the CoE to evaluating when AI is appropriate (and when it isn't), designing systems where humans and AI agents collaborate effectively, and establishing governance that balances autonomy with control.

Why you need a center of excellence

Whether you're working with traditional RPA or expanding into agentic AI, your CoE defines the difference between deploying technology and delivering business outcomes.

Without structured governance, what started as efficiency gains turns into technical debt, compliance risk, and wasted investment as different teams build redundant solutions and inconsistent practices create security gaps.

The CoE's value is providing the unifying structure that makes automation programs successful. The benefits of this structure stem from five core capabilities: Governance, scalability, visibility, efficiency, and continuous improvement./p>

Risk reduction with standardized governance

When different teams build automations their own way, it creates technical, security, and compliance problems that nobody understood they were creating.

A CoE establishes one set of standards for development, testing, deployment, and maintenance. It ensures every automation meets security requirements, follows regulatory guidelines, and aligns with internal policies before it goes into production.

CoEs track governance effectiveness with:

• Percentage of automations following development standards
• Security audit pass rates
• Compliance incidents related to automation
• Time to remediate policy violations

Scale and consistency that multiply value

Without a CoE, finance may build an invoice processing automation that saves 200 hours per month. Meanwhile, HR and procurement each build their own versions. You've now built (and paid for) the same core solution three times.

With a CoE, this invoice processing logic becomes a template.

CoEs capture what works, document it, and provide the means for teams to apply it elsewhere, consistently. That means the second deployment takes weeks instead of months, and the third happens in a matter of days. And so on.

Key scalability indicators include:

• Number of automations replicated across business units
• Time to deploy similar automations (second, third, fourth instances)
• Percentage of new automations using reusable components
• Reduction in redundant development effort

Visible results and ROI

Without a CoE, the answer to "What's our return on investment?" takes time and effort to come up with, and the number you eventually produce lacks credibility because it wasn't tracked systematically.

A CoE centralizes tracking from day one: Every automation has defined metrics, and every metric gets measured. Communicating ROI is straightforward and defensible because it's based on real-time data on portfolio health, value delivered, and what's coming next, rather than post-facto estimates.

CoEs monitor ROI metrics such as:

• Total hours saved across all automations
• Cost savings and cost avoidance by business unit
• Error rate reduction in automated processes
• Forecasted value vs. realized value (accuracy of predictions)

Accelerated value through deployment efficiency

CoEs provide standard methods and reusable components to reduce rework and drive faster development. This pairs with centralized infrastructure that eliminates redundant setup, so developers spend time solving new problems instead of rebuilding what already exists.

CoEs follow efficiency gains with:

• Average time from idea to production deployment
• Development time reduction for similar automation types
• Developer time spent on new development vs. maintenance
• Automations delivered per developer (compared to baseline)

Performance that improves over time

CoEs create a closed loop for improving automation standards and performance by centralizing feedback and propagating optimizations/new versions. They centralize performance data and post-implementation reviews so that insights can feed back into standards and practices, so that each automation improves the next one.

This matters more as programs mature. The difference between a mediocre automation and an excellent one is whether it keeps improving or degrades with time — CoEs create the conditions for ongoing optimization.

Continuous improvement gets measured by:

• Frequency of optimization cycles
• Year-over-year improvement in automation performance
• Lessons learned documented and applied
• Rate of new technology adoption

Running automation as a business function, with clear organizational structure, defined roles, and accountability for these outcomes, is what separates successful programs.

Roles and structure of an automation CoE

The roles and structure of the CoE work best when they balance strategic direction with operational execution. Most successful CoEs organize around three layers, each with separate responsibilities that support the others.

1: Strategic layer: Leadership and Governance

This top-level layer consists of executives and steering committee members who set direction. They define the vision for automation, establish policies that govern its use, and determine the KPIs that measure success.

This layer answers questions like: Which business objectives should automation support? How much should we invest? Where do we draw boundaries on what can be automated? What risks are we willing to accept?

Typically, this group meets quarterly to review portfolio performance, approve major initiatives, and adjust direction to align with business priorities.

2: Tactical layer: CoE core team

Here's where strategy becomes execution. The core team manages the automation pipeline, deciding what gets built and when. They establish the standards everyone follows. They coordinate resources, troubleshoot problems, and ensure delivery happens on schedule.

