GBS Tower Lead Guide:
Shared Services & GBS

Selecting the right technology stack is critical for GBS maturity. The market is crowded, and the wrong choice can lock you into years of technical debt. Here is a neutral evaluation of the primary architectural approaches.

1. The Platform Approach (Service Management Layers)

This involves overlaying a unified service management platform (like ESM tools) across all towers (Finance, HR, IT).

• Pros: Single 'pane of glass' for the business; unified reporting; easier to implement cross-functional workflows (e.g., onboarding involves IT, HR, and Finance).

• Cons: High initial licensing costs; requires integration with core systems of record (ERPs, HRIS).

• Best For: Mature GBS organizations managing multiple towers who need to improve customer experience (CX).

2. The ERP-Native Approach

Leveraging the service modules built into your existing ERP or HRIS ecosystem.

• Pros: Lower integration friction; data stays in the system of record; often included in existing enterprise licenses.

• Cons: Poor user experience (UX) for business partners; limited capability to handle cross-functional requests (e.g., an ERP tool handles Finance well but struggles with IT requests).

• Best For: Single-function shared services (e.g., Finance-only SSCs) heavily reliant on a single ERP.

3. Point Solutions & Best-of-Breed

Buying specific tools for specific problems (e.g., a specialized AP automation tool, a separate ticketing tool for IT).

• Pros: Deep functionality for specific tasks; rapid deployment for isolated problems.

• Cons: Creates data silos; fragmentation of the user experience (users need 5 logins); nightmare for global reporting.

• Best For: Solving urgent, specific operational fires (e.g., a tax compliance crisis) where speed outweighs integration.

Build vs. Buy Considerations

• Buy: Standard processes (AP, Payroll, Helpdesk). The market maturity is high; building your own is wasteful.

• Build: Strategic differentiators or highly proprietary industry-specific workflows.

Evaluation Checklist

When vetting vendors, GBS Tower Leads should ask:

1. Integration: 'Do you offer pre-built connectors to our specific ERP versions, or is this a custom API build?'

2. Intake: 'Can your tool ingest and structure requests from email and Teams/Slack natively?'

3. Regionality: 'How does your platform handle data residency requirements for China and the EU simultaneously?'

4. Time-to-Value: 'What is the average time from contract to first transaction processed?' (Target: <4 months).

FAQs

How long does a typical GBS transformation take to show ROI?

While a full transformation is a 12-18 month journey, you should expect to see 'Quick Win' ROI within 4-6 months. This typically comes from the 'Stabilize' phase where you eliminate the 'Hidden Factory' of unstructured work. By simply routing email traffic into a structured triage system, organizations often see a 10-15% capacity release immediately. However, deep financial ROI (20%+ savings) usually matures around the 12-month mark once automation scaling begins. If you aren't seeing trend-line improvements by month 6, you are likely stuck in 'Pilot Purgatory'.

Do I need to replace my existing ERP to fix the intake problem?

Absolutely not. In fact, replacing the ERP is often the wrong place to start. The most effective approach is to implement a 'Service Layer' or 'Engagement Layer' that sits *on top* of your ERPs. This layer handles the intake, triage, and orchestration, while the ERP remains the system of record. This 'wrap and extend' strategy is faster (months vs. years), cheaper, and less risky than a rip-and-replace ERP project. It allows you to unify the experience without disrupting the core accounting or HR engines.

How do we handle the resistance from business units who want to keep their 'special' processes?

Data is your best weapon against emotional resistance. Do not argue opinions; argue metrics. Show them the 'Price of Variation.' For example: 'Standard Process A costs $12 per transaction and takes 2 days. Your Custom Process B costs $45 and takes 5 days.' Give them the choice: they can keep their special process, but they must pay the fully loaded cost for it (Chargeback Model). Most business units will voluntarily standardize once the true cost of their customization is revealed on their P&L.

How does GenAI fit into the GBS roadmap if our data isn't clean?

You cannot effectively deploy GenAI on dirty data. However, you can use GenAI to *clean* your data. This is a critical distinction. Use GenAI early in the roadmap for 'Intake Triage'—interpreting unstructured emails and mapping them to structured fields. This cleans the data at the source. Do not attempt high-stakes GenAI use cases (like autonomous financial forecasting) until your foundational data governance is mature (typically Phase 3). Start with 'Assisted' AI (Copilots) before moving to 'Autonomous' AI.

Should we build our own automation capabilities or outsource to a vendor?

The trend for 2025 is toward a 'Hybrid Center of Excellence.' You should build a small, core internal team of 'Architects' who own the governance, security, and strategy. This ensures you retain the IP and control. However, for the execution (coding the bots, training the models), it is often more cost-effective to leverage partners or vendors who can scale up and down. Relying 100% on outsourcing creates a 'Black Box' where you lose understanding of your own operations; relying 100% on internal build is often too slow and hard to recruit for.