Scaling RPA in a Regulated Bank: Governance and CoE
A mid-sized bank has successfully deployed 15 RPA bots in its retail banking division over the past year, achieving 20% reduction in manual effort for account opening and loan processing. The bots are managed by a central RPA team of 5 people. Now, the bank wants to scale RPA to 100 bots across retail, commercial, and wealth management divisions within 18 months. Each division has its own IT and compliance teams. The bank operates in a regulated environment with strict data privacy and audit requirements. The central RPA team proposes a federated CoE model where each division has a dedicated RPA champion and the central team provides standards, tools, and training. However, the central team is concerned about maintaining control and avoiding shadow IT. They also need to ensure bots comply with changing regulations. The bank's CIO is skeptical about the cost and wants a clear ROI model. The central team has a budget of $2 million for the next 18 months.
1. Evaluate the proposed federated CoE model versus a centralized CoE model for this bank. What are the risks and benefits of each? Recommend a model and justify your choice, considering the regulatory environment and scaling targets.
Read the worked answer
Centralized CoE:
- Benefits: Strong control, consistent standards, easier compliance, centralized expertise, economies of scale.
- Risks: Bottleneck for 100 bots, slow response to division-specific needs, potential lack of business ownership.
Federated CoE:
- Benefits: Faster deployment, business alignment, distributed ownership, scalability.
- Risks: Inconsistent standards, shadow IT, compliance gaps, duplication of effort.
Recommendation: A hybrid model: central CoE sets standards, provides shared services (e.g., tooling, training, compliance framework), and maintains a registry of all bots. Each division has a champion responsible for pipeline and local governance, but with mandatory adherence to central standards and regular audits. This balances control and speed. Given the regulatory environment, central oversight of compliance is non-negotiable. The central team should focus on enabling, not gatekeeping. For scaling to 100 bots, a purely centralized model would be a bottleneck; federated with strong central governance is optimal.
Justification: The bank needs both speed and compliance. A hybrid model allows divisions to move quickly while ensuring central visibility and control. The central team can use a 'governance as a service' approach, providing templates, automated compliance checks, and training. This also addresses CIO's ROI concern by distributing costs and accelerating benefits.
2. The CIO wants a clear ROI model for the $2 million investment. What factors should be included in the ROI calculation? How would you structure the business case to gain approval?
Read the worked answer
ROI factors:
- Costs: Technology licenses (RPA software, AI services), infrastructure, central team salaries, training, change management, compliance audits, ongoing maintenance (typically 15-20% of development cost per year).
- Benefits: Labor cost savings (FTE reduction or redeployment), error reduction (cost of rework), cycle time improvement (revenue impact), scalability (avoided hiring), compliance risk reduction (avoided fines).
- Intangibles: Employee satisfaction, customer experience.
Structure:
- Phase the investment: pilot in one division, then scale.
- Show quick wins: automate high-volume, low-complexity processes first to demonstrate ROI.
- Use a phased budget: allocate $500k for first 6 months for 20 bots, then scale.
- Include a sensitivity analysis: show ROI under different adoption rates.
- Benchmark against industry: typical RPA ROI is 6-12 months for simple processes.
- Emphasize risk mitigation: compliance bots reduce regulatory risk, which has high cost.
Judgment: The $2M for 100 bots over 18 months is aggressive. Average cost per bot including licenses and development might be $20k-$50k, so 100 bots could cost $2M-$5M. The budget may be insufficient. The business case should prioritize processes with highest ROI and consider a phased approach. Also, include a governance cost to avoid shadow IT risks. The CIO will want to see a payback period of less than 18 months. So focus on processes with high manual effort and error rates.