This layer manages day-to-day CoE operations: prioritizing incoming requests, overseeing development and deployment, monitoring automation health, and sharing knowledge across the organization. This is the operational heart of the CoE.

3: Operational layer: Business units

Business units aren't part of the CoE structure, but they are central to its success. They identify automation opportunities in their domains, provide process expertise, validate solutions, and ultimately realize the value of automation.

Mature CoEs cultivate automation champions within these units. Champions understand both business processes within their domain and what automation is capable of, which makes them effective as bridges between their departments and the CoE, translating needs in both directions.

Essential roles within the CoE

Executive Sponsor: Usually a C-level executive or senior VP who provides strategic oversight. Their role is to secure funding, resolve conflicts that escalate beyond the CoE, and communicate automation's strategic importance to the organization.

It's important that the sponsor has enough authority to drive cross-functional initiatives and influence how resources get allocated.

CoE Lead / Program Manager: Runs the CoE day-to-day. This person oversees the automation pipeline and prioritization, manages the team, tracks ROI, coordinates with business units and IT, and establishes governance.

Strong program management matters here, but it also requires technical understanding to make informed decisions, along with enough political skill to navigate stakeholder relationships.

Solution Architects / Developers: The automation builders. They assess technical feasibility, design automations, write code that follows CoE standards, and troubleshoot when things break.

Mature CoEs mix experience levels — senior architects who handle complex designs, mid-level developers who execute most builds, and junior developers who learn while contributing.

And architects also act as mentors for citizen developers when the CoE enables broader participation.

For agentic automation, development roles also involve agent orchestration to coordinate multiple agents working together, managing dependencies between agents, and preventing conflicts when agents have competing objectives. This goes beyond workflow design to understanding how autonomous systems collaborate.

And this extends to understanding AI model behavior — the ability to evaluate when AI is appropriate and when it isn't, understand model limitations and failure modes, and assess whether agent decisions align with business intent. Note that this isn't about coding but rather understanding how AI reasons and where it breaks down.

Business Analysts: Bridge the gap between business needs and technical solutions. They identify opportunities, document processes, calculate expected ROI, and verify whether the value actually materializes after implementation.

Strong analysts help to prevent a common pitfall: building automations that are technically impressive but solve the wrong problem.

Automation Champions: Representatives within business units who advocate for automation in their areas. They spot opportunities, facilitate access to SMEs, help prioritize from a business perspective, and support adoption.

Champions typically aren't full-time CoE members, but they're indispensable for gaining grassroots momentum.

Agentic AI roles: As CoEs expand into agentic automation, additional role types are joining the core team roster. Specifically, observability/explainability specialists and AI ethics and compliance roles.

Starting size for your CoE

How large should your CoE team be initially? The answer is likely to be smaller than you may think. Many successful CoEs start with three to five people: a program manager, one or two developers, and a business analyst.

As the program demonstrates value and demand grows, add capacity and specialized roles — infrastructure engineers, security specialists, change management professionals. For agentic automation, add observability specialists and ethics officers as agent deployments start to scale.

Building an automation CoE framework

No matter the starting size of your CoE, expect it to grow and change as your capabilities mature and business needs shift.

With the goal of making the CoE journey as effective as possible, this framework walks through five essential steps, from defining governance through scaling operations. All of these pieces form a strong CoE foundation that’s ready to expand into agentic automation.

Step 1: Define the governance model

Before building any automation, first answer three questions: Who decides? How do they decide? What happens when things go wrong? This, in a nutshell, is what your governance model is for.

CoE ownership and reporting structure

The first step in setting up your governance model is determining where the CoE sits organizationally. The answer shapes everything from budget authority to decision speed to stakeholder relationships.

Three models work in practice:

• CoEs that report to IT tend to have a strong technical governance and risk management focus. This works well when automation is primarily about system integration and infrastructure reliability.
• CoEs that report to operations or business transformation get faster alignment on business priorities. This works when automation is strategic to the business model.
• CoEs that report to a digital transformation office balance both approaches, connecting automation to broader transformation initiatives while maintaining technical discipline.

None is inherently better. The right choice depends on where automation expertise lives today, where budget authority sits, and whether your organization views automation as primarily technical capability or business transformation.

From a governance perspective, the key is to decide and document the reporting relationship explicitly, including how the CoE interacts with business units, IT, security, compliance, legal, and HR. Ambiguity here causes friction later when you need those groups to make choices and/or act quickly.

Decision rights

Who has final say when two departments compete for the same automation resources? Who approves deviations from standards when a business urgency crops up? Who can pull developers off a project to address an emergency? Ideally, these questions will have answers before conflicts arise.

Start by creating a RACI matrix (Responsible, Accountable, Consulted, Informed) for major decision types: automation prioritization, technical architecture, resource allocation, exception handling, and investment decisions. The format matters less than the clarity — the goal is that everyone knows exactly who makes which decisions.

For agentic automation, add decision rights for agent accountability.

• Who authorizes agents to make autonomous decisions?
• At what monetary thresholds do agents require human oversight?
• Who decides which AI models to use and approves their training data?
• When an agent makes a wrong decision, who has the authority to shut it down, investigate what happened, and determine when it can restart?

Because agentic systems can affect customers, employees, partners, and regulators, governance requires broader representation than traditional automation, which means adding the right teams to review and approval processes. For example, legal and compliance would be part of agent deployment approvals, and customer advocacy should be at the table when agents make customer-facing decisions.

Step 2: Establish intake and prioritization

Ad hoc requests and political pressure can derail even the best-run CoE if there isn't an accepted process for saying "yes," "not yet," or "no" to automation ideas.

Standardize intake

Build an intake form that captures both technical requirements and business value. Ask for a detailed process description with current pain points, estimated volume and frequency, systems and applications involved, and expected business impacts/benefits.

Make the intake process accessible to anyone in the organization, not just licensed automation platform users, to build a healthy automation pipeline that reflects the needs and opportunities of the entire organization. Consider choosing a platform that includes configurable intake forms that can be submitted by any employee, removing barriers to identifying candidate processes.

Scoring and prioritization

Define objective criteria for evaluating opportunities. Common criteria include:

• Expected ROI (time savings, cost reduction, revenue impact)
• Strategic alignment (supports key business objectives)
• Technical complexity (effort required to implement)
• Implementation risk (technical, operational, change management)
• Business impact (number of people affected, criticality of process)

Weight these factors based on business priorities, for example, a scoring model might weight ROI at 40%, strategic alignment at 25%, technical feasibility at 20%, and business impact at 15%.

For agentic automation, additional scoring factors become relevant, in particular:

• Decision complexity (does this require judgment calls, or can rules handle it?)
• Exception handling needs (how much variation exists in decision-making?)
• Risk tolerance (what happens if the agent makes wrong decisions — financial exposure, regulatory risk, reputation impact?)
• Data readiness (is training data available, sufficient, representative, and clean?)
• Explainability requirements (will agent decisions need to be explained to regulators, customers, or internal auditors?)

Document the scoring model and communicate it clearly so stakeholders understand how decisions get made, so that when the CoE deprioritizes someone's request, they understand why based on objective criteria.

Review cadence

Regular pipeline reviews help maintain the health of the program and keep ideas from becoming stale. Set a schedule for when the CoE team evaluates new submissions using the scoring model, updates scores for existing pipeline items based on current business needs, commits resources to top-priority items, and communicates decisions.

Step 3: Develop standards and playbooks

Standardization is what transforms a collection of individual automations into a scalable program. Without standards, every developer does things differently, creating maintenance potholes and barriers to scaling.

Development standards

Document detailed guidelines covering:

• Naming conventions — consistency enables searching, troubleshooting, and maintenance
• Code structure and organization (where do configurations live, how do you separate business logic from infrastructure?)
• Error handling and logging (what errors get caught, what gets logged, where, and for how long?)
• Security and credential management (which vault, who has access, how do you rotate secrets?)
• Documentation requirements (what must be documented, where does it live, how is it maintained?)
• Version control practices (which system, what branching strategy, how do you tag releases?)

Make standards detailed enough to ensure consistency but flexible enough to accommodate different automation types. Include examples of good and bad implementations so developers understand both the rules and the reasoning behind them.

For agentic automation, add standards for:

• Model selection criteria (when do you use which AI models, what performance thresholds must models meet?)
• Training data management (how do you source, validate, version, and monitor for drift?)
• Decision transparency (how do you document agent reasoning, what confidence thresholds trigger review?)
• Agent testing protocols (how many scenarios, what edge cases, who approves results?)
• Monitoring requirements (what behaviors get monitored, what patterns trigger alerts, who responds, and how quickly?)

Reusable component libraries

Create libraries of reusable components for common tasks like credential management, email and database operations, API integrations, and error handling. Document each component clearly with purpose and use cases, input/output specifications, configuration options, usage examples, known limitations, and maintenance requirements.

This library becomes increasingly valuable as the program matures. New automations can leverage these components rather than rebuilding functionality from scratch.

End-to-end delivery methodology

The aim here is to have consistent project execution regardless of who's doing the work.

Document the complete delivery process — from opportunity identification through post-production support — and include templates for each component. And for agentic automation, there will be additional delivery phases for agent decision points to ensure governance, risk management, and quality.

Documentation requirements

Like for delivery methods, good documentation supports long-term sustainability by enabling others to maintain and enhance automations when original developers move on. This is something that becomes even more important for agentic systems, where understanding decision logic matters.

Require comprehensive documentation for every automation:

• High-level process overview (what problem does this solve, who benefits, what's the business value?)
• Technical architecture (how does it work, what systems does it touch, what are the dependencies?)
• User instructions (how do you run it, what inputs does it need, what outputs does it produce?)
• Troubleshooting guides (what can go wrong, how do you diagnose problems, how do you fix them?)
• Maintenance procedures (how do you update it, how do you test changes, when does it need updates?).

For agentic automation, add documentation for decision authority boundaries and reasoning, as well as model specifications (which model, what version and data inputs). Document oversight and incident response protocols too, so that who monitors, how often, and what triggers review/response is well established.

Step 4: Measure outcomes

The adage “what gets measured gets improved” is especially true for centers of excellence. In fact, a central part of the value of the CoE is the centralized monitoring and measurement of automation performance.

That means measuring outcomes is meant to be built into your CoE framework from the beginning. Without systematic measurement, you can't demonstrate value, optimize performance, or justify continued investment.

Track value realization by comparing projected savings to actual savings. Monitor whether automations are being used as expected. Conduct post-implementation reviews 30 days, 90 days, and six months after deployment to verify that benefits materialized, and apply this realization data to improve benefit projections.

Apart from being the CoE’s mandate, making measurable accountability a core pillar of CoE operations builds credibility with business stakeholders — when you say an automation will save $200K and you can later prove it did, stakeholders will trust your next business case.

Define KPIs upfront

CoEs are the gatekeepers here, making sure baseline metrics are clear before implementation so actual improvement is measured, not estimated after the fact.

For each automation, establish success metrics before any development begins. These might include hours saved per month (transaction volume × time savings per transaction), error rate reduction (baseline error rate minus post-automation error rate), processing time improvement (baseline processing time minus automated processing time), cost savings or avoidance (labor cost reduction plus error cost reduction plus capacity cost avoidance), and customer satisfaction impact (survey scores, NPS changes, complaint reductions).

For agentic automation, metrics will expand to include decision quality, autonomous decision rate, escalations, decision transparency, and model performance stability (how consistently does decision quality maintain over time?).

Role-specific dashboards

Share accurate performance by implementing dashboards according to stakeholder needs.

Executive dashboards will focus on portfolio-level ROI/payback period and business impact: Revenue enabled, cost reduced, customer experience improvements. They also need to see risk indicators like security incidents, compliance violations, and high-severity failures.

Where possible, connect automation metrics to broader business KPIs:

• How does reducing invoice processing time contribute to days sales outstanding (DSO)?
• How does automating customer onboarding impact customer acquisition cost (CAC)?
• How does eliminating manual data entry affect employee retention in high-turnover roles?
• How does faster quote generation influence win rates?

These connections help executives — and the organization — see beyond operational efficiency to the strategic value and impact of automation.

Business unit dashboards will focus on their specific automations and current status, realized benefits versus forecasts, upcoming opportunities in their pipeline, usage patterns, and adoption metrics.

CoE leaders need a set of dashboards covering pipeline, resource use, automation delivery, and quality metrics.

Step 5: Scale and evolve

Perhaps it goes without saying, but a successful CoE is one that continuously evolves based on lessons learned, changing business needs, and new technology capabilities. The rule is to make sure automations don’t become "set it and forget it" — they should improve over time.

Use data from your automation platform to identify patterns and opportunities. Which types of automations deliver the highest ROI? Double down on those patterns. Where are common failure points? Invest in making those more robust. Which business processes show the most automation potential? Prioritize those areas.

The theme is the same: Let data guide strategic decisions wherever possible.

To that end, make reviews a proactive part of CoE routine — for example, schedule quarterly reviews of high-value automations to identify optimization potential and create feedback mechanisms for business users, IT, and executive sponsors alike. Use feedback to refine CoE processes, adjust priorities, and identify new capabilities to build.

Gather business user feedback with quarterly surveys on automation usefulness, ease of use, and pain points, and establish regular touchpoints with IT partners to review infrastructure performance, integration challenges, and security concerns.

Likewise, set a regular cycle for communicating with executive sponsors with formal reviews on strategic value, investment priorities, and capability gaps.

The evolution to agentic automation

In terms of advancing capabilities, most CoEs evolve through traditional RPA to adding AI for specific capabilities like document understanding, and then to AI agent-driven automation.

Introducing AI agents typically starts with assisted agentic workflows where agents make recommendations but humans make decisions, then to human validation before execution, and then, where possible, to autonomous agentic workflows (that operate within defined boundaries).

As of early 2025, about 65% of large enterprises are in the early phases with AI agents — experimentation to pilots — with only 11% achieving full deployment. This suggests most CoEs are still establishing governance for agentic systems rather than operating them at scale.

The implication here is that if you're just starting to think about agentic automation governance, you're not behind — yet. But organizations that define frameworks now, in particular for governance, before scaling agentic deployments, will avoid painful retrofitting later when agents are already making autonomous decisions in production.

Measuring the success of a CoE

The difference between a struggling CoE and a thriving one often comes down to whether leadership understands what the numbers mean, not just that they exist.

Step 4 in the framework covered how to build measurement into CoE operations. This section addresses what those measurements actually tell you about CoE health, how to interpret them over time, and how measurement capabilities themselves evolve as your CoE matures.

Understanding CoE metrics

Automation pipeline value

This is a ratio of forecasted-to-realized value. A healthy pattern shows realized value meeting or slightly exceeding forecasts, suggesting accurate estimation with occasional upside. You should expect the ratio to improve over time as your CoE learns what drives value and gets better at estimating it.

If realized value consistently runs at 60% of forecasts, it can signal that the team is overselling benefits to get projects approved. While building momentum is important, you don’t want to erode it later with underwhelming results.

If realized value consistently exceeds forecasts by 20-30%, it’s likely estimates are too conservative, leaving value on the table during prioritization, or you're genuinely discovering additional benefits during implementation that weren't visible upfront.

ROI and payback period

Early-stage CoEs often see longer payback periods — around twelve months — versus individual automation projects, because they're building foundational capabilities and establishing credibility that will enable CoE-led automation to quickly outpace and outperform disconnected projects.

As the CoE matures, payback periods should shrink. Mid-stage CoEs typically achieve six to twelve month payback as they leverage reusable components and established patterns. CoEs sometimes see payback in three to six months for certain automation types because they've thoroughly optimized delivery.

The main thing is to see payback periods shortening over time. If that’s not happening, it’s a red flag. Either complexity is increasing faster than capability — as in, you're tackling harder problems without sufficient skill growth — or efficiency isn't improving, which is often from missing learning from past implementations.

Process throughput and quality

These metrics show whether automation is transforming operations or only marginally improving them:

• Cycle time reduction (before vs. after automation)
• Error rates in automated processes
• Processing volume capacity
• Exception handling rates

For example, invoice processing dropping from three days to two days is a significant but still incremental improvement; dropping from three days to three hours is transformational. The size of improvement can reveal whether you're automating at the surface or redesigning core processes.

Consistency across automations also signals program health. If some automations deliver 10x improvements while others deliver 10%, there could be a problem with how you’re selecting processes to automate, or automations are built at surface level only, without reimagining processes before automating them.

Adoption rate

Adoption patterns signal how the organization trusts your CoE. Slow adoption despite demonstrated ROI suggests change management problems or perceived complexity.

Rapid adoption in some business units but resistance in others often indicates that champions matter more than actual results. It isn’t unusual for key personalities to have an outsized impact — but what you want is for proven value to be clear to everyone, regardless of who is championing your cause.

Broad, sustained adoption across multiple business units indicates the CoE has built systemic credibility. People believe automation works because they're seeing it work, time and again.

Cost savings and hard versus soft benefits

The ratio between hard savings (actual cost reduction) and soft benefits (improved satisfaction, faster decisions) can help gauge CoE maturity.

Early-stage CoEs often rely heavily on soft benefits because they're easier to claim but harder to verify. And mature CoEs showcase hard savings because they are tackling more substantive automations.

This is an opportunity to assess your business case mix. Leaning into soft benefits beyond year one can indicate capability and/or credibility gaps causing the CoE to avoid harder problems that deliver more concrete value.

Enabling effective measurement

Spreadsheets work for five automations, but at five hundred automated workflows, they're simply not an option. As programs scale, you will need the right CoE tools to support effective measurement.

Automation Anywhere CoE Manager is made for measuring CoE performance — it creates a unified view of your automation program from idea submission through value realization, pulling data automatically from automations in production.

This automated tracking means when executives ask, "What's our ROI?" you have answers right away.

Configurable stakeholder views

CoE Manager's dashboard customization enables each audience to see what matters to them without excess detail. You configure business information at the outset — process names, departments, savings calculations — and it generates appropriate views automatically:

• Executives can see portfolio-level ROI calculated automatically from actual automation performance, strategic alignment against business objectives, and risk indicators flagging issues requiring attention.
• CoE leads can see pipeline velocity showing how quickly opportunities move through stages, resource use showing bottlenecks and capacity constraints, and delivery metrics tracking efficiency trends.
• Business units can see their specific automations with realized benefits compared to forecasts, upcoming opportunities in their queue, and adoption patterns showing actual usage.

You can create multiple dashboard copies for different analytical views, add custom fields specific to your business, and publish different versions to different audiences, all from the same underlying data.

Connecting automation to business outcomes

CoE Manager's integration with business intelligence platforms like Power BI and Tableau allows you to see automation impact in the context of overall business performance. Export data in HTML, CSV, or JSON formats, or use connector URLs and FastTable API for direct integration with BI tools.

This means you can correlate automation deployment with customer satisfaction trends, analyze automation ROI alongside revenue performance by business unit, track automation's contribution to strategic KPIs in executive dashboards, and create custom visualizations that answer specific leadership questions.

The platform also automatically pulls automation performance data to calculate actual ROI without manual data entry. This native integration eliminates the gap between automation execution and measurement, as savings calculations update as automations run.

Tracking CoE progress

As your CoE matures, what you measure and how you interpret it will evolve. With effective measurement and analysis, you create a continuous loop of improvement — revealing performance, patterns, and guiding decisions; decisions drive improvements, which lead to better performance and metrics.

Quarterly maturity assessments are a standard approach to assess progress across dimensions to understand your current state and plan deliberate next steps. Use quarterly reviews to spot gaps between current state and where you need to be in six months, set specific improvement goals, track progress over time, and adjust investment priorities based on areas of weakness/opportunity.

CoEs with this loop in place will keep optimizing, demonstrating value, and evolving with purpose. The goal is to advance based on your needs. For example, an early-stage CoE wouldn’t dive into advanced governance when basic automation delivery is still inconsistent.

How Automation Anywhere CoE Manager supports your CoE

The strategic framework described here provides the blueprint for CoE success. Executing effectively hinges on having the right tools.

Automation Anywhere CoE Manager is purpose-built to operationalize CoE governance, visibility, and measurement from a single environment, and it’s made for managing automation programs, not a general project management tool adapted for automation.

Intake and prioritization

Traditional CoE intake processes often require platform licenses, limiting participation. CoE Manager removes this barrier with configurable intake forms accessible to any employee (no license required) so that front-line workers can easily submit automation ideas.

Employees describe their process, explain pain points, estimate volume, and identify expected benefits. Submissions flow directly into the CoE's pipeline.

And it supports opportunity evaluation through configurable scoring models where you define criteria — ROI potential, strategic alignment, technical complexity, business impact — weight them according to priorities, and the platform automatically scores each opportunity. Opportunities move through clear lifecycle stages — Ideas, Pipeline, In Progress, Deployed — with tracking across all stages for visibility into pipeline health.

Portfolio management

CoE Manager provides dashboard visibility across the entire automation lifecycle, displaying idea velocity, opportunities by stage, forecasted savings versus targets, value by stage, and benefits realization comparing forecasted to actual.

Visualization options give CoE leads a single source of truth to identify bottlenecks, rebalance resources, and track progress.

ROI and benefits realization

CoE Manager links delivery timelines directly with savings projections, then tracks actual realization. Integration with Automation Anywhere Control Room enables automatic ROI calculation, pulling actual performance data and calculating realized savings.

The platform tracks person hours saved per automation, total cost savings, cost savings per automation and process, success ratios, monthly ROI trends, and comparison of forecasted versus realized benefits.

Custom reporting and presentations

The platform automates report generation entirely so CoE teams can produce status reports, executive presentations, and stakeholder updates in minutes. It also provides role-based access control for customizable report views so that each stakeholder sees relevant information.

Integration ecosystem

CoE Manager integrates with Power BI and Tableau through data export in HTML, CSV, or JSON formats, or via connector URLs and FastTable API for direct integration. This enables automation dashboards alongside other business intelligence, combining automation metrics with broader performance data. Native Control Room integration automatically pulls performance data, and integration with discovery pulls in opportunities identified through process mining to submit to CoE Manager with one click.

Product tiers

• CoE Manager works with the Automation Anywhere platform, providing intake and prioritization, portfolio management, ROI tracking via Control Room integration, and support for up to three Control Room instances.
• CoE Manager Plus supports multi-vendor automation environments, with unlimited Control Room integrations, native integrations for Microsoft Power Automate and UiPath, more customization, and advanced analytics. Organizations can use CoE Manager Plus without Automation Anywhere, making it suitable for heterogeneous automation environments.

Both versions are powered by Shibumi platform and include role-based access control, built-in approval workflows, comprehensive audit trails, and SOC 2 Type 2 certification compliance.

Real-world customer results with CoE Manager

A global conglomerate that had automated 140 processes across 11 business functions lacked visibility into activities and couldn't quantify contributions. With CoE Manager, they created custom dashboards measuring performance, improvements, and ROI — and could now quantify their results: 230,000 hours automated, 95% automation efficiency, and 24/7 execution of core business functions.

CoE Manager is available out of the box with minimal setup, enabling organizations to start managing their automation pipeline within days. Get started now — talk to us about your CoE’s needs or set up a live demo.

FAQs

What defines a center of excellence?

A center of excellence is an operating model — not just a team or department — that establishes governance, best practices, and standards to enable a specialized capability across an organization. In the context of automation, the CoE includes dedicated teams, governance structures, and standardized processes that determine how work gets done consistently and at scale. Unlike operational departments that execute tasks, a CoE focuses on establishing standards, sharing knowledge, driving continuous improvement, and enabling others across the organization.

What is the purpose of the CoE?

The CoE transforms scattered automation efforts into a cohesive, value-driven program that delivers measurable business impact. It ensures governance and standardization across initiatives, scales successful automations efficiently across departments, provides visibility into performance and ROI, accelerates deployment through reusable components and methodologies, and drives continuous improvement through systematic learning. Without a CoE, organizations face duplicated efforts, inconsistent standards, invisible ROI, and automation programs that stall after initial pilots despite early success.

What are examples of centers of excellence?

Organizations establish CoEs around strategic capabilities to support business priorities. Common examples include automation and RPA CoEs that govern automation programs, data and analytics CoEs that establish data governance and promote data-driven decision-making, and customer experience CoEs that standardize approaches to improving customer interactions. Other examples include cloud CoEs that guide cloud adoption and optimization, cybersecurity CoEs that establish security standards and practices, and agile or DevOps CoEs that promote modern software development practices. Each shares the common purpose of establishing excellence through governance, standardization, and knowledge sharing rather than performing operational tasks.

What's the difference between an RPA CoE and an Automation CoE?

The key difference lies in scope and strategic focus. An RPA CoE concentrates specifically on robotic process automation— software that automates rule-based, repetitive tasks involving structured data and defined business rules. An automation CoE takes a broader view, encompassing RPA plus workflow automation, business process management, integration platforms, and increasingly, advanced capabilities including AI, machine learning, and document understanding. While an RPA CoE would focus on deploying bots efficiently, an automation CoE looks at end-to-end process transformation, considering which combination of technologies delivers optimal business outcomes. Many organizations start with an RPA CoE and evolve it into a broader automation CoE over time.

How large should a CoE team be to start?

A starting CoE can be surprisingly lean; many successful programs begin with just three to five people. The typical initial structure includes a program manager to lead the CoE and manage stakeholder relationships, one or two automation developers to build solutions, and a business analyst to evaluate opportunities and document requirements. As the program demonstrates value and scales, the team grows to include additional developers, architects, process analysts, and specialized roles like infrastructure engineers, security specialists, or change management professionals. The key is starting with the right skills and clear roles rather than a large team, then scaling resources deliberately as automation adoption grows and ROI is proven.

How can I demonstrate ROI from my automation CoE?

Demonstrating ROI requires systematic measurement from the start. First, establish clear baseline metrics before implementing automations: Current processing times, error rates, labor hours, and costs. Second, define expected benefits for each automation during planning, using conservative estimates to maintain credibility. Third, track actual benefits realization after deployment, measuring both hard savings (quantifiable cost reduction) and soft benefits (employee satisfaction, faster decision-making).

Research shows typical RPA payback periods range from 3 to 9 months, with expected ROI from 30% to 200% in the first year. Fourth, calculate comprehensive ROI by comparing total benefits against all costs, including platform licenses, development resources, infrastructure, and ongoing maintenance. Finally, use tools like Automation Anywhere CoE Manager to maintain dashboards showing portfolio-level ROI, individual automation performance, and forecasted versus realized value, enabling immediate answers when executives ask about returns.

How does a CoE evolve as automation maturity increases?

CoE evolution typically follows a predictable path aligned with automation maturity. In the initial stage, the CoE focuses on tactical execution, focused on building automations, establishing basic standards, and demonstrating quick wins to build momentum and trust.

As the program scales, the CoE shifts to standardization and optimization, developing reusable components, formalizing governance, and scaling proven automations across business units. In the mature stage, the CoE becomes a strategic enabler, driving business transformation through automation, enabling citizen development while maintaining appropriate controls, and integrating automation with broader digital transformation initiatives.

In the most advanced stage, the CoE evolves into an innovation hub, exploring emerging technologies like process mining and agentic AI, and establishing automation as a core competency that shapes business strategy. This evolution typically happens over two to four years, with early-stage CoEs laser-focused on how to build automations correctly, and mature CoEs debating which business problems to solve next.

What tools support governance and reporting in a CoE?

Effective CoE governance requires purpose-built tools because spreadsheets and homegrown systems aren’t reliable at scale. Automation Anywhere CoE Manager is designed specifically for automation governance, offering intake and prioritization workflows, portfolio management and tracking, ROI calculation and benefits realization monitoring, and customizable dashboards with automated reporting.

CoE Manager integrates with complementary tools, including process mining platforms like Celonis to identify opportunities, project management tools like Jira for development tracking, business intelligence platforms like Power BI and Tableau for advanced analytics, and IT service management systems like ServiceNow for incident management. This integrated ecosystem provides end-to-end visibility from opportunity identification through ongoing performance monitoring and eliminates the manual overhead that typically burdens automation teams.

How does Automation Anywhere CoE Manager simplify CoE operations?

CoE Manager eliminates the manual overhead that can consume days every month for a CoE team. Instead of maintaining spreadsheets for opportunity tracking, the platform provides automated intake workflows with configurable forms and scoring accessible to any employee (no license requirements). Rather than compiling ROI reports manually from multiple sources, it generates comprehensive dashboards automatically with real-time data pulled from Control Room integration. It provides integrated portfolio management with data visualizations and dependency tracking.

The platform's role-based dashboards ensure executives see portfolio-level ROI, CoE leads see pipeline health and resource utilization, and business units see their specific automations and realized benefits. Customer results show that CoE Manager transforms automation from an IT project into a visible, governed business function that demonstrates clear strategic value with minimal administrative burden.